{"passed":true,"scope":"Real HTTPS / ASR / ECAPA / local+API LLM / TTS diagnostics.\n\nDigital PCM injection uses the actual browser playback clock. Not a physical\nmicrophone/acoustic echo certification. Every run is archived, including failures.\nPublic ESC-50 inputs stay local; the receipt includes provenance, never raw audio.\n","started":"20261027T134007Z","group":"all","source_sha256":"f086cee87229863d2051e6ca20d413cc6cd281f533f2bb8f03543345f408b82a","injections":[{"browser_start_ms":4900.299999237061,"browser_end_ms":9900.199996948242,"packets":250,"max_schedule_lag_ms":1.8999977111816406,"mean_schedule_lag_ms":0.09879975891113281,"stream_start_sample":38400,"samples":80000,"sha256":"474c118b913bd8c001d69ff27613cee094b8afecd9cff9c1a044bba901e41189"},{"browser_start_ms":5003.600002288818,"browser_end_ms":21002.79999923706,"packets":800,"max_schedule_lag_ms":63.69999694824219,"mean_schedule_lag_ms":1.7046244955062866,"stream_start_sample":6400,"samples":256000,"sha256":"3225a2bc288de755d3bdc17787d198b6e61fc8463ae5a76d7c775ee957fda44d"},{"browser_start_ms":4124.899997711182,"browser_end_ms":9124.29999923706,"packets":250,"max_schedule_lag_ms":5.600002288818359,"mean_schedule_lag_ms":0.8124002075195312,"stream_start_sample":38400,"samples":80000,"sha256":"7d3b3ad57d3c01aeecde256dd1656177f845391c1b11f9ee4e501537fe7d0bab"},{"browser_start_ms":4012.2999992370605,"browser_end_ms":9011.5,"packets":250,"max_schedule_lag_ms":68.10000228881836,"mean_schedule_lag_ms":3.3448005981445315,"stream_start_sample":38400,"samples":80000,"sha256":"38d153ad9d1a314e615fd8c84459cef4b000b3c044af3397e12529e02b16a028"},{"browser_start_ms":3919.900001525879,"browser_end_ms":8939.20000076294,"packets":250,"max_schedule_lag_ms":61.400001525878906,"mean_schedule_lag_ms":1.8007999420166017,"stream_start_sample":38400,"samples":80000,"sha256":"2d035982c15f176648ffe3107236fcb1df5e864c9b0bace93ff89c63bbe1f6ce"},{"browser_start_ms":3906.1000022888184,"browser_end_ms":8905.900001525879,"packets":250,"max_schedule_lag_ms":3.5999984741210938,"mean_schedule_lag_ms":0.3991997528076172,"stream_start_sample":38400,"samples":80000,"sha256":"048b79fbc94b62cce2f131873bf1bace053c75b9eb6cc3942e509ad6fa1275e7"},{"browser_start_ms":3955.7000007629395,"browser_end_ms":6954.799999237061,"packets":150,"max_schedule_lag_ms":7.799999237060547,"mean_schedule_lag_ms":0.7546662902832031,"stream_start_sample":41600,"samples":48000,"sha256":"917ef60d2530297772bfcf56a58dc10c27a6188586bea9316f0e5875fefb3cb4"},{"browser_start_ms":6420,"browser_end_ms":9419.29999923706,"packets":150,"max_schedule_lag_ms":6.299999237060547,"mean_schedule_lag_ms":0.7153329213460287,"stream_start_sample":41600,"samples":48000,"sha256":"917ef60d2530297772bfcf56a58dc10c27a6188586bea9316f0e5875fefb3cb4"},{"browser_start_ms":3902.5,"browser_end_ms":7386.099998474121,"packets":174,"max_schedule_lag_ms":7.899997711181641,"mean_schedule_lag_ms":0.8793101913627537,"stream_start_sample":38400,"samples":55520,"sha256":"b196013e2f7f5bac54a115f3a2d7621c01c79a8ddb66e453fa62d20f80c99321"},{"browser_start_ms":7820.200000762939,"browser_end_ms":12872.70000076294,"packets":250,"max_schedule_lag_ms":61.79999923706055,"mean_schedule_lag_ms":0.782399658203125,"stream_start_sample":99520,"samples":80000,"sha256":"474c118b913bd8c001d69ff27613cee094b8afecd9cff9c1a044bba901e41189"},{"browser_start_ms":4454.299999237061,"browser_end_ms":8894.400001525879,"packets":222,"max_schedule_lag_ms":4.5,"mean_schedule_lag_ms":0.29549579792194536,"stream_start_sample":38400,"samples":70880,"sha256":"5d665c310e221c7e3af7ac9ed35c4762517c8a14d93056ea055ebafca2b766b7"},{"browser_start_ms":22799.70000076294,"browser_end_ms":29159.70000076294,"packets":318,"max_schedule_lag_ms":4,"mean_schedule_lag_ms":0.5729557493197843,"stream_start_sample":323680,"samples":101600,"sha256":"422772ed90db71170764837bcd8dbc42cf8ed947e0cd93d788cb37ca1b2fd150"},{"browser_start_ms":33243.59999847412,"browser_end_ms":37202.79999923706,"packets":198,"max_schedule_lag_ms":5.400001525878906,"mean_schedule_lag_ms":0.6040408201891967,"stream_start_sample":489280,"samples":63200,"sha256":"dd2ca90a85488b86997a05c74e6fe3017ea091f6832a627d073e141fc3738249"},{"browser_start_ms":51821.70000076294,"browser_end_ms":56840.70000076294,"packets":250,"max_schedule_lag_ms":34.79999923706055,"mean_schedule_lag_ms":1.5207996520996094,"stream_start_sample":779680,"samples":79840,"sha256":"4dcb7fd02b14b2e29cbc064502468df8c18d7b862839701a500f21f6821559fd"},{"browser_start_ms":64483.90000152588,"browser_end_ms":68443.5,"packets":198,"max_schedule_lag_ms":6.899997711181641,"mean_schedule_lag_ms":0.7550501582598446,"stream_start_sample":977920,"samples":63200,"sha256":"dd2ca90a85488b86997a05c74e6fe3017ea091f6832a627d073e141fc3738249"}],"cases":[{"case":"music_onset","events":[{"type":"interaction","enabled":true,"helpful_enabled":false,"rules":[],"decision_model":"convaiinnovations/laya-multilingual (Stop) · dialogue model (context)","decision_interval_ms":320,"checks":0,"last_decision":null,"waiting_for":"No agreed rules yet","attempts":0,"semantic_error":null,"scope":"session-only conversational rules; acoustic verification is separate","at":2856.199996948242},{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.44,"tokens":35,"preprocess_ms":10.75,"encoder_gpu_ms":8.08,"decoder_gpu_ms":274.54,"inference_ms":282.82,"total_ms":293.57,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":38.72,"window_s":3.44,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1342},{"label":"Drum machine","score":0.1279},{"label":"Synthesizer","score":0.1235},{"label":"Keyboard (musical)","score":0.1138},{"label":"Electric piano","score":0.1082}],"music_score":0.9159,"speech_score":0.0116,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8257},{"label":"Ringtone","score":0.0555},{"label":"Musical instrument","score":0.0549}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1042},{"label":"Synthesizer","score":0.0667}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Music","score":0.8667},{"label":"Musical instrument","score":0.1296},{"label":"Drum machine","score":0.1279}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Music","score":0.8513},{"label":"Musical instrument","score":0.1342},{"label":"Synthesizer","score":0.1235}]},{"start_s":2.16,"end_s":3.44,"candidates":[{"label":"Music","score":0.8948},{"label":"Musical instrument","score":0.1224},{"label":"Synthesizer","score":0.0759}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":46.98},"type":"sound_caption","request_ms":310.26,"capture_ms":3240,"event_age_ms":3550.26,"kind":"music","eligible":true,"suppression_reason":null,"at":8493.29999923706},{"type":"environment_observation","kind":"music","status":"responding","at":8495.399997711182},{"type":"latency_start","reply_id":1,"input":{"kind":"environment","sound_caption_ms":310.26},"at":8495.5},{"type":"reply","text":"Oh, da läuft gerade ein ziemlich treibender elektronischer Track.","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Oh, da läuft gerade ein ziemlich treibender elektronischer Track.","start_s":0,"end_s":3.06}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Oh, da läuft gerade ein ziemlich treibender elektronischer Track.","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":628.26,"finish_reason":"stop","usage":{"prompt_tokens":2376,"completion_tokens":20,"total_tokens":2396,"cost":0.0023582196,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00238204,"upstream_inference_prompt_cost":0.00235224,"upstream_inference_completions_cost":2.98e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":711.14,"generation_calls":1,"llm_total_ms":711.18,"response_complete":true,"sentence_ready_ms":724.17,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":9223.199996948242},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":628.26,"finish_reason":"stop","usage":{"prompt_tokens":2376,"completion_tokens":20,"total_tokens":2396,"cost":0.0023582196,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00238204,"upstream_inference_prompt_cost":0.00235224,"upstream_inference_completions_cost":2.98e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":711.14,"generation_calls":1,"llm_total_ms":711.18,"response_complete":true,"answer_first_audio_ms":1526.53,"tts_first_audio_ms":802.17,"first_audio_ms":1526.54},"at":10026.099998474121},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":628.26,"finish_reason":"stop","usage":{"prompt_tokens":2376,"completion_tokens":20,"total_tokens":2396,"cost":0.0023582196,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00238204,"upstream_inference_prompt_cost":0.00235224,"upstream_inference_completions_cost":2.98e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":711.14,"generation_calls":1,"llm_total_ms":711.18,"response_complete":true,"answer_first_audio_ms":1526.53,"tts_first_audio_ms":802.17,"first_audio_ms":1526.54,"tts_server_stages":[{"generation_ms":1492,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1673,"performance":{"device_seconds":1.1991299324035645,"raw_audio_seconds":4.16,"aggregate_rtf":0.28825238759701066,"audio_seconds_per_second":3.469182019050761,"batches":[{"batch_size":1,"device_seconds":1.1991299324035645,"audio_seconds":[4.16],"model_wall_seconds":1.3638299107551575,"first_audio_seconds":0.478538818359375,"graph_capture_seconds":0,"live_callback_seconds":0.16018366813659668}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1688.76,"audio_s":4.03}],"total_ms":2413.13,"audio_s":4.03,"sentences":1,"generation_rtf":0.5988},"at":10928.099998474121},{"type":"turn_done","barged":false,"at":14493.899997711182}],"reply":"Oh, da läuft gerade ein ziemlich treibender elektronischer Track.","captions":[{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.44,"tokens":35,"preprocess_ms":10.75,"encoder_gpu_ms":8.08,"decoder_gpu_ms":274.54,"inference_ms":282.82,"total_ms":293.57,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":38.72,"window_s":3.44,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1342},{"label":"Drum machine","score":0.1279},{"label":"Synthesizer","score":0.1235},{"label":"Keyboard (musical)","score":0.1138},{"label":"Electric piano","score":0.1082}],"music_score":0.9159,"speech_score":0.0116,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8257},{"label":"Ringtone","score":0.0555},{"label":"Musical instrument","score":0.0549}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1042},{"label":"Synthesizer","score":0.0667}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Music","score":0.8667},{"label":"Musical instrument","score":0.1296},{"label":"Drum machine","score":0.1279}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Music","score":0.8513},{"label":"Musical instrument","score":0.1342},{"label":"Synthesizer","score":0.1235}]},{"start_s":2.16,"end_s":3.44,"candidates":[{"label":"Music","score":0.8948},{"label":"Musical instrument","score":0.1224},{"label":"Synthesizer","score":0.0759}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":46.98},"type":"sound_caption","request_ms":310.26,"capture_ms":3240,"event_age_ms":3550.26,"kind":"music","eligible":true,"suppression_reason":null,"at":8493.29999923706}],"pcm_chunks":5,"page_errors":[],"passed":true},{"case":"music_already_playing","events":[{"type":"interaction","enabled":true,"helpful_enabled":false,"rules":[],"decision_model":"convaiinnovations/laya-multilingual (Stop) · dialogue model (context)","decision_interval_ms":320,"checks":0,"last_decision":null,"waiting_for":"No agreed rules yet","attempts":0,"semantic_error":null,"scope":"session-only conversational rules; acoustic verification is separate","at":4878.799999237061},{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.44,"tokens":35,"preprocess_ms":10.56,"encoder_gpu_ms":16.92,"decoder_gpu_ms":413.32,"inference_ms":430.45,"total_ms":441.01,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":32.4,"window_s":3.44,"candidates":[{"label":"Music","score":0.881},{"label":"Synthesizer","score":0.1634},{"label":"Musical instrument","score":0.1571},{"label":"Electric piano","score":0.1481},{"label":"Keyboard (musical)","score":0.1456},{"label":"Drum machine","score":0.1308}],"music_score":0.881,"speech_score":0.0153,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8649},{"label":"Synthesizer","score":0.1634},{"label":"Musical instrument","score":0.1571}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.8796},{"label":"Musical instrument","score":0.1454},{"label":"Synthesizer","score":0.1305}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Music","score":0.7893},{"label":"Ringtone","score":0.113},{"label":"Musical instrument","score":0.0619}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Music","score":0.881},{"label":"Musical instrument","score":0.1449},{"label":"Electric piano","score":0.128}]},{"start_s":2.16,"end_s":3.44,"candidates":[{"label":"Music","score":0.8515},{"label":"Electric piano","score":0.1238},{"label":"Keyboard (musical)","score":0.1092}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":40.71},"type":"sound_caption","request_ms":457.78,"capture_ms":3240,"event_age_ms":3697.79,"kind":"music","eligible":true,"suppression_reason":null,"at":9995.70000076294},{"type":"environment_observation","kind":"music","status":"responding","at":10059.79999923706},{"type":"latency_start","reply_id":1,"input":{"kind":"environment","sound_caption_ms":457.78},"at":10059.900001525879},{"type":"reply","text":"Oh, das klingt nach fröhlicher Musik!","reply_id":1,"performance":{"sentences":[{"emotion":"contentment","secondary":null,"blend":0,"intensity":0.55,"direction":"curiosity, soft","text":"Oh, das klingt nach fröhlicher Musik!","start_s":0,"end_s":2.04}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Oh, das klingt nach fröhlicher Musik!","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":707.33,"finish_reason":"stop","usage":{"prompt_tokens":2369,"completion_tokens":16,"total_tokens":2385,"cost":0.0023454585,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00236915,"upstream_inference_prompt_cost":0.00234531,"upstream_inference_completions_cost":2.384e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":732.74,"generation_calls":1,"llm_total_ms":732.76,"response_complete":true,"sentence_ready_ms":733.75,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.soft.soft"},"sources":[],"lookups":[],"at":10795.400001525879},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":707.33,"finish_reason":"stop","usage":{"prompt_tokens":2369,"completion_tokens":16,"total_tokens":2385,"cost":0.0023454585,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00236915,"upstream_inference_prompt_cost":0.00234531,"upstream_inference_completions_cost":2.384e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":732.74,"generation_calls":1,"llm_total_ms":732.76,"response_complete":true,"answer_first_audio_ms":1499.69,"tts_first_audio_ms":765.84,"first_audio_ms":1499.7},"at":11560.400001525879},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":707.33,"finish_reason":"stop","usage":{"prompt_tokens":2369,"completion_tokens":16,"total_tokens":2385,"cost":0.0023454585,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00236915,"upstream_inference_prompt_cost":0.00234531,"upstream_inference_completions_cost":2.384e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":732.74,"generation_calls":1,"llm_total_ms":732.76,"response_complete":true,"answer_first_audio_ms":1499.69,"tts_first_audio_ms":765.84,"first_audio_ms":1499.7,"tts_server_stages":[{"generation_ms":1092,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1253,"performance":{"device_seconds":0.8731719608306886,"raw_audio_seconds":3.04,"aggregate_rtf":0.28722761869430546,"audio_seconds_per_second":3.4815593449747384,"batches":[{"batch_size":1,"device_seconds":0.8731719608306886,"audio_seconds":[3.04],"model_wall_seconds":0.9711175337433815,"first_audio_seconds":0.46512939453125,"graph_capture_seconds":0,"live_callback_seconds":0.09767786413431168}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1266.26,"audio_s":2.51}],"total_ms":2000.13,"audio_s":2.51,"sentences":1,"generation_rtf":0.7969},"at":12078.600002288818},{"type":"turn_done","barged":false,"at":14507.70000076294},{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a simple, repetitive drum pattern.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.2,"tokens":37,"preprocess_ms":8.2,"encoder_gpu_ms":8.56,"decoder_gpu_ms":371.11,"inference_ms":379.87,"total_ms":388.07,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":30.79,"window_s":3.2,"candidates":[{"label":"Music","score":0.9043},{"label":"Electric piano","score":0.342},{"label":"Keyboard (musical)","score":0.2976},{"label":"Piano","score":0.25},{"label":"Musical instrument","score":0.1913},{"label":"Synthesizer","score":0.1498}],"music_score":0.9043,"speech_score":0.0129,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.875},{"label":"Electric piano","score":0.342},{"label":"Keyboard (musical)","score":0.2976}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.8821},{"label":"Electric piano","score":0.188},{"label":"Keyboard (musical)","score":0.18}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Music","score":0.7679},{"label":"Ringtone","score":0.1093},{"label":"Musical instrument","score":0.0775}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Music","score":0.9043},{"label":"Musical instrument","score":0.1529},{"label":"Electric piano","score":0.1083}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":41.52},"type":"sound_caption","request_ms":406.85,"capture_ms":2960,"event_age_ms":3366.85,"kind":"music","eligible":false,"suppression_reason":"music_cooldown","at":21738}],"reply":"Oh, das klingt nach fröhlicher Musik!","captions":[{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.44,"tokens":35,"preprocess_ms":10.56,"encoder_gpu_ms":16.92,"decoder_gpu_ms":413.32,"inference_ms":430.45,"total_ms":441.01,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":32.4,"window_s":3.44,"candidates":[{"label":"Music","score":0.881},{"label":"Synthesizer","score":0.1634},{"label":"Musical instrument","score":0.1571},{"label":"Electric piano","score":0.1481},{"label":"Keyboard (musical)","score":0.1456},{"label":"Drum machine","score":0.1308}],"music_score":0.881,"speech_score":0.0153,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8649},{"label":"Synthesizer","score":0.1634},{"label":"Musical instrument","score":0.1571}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.8796},{"label":"Musical instrument","score":0.1454},{"label":"Synthesizer","score":0.1305}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Music","score":0.7893},{"label":"Ringtone","score":0.113},{"label":"Musical instrument","score":0.0619}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Music","score":0.881},{"label":"Musical instrument","score":0.1449},{"label":"Electric piano","score":0.128}]},{"start_s":2.16,"end_s":3.44,"candidates":[{"label":"Music","score":0.8515},{"label":"Electric piano","score":0.1238},{"label":"Keyboard (musical)","score":0.1092}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":40.71},"type":"sound_caption","request_ms":457.78,"capture_ms":3240,"event_age_ms":3697.79,"kind":"music","eligible":true,"suppression_reason":null,"at":9995.70000076294},{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a simple, repetitive drum pattern.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.2,"tokens":37,"preprocess_ms":8.2,"encoder_gpu_ms":8.56,"decoder_gpu_ms":371.11,"inference_ms":379.87,"total_ms":388.07,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":30.79,"window_s":3.2,"candidates":[{"label":"Music","score":0.9043},{"label":"Electric piano","score":0.342},{"label":"Keyboard (musical)","score":0.2976},{"label":"Piano","score":0.25},{"label":"Musical instrument","score":0.1913},{"label":"Synthesizer","score":0.1498}],"music_score":0.9043,"speech_score":0.0129,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.875},{"label":"Electric piano","score":0.342},{"label":"Keyboard (musical)","score":0.2976}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.8821},{"label":"Electric piano","score":0.188},{"label":"Keyboard (musical)","score":0.18}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Music","score":0.7679},{"label":"Ringtone","score":0.1093},{"label":"Musical instrument","score":0.0775}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Music","score":0.9043},{"label":"Musical instrument","score":0.1529},{"label":"Electric piano","score":0.1083}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":41.52},"type":"sound_caption","request_ms":406.85,"capture_ms":2960,"event_age_ms":3366.85,"kind":"music","eligible":false,"suppression_reason":"music_cooldown","at":21738}],"pcm_chunks":4,"page_errors":[],"passed":true},{"case":"dog","events":[{"type":"interaction","enabled":true,"helpful_enabled":false,"rules":[],"decision_model":"convaiinnovations/laya-multilingual (Stop) · dialogue model (context)","decision_interval_ms":320,"checks":0,"last_decision":null,"waiting_for":"No agreed rules yet","attempts":0,"semantic_error":null,"scope":"session-only conversational rules; acoustic verification is separate","at":2079},{"status":"available","caption":"The audio contains a single, distinct bark. The bark is relatively short in duration and has a medium pitch.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.84,"tokens":23,"preprocess_ms":7.97,"encoder_gpu_ms":9.28,"decoder_gpu_ms":187.25,"inference_ms":196.77,"total_ms":204.75,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":29.86,"window_s":1.84,"candidates":[{"label":"Dog","score":0.7442},{"label":"Animal","score":0.7356},{"label":"Bark","score":0.7163},{"label":"Domestic animals, pets","score":0.6316},{"label":"Silence","score":0.4257},{"label":"Bow-wow","score":0.2777}],"music_score":0.0931,"speech_score":0.0268,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Dog","score":0.7442},{"label":"Animal","score":0.7317},{"label":"Bark","score":0.7027}]},{"start_s":0.56,"end_s":1.84,"candidates":[{"label":"Silence","score":0.4257},{"label":"Music","score":0.0931},{"label":"Sound effect","score":0.0186}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":39.11},"type":"sound_caption","request_ms":222.91,"capture_ms":1640,"event_age_ms":1862.92,"kind":"sound","eligible":true,"suppression_reason":null,"at":8264.79999923706},{"type":"environment_observation","kind":"sound","status":"responding","at":8360.199996948242},{"type":"latency_start","reply_id":1,"input":{"kind":"environment","sound_caption_ms":222.91},"at":8360.399997711182},{"type":"reply","text":"Habt ihr gerade ein Hundegefühl gehört?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Habt ihr gerade ein Hundegefühl gehört?","start_s":0,"end_s":2.04}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Habt ihr gerade ein Hundegefühl gehört?","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":491.08,"finish_reason":"stop","usage":{"prompt_tokens":2098,"completion_tokens":15,"total_tokens":2113,"cost":0.0020783763,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00209937,"upstream_inference_prompt_cost":0.00207702,"upstream_inference_completions_cost":2.235e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":529.6,"generation_calls":1,"llm_total_ms":529.63,"response_complete":true,"sentence_ready_ms":530.56,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":8893.099998474121},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":491.08,"finish_reason":"stop","usage":{"prompt_tokens":2098,"completion_tokens":15,"total_tokens":2113,"cost":0.0020783763,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00209937,"upstream_inference_prompt_cost":0.00207702,"upstream_inference_completions_cost":2.235e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":529.6,"generation_calls":1,"llm_total_ms":529.63,"response_complete":true,"answer_first_audio_ms":1287.77,"tts_first_audio_ms":757.11,"first_audio_ms":1287.78},"at":9656.199996948242},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":491.08,"finish_reason":"stop","usage":{"prompt_tokens":2098,"completion_tokens":15,"total_tokens":2113,"cost":0.0020783763,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00209937,"upstream_inference_prompt_cost":0.00207702,"upstream_inference_completions_cost":2.235e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":529.6,"generation_calls":1,"llm_total_ms":529.63,"response_complete":true,"answer_first_audio_ms":1287.77,"tts_first_audio_ms":757.11,"first_audio_ms":1287.78,"tts_server_stages":[{"generation_ms":1030,"alignment_ms":41,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1193,"performance":{"device_seconds":0.8170067462921142,"raw_audio_seconds":2.8,"aggregate_rtf":0.2917881236757551,"audio_seconds_per_second":3.4271442833335946,"batches":[{"batch_size":1,"device_seconds":0.8170067462921142,"audio_seconds":[2.8],"model_wall_seconds":0.915948748588562,"first_audio_seconds":0.46208615112304685,"graph_capture_seconds":0,"live_callback_seconds":0.09669449180364609}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1206.12,"audio_s":2.35}],"total_ms":1736.8,"audio_s":2.35,"sentences":1,"generation_rtf":0.7391},"at":10155.399997711182},{"type":"turn_done","barged":false,"at":12442.79999923706}],"reply":"Habt ihr gerade ein Hundegefühl gehört?","captions":[{"status":"available","caption":"The audio contains a single, distinct bark. The bark is relatively short in duration and has a medium pitch.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.84,"tokens":23,"preprocess_ms":7.97,"encoder_gpu_ms":9.28,"decoder_gpu_ms":187.25,"inference_ms":196.77,"total_ms":204.75,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":29.86,"window_s":1.84,"candidates":[{"label":"Dog","score":0.7442},{"label":"Animal","score":0.7356},{"label":"Bark","score":0.7163},{"label":"Domestic animals, pets","score":0.6316},{"label":"Silence","score":0.4257},{"label":"Bow-wow","score":0.2777}],"music_score":0.0931,"speech_score":0.0268,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Dog","score":0.7442},{"label":"Animal","score":0.7317},{"label":"Bark","score":0.7027}]},{"start_s":0.56,"end_s":1.84,"candidates":[{"label":"Silence","score":0.4257},{"label":"Music","score":0.0931},{"label":"Sound effect","score":0.0186}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":39.11},"type":"sound_caption","request_ms":222.91,"capture_ms":1640,"event_age_ms":1862.92,"kind":"sound","eligible":true,"suppression_reason":null,"at":8264.79999923706}],"pcm_chunks":4,"page_errors":[],"passed":true},{"case":"car_horn","events":[{"type":"interaction","enabled":true,"helpful_enabled":false,"rules":[],"decision_model":"convaiinnovations/laya-multilingual (Stop) · dialogue model (context)","decision_interval_ms":320,"checks":0,"last_decision":null,"waiting_for":"No agreed rules yet","attempts":0,"semantic_error":null,"scope":"session-only conversational rules; acoustic verification is separate","at":1965.900001525879},{"status":"available","caption":"The audio contains a single, distinct, and loud honking sound. The sound is short in duration and has a clear, resonant quality.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.84,"tokens":30,"preprocess_ms":8.08,"encoder_gpu_ms":9.49,"decoder_gpu_ms":241.31,"inference_ms":251.06,"total_ms":259.13,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":23.89,"window_s":1.84,"candidates":[{"label":"Vehicle","score":0.2104},{"label":"Vehicle horn, car horn, honking","score":0.1379},{"label":"Car","score":0.0793},{"label":"Bouncing","score":0.0575},{"label":"Music","score":0.0387},{"label":"Outside, urban or manmade","score":0.0371}],"music_score":0.0387,"speech_score":0.0207,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Vehicle","score":0.061},{"label":"Bouncing","score":0.0575},{"label":"Silence","score":0.028}]},{"start_s":0.56,"end_s":1.84,"candidates":[{"label":"Vehicle","score":0.1837},{"label":"Car","score":0.0564},{"label":"Music","score":0.0387}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":34.43},"type":"sound_caption","request_ms":278.07,"capture_ms":1640,"event_age_ms":1918.08,"kind":"sound","eligible":true,"suppression_reason":null,"at":5958.100002288818},{"type":"environment_observation","kind":"sound","status":"responding","at":6025.600002288818},{"type":"latency_start","reply_id":1,"input":{"kind":"environment","sound_caption_ms":278.07},"at":6025.700000762939},{"type":"reply","text":"Hörst du das?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Hörst du das?","start_s":0,"end_s":1.02}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Hörst du das?","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":346.72,"llm_first_sentence_ms":383.1,"sentence_ready_ms":384.12,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":6412.799999237061},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":346.72,"llm_first_sentence_ms":383.1,"finish_reason":"stop","usage":{"prompt_tokens":2108,"completion_tokens":26,"total_tokens":2134,"cost":0.0021044034,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00212566,"upstream_inference_prompt_cost":0.00208692,"upstream_inference_completions_cost":3.874e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":388.15,"response_complete":true,"answer_first_audio_ms":1126.31,"tts_first_audio_ms":742.1,"first_audio_ms":1126.32},"at":7152.900001525879},{"type":"reply","text":"Hörst du das? Ein kurzes Hupen – war das vielleicht gerade ein Auto draußen?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Hörst du das?","start_s":0,"end_s":1.95},{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Ein kurzes Hupen – war das vielleicht gerade ein Auto draußen?","start_s":1.95,"end_s":5.69}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Ein kurzes Hupen – war das vielleicht gerade ein Auto draußen?","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":346.72,"llm_first_sentence_ms":383.1,"finish_reason":"stop","usage":{"prompt_tokens":2108,"completion_tokens":26,"total_tokens":2134,"cost":0.0021044034,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00212566,"upstream_inference_prompt_cost":0.00208692,"upstream_inference_completions_cost":3.874e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":388.15,"response_complete":true,"answer_first_audio_ms":1126.31,"tts_first_audio_ms":742.1,"first_audio_ms":1126.32,"tts_server_stages":[{"generation_ms":1014,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1175,"performance":{"device_seconds":0.8132195658683776,"raw_audio_seconds":2.8,"aggregate_rtf":0.2904355592387063,"audio_seconds_per_second":3.4431045655057315,"batches":[{"batch_size":1,"device_seconds":0.8132195658683776,"audio_seconds":[2.8],"model_wall_seconds":0.900035448372364,"first_audio_seconds":0.455041015625,"graph_capture_seconds":0,"live_callback_seconds":0.0865948423743248}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1184.27,"audio_s":1.95}],"sentence_ready_ms":1569.35,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":7638},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":346.72,"llm_first_sentence_ms":383.1,"finish_reason":"stop","usage":{"prompt_tokens":2108,"completion_tokens":26,"total_tokens":2134,"cost":0.0021044034,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00212566,"upstream_inference_prompt_cost":0.00208692,"upstream_inference_completions_cost":3.874e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":388.15,"response_complete":true,"answer_first_audio_ms":1126.31,"tts_first_audio_ms":742.1,"first_audio_ms":1126.32,"tts_server_stages":[{"generation_ms":1014,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1175,"performance":{"device_seconds":0.8132195658683776,"raw_audio_seconds":2.8,"aggregate_rtf":0.2904355592387063,"audio_seconds_per_second":3.4431045655057315,"batches":[{"batch_size":1,"device_seconds":0.8132195658683776,"audio_seconds":[2.8],"model_wall_seconds":0.900035448372364,"first_audio_seconds":0.455041015625,"graph_capture_seconds":0,"live_callback_seconds":0.0865948423743248}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1184.27,"audio_s":1.95}]},"at":8367.20000076294},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":346.72,"llm_first_sentence_ms":383.1,"finish_reason":"stop","usage":{"prompt_tokens":2108,"completion_tokens":26,"total_tokens":2134,"cost":0.0021044034,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00212566,"upstream_inference_prompt_cost":0.00208692,"upstream_inference_completions_cost":3.874e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":388.15,"response_complete":true,"answer_first_audio_ms":1126.31,"tts_first_audio_ms":742.1,"first_audio_ms":1126.32,"tts_server_stages":[{"generation_ms":1014,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1175,"performance":{"device_seconds":0.8132195658683776,"raw_audio_seconds":2.8,"aggregate_rtf":0.2904355592387063,"audio_seconds_per_second":3.4431045655057315,"batches":[{"batch_size":1,"device_seconds":0.8132195658683776,"audio_seconds":[2.8],"model_wall_seconds":0.900035448372364,"first_audio_seconds":0.455041015625,"graph_capture_seconds":0,"live_callback_seconds":0.0865948423743248}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1539,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1698,"performance":{"device_seconds":1.2696923303604126,"raw_audio_seconds":4.56,"aggregate_rtf":0.2784413005176344,"audio_seconds_per_second":3.591421237226507,"batches":[{"batch_size":1,"device_seconds":1.2696923303604126,"audio_seconds":[4.56],"model_wall_seconds":1.4270158782601357,"first_audio_seconds":0.46315826416015626,"graph_capture_seconds":0,"live_callback_seconds":0.15279464423656464}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1184.27,"audio_s":1.95},{"sentence":1,"total_ms":1709.98,"audio_s":4.56}],"total_ms":3279.47,"audio_s":6.51,"sentences":2,"generation_rtf":0.5038},"at":9317.900001525879},{"type":"turn_done","barged":false,"at":14093.70000076294}],"reply":"Hörst du das? Ein kurzes Hupen – war das vielleicht gerade ein Auto draußen?","captions":[{"status":"available","caption":"The audio contains a single, distinct, and loud honking sound. The sound is short in duration and has a clear, resonant quality.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.84,"tokens":30,"preprocess_ms":8.08,"encoder_gpu_ms":9.49,"decoder_gpu_ms":241.31,"inference_ms":251.06,"total_ms":259.13,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":23.89,"window_s":1.84,"candidates":[{"label":"Vehicle","score":0.2104},{"label":"Vehicle horn, car horn, honking","score":0.1379},{"label":"Car","score":0.0793},{"label":"Bouncing","score":0.0575},{"label":"Music","score":0.0387},{"label":"Outside, urban or manmade","score":0.0371}],"music_score":0.0387,"speech_score":0.0207,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Vehicle","score":0.061},{"label":"Bouncing","score":0.0575},{"label":"Silence","score":0.028}]},{"start_s":0.56,"end_s":1.84,"candidates":[{"label":"Vehicle","score":0.1837},{"label":"Car","score":0.0564},{"label":"Music","score":0.0387}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":34.43},"type":"sound_caption","request_ms":278.07,"capture_ms":1640,"event_age_ms":1918.08,"kind":"sound","eligible":true,"suppression_reason":null,"at":5958.100002288818}],"pcm_chunks":9,"page_errors":[],"passed":true},{"case":"glass_breaking","events":[{"type":"interaction","enabled":true,"helpful_enabled":false,"rules":[],"decision_model":"convaiinnovations/laya-multilingual (Stop) · dialogue model (context)","decision_interval_ms":320,"checks":0,"last_decision":null,"waiting_for":"No agreed rules yet","attempts":0,"semantic_error":null,"scope":"session-only conversational rules; acoustic verification is separate","at":1868.8000030517578},{"status":"available","caption":"The audio contains a single, sharp, percussive sound. It sounds like a single, forceful impact, possibly a drum hit or a similar sound.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.86,"tokens":33,"preprocess_ms":13.53,"encoder_gpu_ms":11.59,"decoder_gpu_ms":252.74,"inference_ms":264.59,"total_ms":278.12,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":25.43,"window_s":1.86,"candidates":[{"label":"Breaking","score":0.296},{"label":"Shatter","score":0.174},{"label":"Glass","score":0.0906},{"label":"Sound effect","score":0.0666},{"label":"Silence","score":0.0605},{"label":"Chink, clink","score":0.0504}],"music_score":0.0468,"speech_score":0.0347,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Silence","score":0.0605},{"label":"Vehicle","score":0.0456},{"label":"Music","score":0.0382}]},{"start_s":0.58,"end_s":1.86,"candidates":[{"label":"Sound effect","score":0.0666},{"label":"Music","score":0.0468},{"label":"Chink, clink","score":0.0393}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.98},"type":"sound_caption","request_ms":298.77,"capture_ms":1660,"event_age_ms":1958.77,"kind":"sound","eligible":true,"suppression_reason":null,"at":6013.400001525879},{"type":"environment_observation","kind":"sound","status":"responding","at":6036.5},{"type":"latency_start","reply_id":1,"input":{"kind":"environment","sound_caption_ms":298.77},"at":6036.600002288818},{"type":"reply","text":"Hi, ich habe gerade kurz ein klares Klanggeräusch gehört, war das etwa ein Drum-Hit?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Hi, ich habe gerade kurz ein klares Klanggeräusch gehört, war das etwa ein Drum-Hit?","start_s":0,"end_s":4.76}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Hi, ich habe gerade kurz ein klares Klanggeräusch gehört, war das etwa ein Drum-Hit?","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1320.21,"finish_reason":"stop","usage":{"prompt_tokens":2109,"completion_tokens":29,"total_tokens":2138,"cost":0.0021098088,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00213112,"upstream_inference_prompt_cost":0.00208791,"upstream_inference_completions_cost":4.321e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":1363.76,"generation_calls":1,"llm_total_ms":1363.79,"response_complete":true,"sentence_ready_ms":1364.83,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":7403},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1320.21,"finish_reason":"stop","usage":{"prompt_tokens":2109,"completion_tokens":29,"total_tokens":2138,"cost":0.0021098088,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00213112,"upstream_inference_prompt_cost":0.00208791,"upstream_inference_completions_cost":4.321e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":1363.76,"generation_calls":1,"llm_total_ms":1363.79,"response_complete":true,"answer_first_audio_ms":2116.55,"tts_first_audio_ms":751.62,"first_audio_ms":2116.56},"at":8154.400001525879},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1320.21,"finish_reason":"stop","usage":{"prompt_tokens":2109,"completion_tokens":29,"total_tokens":2138,"cost":0.0021098088,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00213112,"upstream_inference_prompt_cost":0.00208791,"upstream_inference_completions_cost":4.321e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":1363.76,"generation_calls":1,"llm_total_ms":1363.79,"response_complete":true,"answer_first_audio_ms":2116.55,"tts_first_audio_ms":751.62,"first_audio_ms":2116.56,"tts_server_stages":[{"generation_ms":2083,"alignment_ms":42,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":2240,"performance":{"device_seconds":1.7334446887969972,"raw_audio_seconds":6.24,"aggregate_rtf":0.27779562320464696,"audio_seconds_per_second":3.599768738124856,"batches":[{"batch_size":1,"device_seconds":1.7334446887969972,"audio_seconds":[6.24],"model_wall_seconds":1.9672154486179352,"first_audio_seconds":0.4715386962890625,"graph_capture_seconds":0,"live_callback_seconds":0.23341821134090424}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":2258.05,"audio_s":6.24}],"total_ms":3622.99,"audio_s":6.24,"sentences":1,"generation_rtf":0.5806},"at":9667.70000076294},{"type":"turn_done","barged":false,"at":14840}],"reply":"Hi, ich habe gerade kurz ein klares Klanggeräusch gehört, war das etwa ein Drum-Hit?","captions":[{"status":"available","caption":"The audio contains a single, sharp, percussive sound. It sounds like a single, forceful impact, possibly a drum hit or a similar sound.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.86,"tokens":33,"preprocess_ms":13.53,"encoder_gpu_ms":11.59,"decoder_gpu_ms":252.74,"inference_ms":264.59,"total_ms":278.12,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":25.43,"window_s":1.86,"candidates":[{"label":"Breaking","score":0.296},{"label":"Shatter","score":0.174},{"label":"Glass","score":0.0906},{"label":"Sound effect","score":0.0666},{"label":"Silence","score":0.0605},{"label":"Chink, clink","score":0.0504}],"music_score":0.0468,"speech_score":0.0347,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Silence","score":0.0605},{"label":"Vehicle","score":0.0456},{"label":"Music","score":0.0382}]},{"start_s":0.58,"end_s":1.86,"candidates":[{"label":"Sound effect","score":0.0666},{"label":"Music","score":0.0468},{"label":"Chink, clink","score":0.0393}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.98},"type":"sound_caption","request_ms":298.77,"capture_ms":1660,"event_age_ms":1958.77,"kind":"sound","eligible":true,"suppression_reason":null,"at":6013.400001525879}],"pcm_chunks":6,"page_errors":[],"passed":true},{"case":"cat","events":[{"type":"interaction","enabled":true,"helpful_enabled":false,"rules":[],"decision_model":"convaiinnovations/laya-multilingual (Stop) · dialogue model (context)","decision_interval_ms":320,"checks":0,"last_decision":null,"waiting_for":"No agreed rules yet","attempts":0,"semantic_error":null,"scope":"session-only conversational rules; acoustic verification is separate","at":1861.5},{"status":"available","caption":"The audio contains a single, clear meow. The meow is relatively high-pitched and has a slightly nasal quality.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.44,"tokens":25,"preprocess_ms":13.04,"encoder_gpu_ms":8.14,"decoder_gpu_ms":228.51,"inference_ms":236.91,"total_ms":249.95,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":38.62,"window_s":3.44,"candidates":[{"label":"Animal","score":0.6495},{"label":"Cat","score":0.5755},{"label":"Meow","score":0.4905},{"label":"Domestic animals, pets","score":0.4871},{"label":"Siren","score":0.4751},{"label":"Emergency vehicle","score":0.249}],"music_score":0.2487,"speech_score":0.0808,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Siren","score":0.4751},{"label":"Emergency vehicle","score":0.249},{"label":"Vehicle","score":0.1209}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Siren","score":0.33},{"label":"Emergency vehicle","score":0.2275},{"label":"Domestic animals, pets","score":0.1647}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Animal","score":0.2332},{"label":"Domestic animals, pets","score":0.0807},{"label":"Cat","score":0.0733}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Animal","score":0.5335},{"label":"Cat","score":0.5001},{"label":"Meow","score":0.4905}]},{"start_s":2.16,"end_s":3.44,"candidates":[{"label":"Cat","score":0.4657},{"label":"Animal","score":0.3949},{"label":"Domestic animals, pets","score":0.3506}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":49.28},"type":"sound_caption","request_ms":266.23,"capture_ms":3240,"event_age_ms":3506.24,"kind":"sound","eligible":true,"suppression_reason":null,"at":7445.200000762939},{"type":"environment_observation","kind":"sound","status":"responding","at":7534.800003051758},{"type":"latency_start","reply_id":1,"input":{"kind":"environment","sound_caption_ms":266.23},"at":7534.900001525879},{"type":"reply","text":"Hallo?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Hallo?","start_s":0,"end_s":0.4}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Hallo?","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":421.38,"llm_first_sentence_ms":447.1,"sentence_ready_ms":447.74,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":7983.900001525879},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":421.38,"llm_first_sentence_ms":447.1,"finish_reason":"stop","usage":{"prompt_tokens":2336,"completion_tokens":27,"total_tokens":2363,"cost":0.0023293413,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00235287,"upstream_inference_prompt_cost":0.00231264,"upstream_inference_completions_cost":4.023e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":500.9,"response_complete":true,"answer_first_audio_ms":1180.09,"tts_first_audio_ms":732.29,"first_audio_ms":1180.1},"at":8718.5},{"type":"reply","text":"Hallo? Ich habe gerade ein leises Meow gehört – ist da etwa eine Katze im Raum?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Hallo?","start_s":0,"end_s":2.43},{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Ich habe gerade ein leises Meow gehört – ist da etwa eine Katze im Raum?","start_s":2.43,"end_s":7.529999999999999}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Ich habe gerade ein leises Meow gehört – ist da etwa eine Katze im Raum?","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":421.38,"llm_first_sentence_ms":447.1,"finish_reason":"stop","usage":{"prompt_tokens":2336,"completion_tokens":27,"total_tokens":2363,"cost":0.0023293413,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00235287,"upstream_inference_prompt_cost":0.00231264,"upstream_inference_completions_cost":4.023e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":500.9,"response_complete":true,"answer_first_audio_ms":1180.09,"tts_first_audio_ms":732.29,"first_audio_ms":1180.1,"tts_server_stages":[{"generation_ms":1236,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1395,"performance":{"device_seconds":0.9930395169258117,"raw_audio_seconds":3.28,"aggregate_rtf":0.3027559502822597,"audio_seconds_per_second":3.3029904088349014,"batches":[{"batch_size":1,"device_seconds":0.9930395169258117,"audio_seconds":[3.28],"model_wall_seconds":1.1209493279457092,"first_audio_seconds":0.44705484008789065,"graph_capture_seconds":0,"live_callback_seconds":0.1276325359940529}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1402.8,"audio_s":2.43}],"sentence_ready_ms":1851.08,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":9398.20000076294},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":421.38,"llm_first_sentence_ms":447.1,"finish_reason":"stop","usage":{"prompt_tokens":2336,"completion_tokens":27,"total_tokens":2363,"cost":0.0023293413,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00235287,"upstream_inference_prompt_cost":0.00231264,"upstream_inference_completions_cost":4.023e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":500.9,"response_complete":true,"answer_first_audio_ms":1180.09,"tts_first_audio_ms":732.29,"first_audio_ms":1180.1,"tts_server_stages":[{"generation_ms":1236,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1395,"performance":{"device_seconds":0.9930395169258117,"raw_audio_seconds":3.28,"aggregate_rtf":0.3027559502822597,"audio_seconds_per_second":3.3029904088349014,"batches":[{"batch_size":1,"device_seconds":0.9930395169258117,"audio_seconds":[3.28],"model_wall_seconds":1.1209493279457092,"first_audio_seconds":0.44705484008789065,"graph_capture_seconds":0,"live_callback_seconds":0.1276325359940529}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1402.8,"audio_s":2.43}]},"at":10131.20000076294},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":421.38,"llm_first_sentence_ms":447.1,"finish_reason":"stop","usage":{"prompt_tokens":2336,"completion_tokens":27,"total_tokens":2363,"cost":0.0023293413,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00235287,"upstream_inference_prompt_cost":0.00231264,"upstream_inference_completions_cost":4.023e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":500.9,"response_complete":true,"answer_first_audio_ms":1180.09,"tts_first_audio_ms":732.29,"first_audio_ms":1180.1,"tts_server_stages":[{"generation_ms":1236,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1395,"performance":{"device_seconds":0.9930395169258117,"raw_audio_seconds":3.28,"aggregate_rtf":0.3027559502822597,"audio_seconds_per_second":3.3029904088349014,"batches":[{"batch_size":1,"device_seconds":0.9930395169258117,"audio_seconds":[3.28],"model_wall_seconds":1.1209493279457092,"first_audio_seconds":0.44705484008789065,"graph_capture_seconds":0,"live_callback_seconds":0.1276325359940529}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1945,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":2104,"performance":{"device_seconds":1.6038379793167112,"raw_audio_seconds":5.76,"aggregate_rtf":0.2784440936313735,"audio_seconds_per_second":3.591385211150788,"batches":[{"batch_size":1,"device_seconds":1.6038379793167112,"audio_seconds":[5.76],"model_wall_seconds":1.8334610313177109,"first_audio_seconds":0.46060028076171877,"graph_capture_seconds":0,"live_callback_seconds":0.2241562157869339}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1402.8,"audio_s":2.43},{"sentence":1,"total_ms":2112.85,"audio_s":5.39}],"total_ms":3964.01,"audio_s":7.82,"sentences":2,"generation_rtf":0.5069},"at":11503.5},{"type":"turn_done","barged":false,"at":16964}],"reply":"Hallo? Ich habe gerade ein leises Meow gehört – ist da etwa eine Katze im Raum?","captions":[{"status":"available","caption":"The audio contains a single, clear meow. The meow is relatively high-pitched and has a slightly nasal quality.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":3.44,"tokens":25,"preprocess_ms":13.04,"encoder_gpu_ms":8.14,"decoder_gpu_ms":228.51,"inference_ms":236.91,"total_ms":249.95,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":38.62,"window_s":3.44,"candidates":[{"label":"Animal","score":0.6495},{"label":"Cat","score":0.5755},{"label":"Meow","score":0.4905},{"label":"Domestic animals, pets","score":0.4871},{"label":"Siren","score":0.4751},{"label":"Emergency vehicle","score":0.249}],"music_score":0.2487,"speech_score":0.0808,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Siren","score":0.4751},{"label":"Emergency vehicle","score":0.249},{"label":"Vehicle","score":0.1209}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Siren","score":0.33},{"label":"Emergency vehicle","score":0.2275},{"label":"Domestic animals, pets","score":0.1647}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Animal","score":0.2332},{"label":"Domestic animals, pets","score":0.0807},{"label":"Cat","score":0.0733}]},{"start_s":1.92,"end_s":3.2,"candidates":[{"label":"Animal","score":0.5335},{"label":"Cat","score":0.5001},{"label":"Meow","score":0.4905}]},{"start_s":2.16,"end_s":3.44,"candidates":[{"label":"Cat","score":0.4657},{"label":"Animal","score":0.3949},{"label":"Domestic animals, pets","score":0.3506}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":49.28},"type":"sound_caption","request_ms":266.23,"capture_ms":3240,"event_age_ms":3506.24,"kind":"sound","eligible":true,"suppression_reason":null,"at":7445.200000762939}],"pcm_chunks":11,"page_errors":[],"passed":true},{"case":"after_question_music_bridge_False","events":[{"type":"environment_observation","kind":"requested_sound","status":"listening_after_question","at":3754.2000007629395},{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.44,"tokens":35,"preprocess_ms":13.04,"encoder_gpu_ms":8.11,"decoder_gpu_ms":261.73,"inference_ms":270.04,"total_ms":283.08,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":32.88,"window_s":2.44,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1602},{"label":"Drum machine","score":0.1019},{"label":"Synthesizer","score":0.0908},{"label":"Keyboard (musical)","score":0.0646},{"label":"Electric piano","score":0.0633}],"music_score":0.9159,"speech_score":0.0139,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8257},{"label":"Ringtone","score":0.0555},{"label":"Musical instrument","score":0.0549}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1042},{"label":"Synthesizer","score":0.0667}]},{"start_s":1.16,"end_s":2.44,"candidates":[{"label":"Music","score":0.8965},{"label":"Musical instrument","score":0.1602},{"label":"Synthesizer","score":0.0908}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":43.84},"type":"sound_caption","request_ms":302.19,"capture_ms":2440,"event_age_ms":2742.19,"kind":"music","eligible":true,"suppression_reason":null,"requested":true,"at":6544.5},{"type":"latency_start","reply_id":1,"input":{"kind":"typed"},"at":6548.099998474121},{"type":"reply","text":"Das klingt nach einem schnellen elektronischen Track mit einer treibenden Bassline und einer sehr repetitiven Synth-Melodie.","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":"contentment","blend":0.25,"intensity":0.72,"direction":"interest, natural","text":"Das klingt nach einem schnellen elektronischen Track mit einer treibenden Bassline und einer sehr repetitiven Synth-Melodie.","start_s":0,"end_s":5.44}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Das klingt nach einem schnellen elektronischen Track mit einer treibenden Bassline und einer sehr repetitiven Synth-Melodie.","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":500.78,"finish_reason":"stop","usage":{"prompt_tokens":2278,"completion_tokens":32,"total_tokens":2310,"cost":0.002279871,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0023029,"upstream_inference_prompt_cost":0.00225522,"upstream_inference_completions_cost":4.768e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":577.53,"generation_calls":1,"llm_total_ms":577.55,"response_complete":true,"sentence_ready_ms":579.28,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":7131.799999237061},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":500.78,"finish_reason":"stop","usage":{"prompt_tokens":2278,"completion_tokens":32,"total_tokens":2310,"cost":0.002279871,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0023029,"upstream_inference_prompt_cost":0.00225522,"upstream_inference_completions_cost":4.768e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":577.53,"generation_calls":1,"llm_total_ms":577.55,"response_complete":true,"answer_first_audio_ms":1367.33,"tts_first_audio_ms":787.94,"first_audio_ms":1367.33},"at":7915.899997711182},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":500.78,"finish_reason":"stop","usage":{"prompt_tokens":2278,"completion_tokens":32,"total_tokens":2310,"cost":0.002279871,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0023029,"upstream_inference_prompt_cost":0.00225522,"upstream_inference_completions_cost":4.768e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":577.53,"generation_calls":1,"llm_total_ms":577.55,"response_complete":true,"answer_first_audio_ms":1367.33,"tts_first_audio_ms":787.94,"first_audio_ms":1367.33,"tts_server_stages":[{"generation_ms":2374,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":2549,"performance":{"device_seconds":1.922487024307251,"raw_audio_seconds":6.64,"aggregate_rtf":0.28953117835952574,"audio_seconds_per_second":3.4538594622726557,"batches":[{"batch_size":1,"device_seconds":1.922487024307251,"audio_seconds":[6.64],"model_wall_seconds":2.258733846247196,"first_audio_seconds":0.4776119079589844,"graph_capture_seconds":0,"live_callback_seconds":0.33365243673324585}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":2566.28,"audio_s":6.64}],"total_ms":3145.68,"audio_s":6.64,"sentences":1,"generation_rtf":0.4737},"at":9699.70000076294},{"type":"turn_done","barged":false,"at":15007.5}],"reply":"Das klingt nach einem schnellen elektronischen Track mit einer treibenden Bassline und einer sehr repetitiven Synth-Melodie.","captions":[{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.44,"tokens":35,"preprocess_ms":13.04,"encoder_gpu_ms":8.11,"decoder_gpu_ms":261.73,"inference_ms":270.04,"total_ms":283.08,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":32.88,"window_s":2.44,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1602},{"label":"Drum machine","score":0.1019},{"label":"Synthesizer","score":0.0908},{"label":"Keyboard (musical)","score":0.0646},{"label":"Electric piano","score":0.0633}],"music_score":0.9159,"speech_score":0.0139,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8257},{"label":"Ringtone","score":0.0555},{"label":"Musical instrument","score":0.0549}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1042},{"label":"Synthesizer","score":0.0667}]},{"start_s":1.16,"end_s":2.44,"candidates":[{"label":"Music","score":0.8965},{"label":"Musical instrument","score":0.1602},{"label":"Synthesizer","score":0.0908}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":43.84},"type":"sound_caption","request_ms":302.19,"capture_ms":2440,"event_age_ms":2742.19,"kind":"music","eligible":true,"suppression_reason":null,"requested":true,"at":6544.5}],"pcm_chunks":7,"page_errors":[],"passed":true},{"case":"after_question_music_bridge_True","events":[{"type":"quick_reply_settings","enabled":true,"at":6214.599998474121},{"type":"environment_observation","kind":"requested_sound","status":"listening_after_question","at":6219.099998474121},{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.4,"tokens":35,"preprocess_ms":9.53,"encoder_gpu_ms":9.09,"decoder_gpu_ms":333.23,"inference_ms":342.55,"total_ms":352.08,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":28.05,"window_s":2.4,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1397},{"label":"Synthesizer","score":0.0709},{"label":"Drum machine","score":0.0705},{"label":"Electric piano","score":0.0644},{"label":"Keyboard (musical)","score":0.0612}],"music_score":0.9159,"speech_score":0.0159,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8257},{"label":"Ringtone","score":0.0555},{"label":"Musical instrument","score":0.0549}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1042},{"label":"Synthesizer","score":0.0667}]},{"start_s":1.12,"end_s":2.4,"candidates":[{"label":"Music","score":0.8883},{"label":"Musical instrument","score":0.1397},{"label":"Synthesizer","score":0.0709}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.42},"type":"sound_caption","request_ms":368.33,"capture_ms":2400,"event_age_ms":2768.34,"kind":"music","eligible":true,"suppression_reason":null,"requested":true,"at":9025.599998474121},{"type":"latency_start","reply_id":1,"input":{"kind":"typed"},"at":9025.899997711182},{"type":"reply","text":"Das klingt nach Musik.","reply_id":1,"performance":{"sentences":[{"emotion":"contentment","secondary":"affection","blend":0.25,"intensity":0.55,"direction":"contentment, soft, warm","text":"Das klingt nach Musik.","start_s":0,"end_s":1.36}],"source":"Contextual reaction + informed Qwen + shared TTS batch","timing":"Sentence starts anchored to preceding PCM; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Das klingt nach Musik.","sentence_role":"reaction","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"selected","tts_backend":"continuous_batch","tts_weight_instances":1,"tts_temperature":0.8,"informed_bridge_status":"replaced_by_informed_main","main_informed_by_reaction":true,"main_launch_ms":382.4,"llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","quick_reply_generation_ms":376.98,"quick_reply_model":"nemotron-bolt","quick_reply_memory_entries":0,"quick_reply_memory_person_matched":false,"quick_reply_words":4,"sentence_ready_ms":385.43,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"contentment.soft"},"sources":[],"lookups":[],"at":9413},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"played","tts_backend":"continuous_batch","tts_weight_instances":1,"tts_temperature":0.8,"informed_bridge_status":"replaced_by_informed_main","main_informed_by_reaction":true,"main_launch_ms":382.4,"llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","quick_reply_generation_ms":376.98,"quick_reply_model":"nemotron-bolt","quick_reply_memory_entries":0,"quick_reply_memory_person_matched":false,"quick_reply_words":4,"first_audio_ms":569.2,"tts_first_audio_ms":184.1},"at":9595.199996948242},{"type":"reply","text":"Das klingt nach Musik.\nEin schneller elektronischer Track mit einer sehr treibenden Basslinie.","reply_id":1,"performance":{"sentences":[{"emotion":"contentment","secondary":"affection","blend":0.25,"intensity":0.55,"direction":"contentment, soft, warm","text":"Das klingt nach Musik.","start_s":0,"end_s":2.72},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"playful, rapid","text":"Ein schneller elektronischer Track mit einer sehr treibenden Basslinie.","start_s":2.72,"end_s":5.78}],"source":"Contextual reaction + informed Qwen + shared TTS batch","timing":"Sentence starts anchored to preceding PCM; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Ein schneller elektronischer Track mit einer sehr treibenden Basslinie.","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"played","tts_backend":"continuous_batch","tts_weight_instances":1,"tts_temperature":0.8,"informed_bridge_status":"replaced_by_informed_main","main_informed_by_reaction":true,"main_launch_ms":382.4,"llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","quick_reply_generation_ms":376.98,"quick_reply_model":"nemotron-bolt","quick_reply_memory_entries":0,"quick_reply_memory_person_matched":false,"quick_reply_words":4,"first_audio_ms":569.2,"tts_first_audio_ms":184.1,"llm_actual_provider":"Cerebras","llm_ttft_ms":388.72,"llm_first_sentence_ms":411.87,"duplicate_first_ack_skipped":true,"finish_reason":"stop","usage":{"prompt_tokens":2360,"completion_tokens":54,"total_tokens":2414,"cost":0.0023926914,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00241686,"upstream_inference_prompt_cost":0.0023364,"upstream_inference_completions_cost":8.046e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":470.39,"response_complete":true,"tts_server_stages":[{"generation_ms":2003.17,"alignment_ms":0.01,"alignment":null,"frames":34,"live_alignment":{"enabled":false,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":0}}],"quick_reply_audio_s":2.72,"quick_reply_tts_ms":2005.44,"sentence_tts":[{"sentence":0,"role":"reaction","audio_s":2.72,"total_ms":2005.44,"prefetched":true}],"sentence_ready_ms":2390.61,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.playful.intense"},"sources":[],"lookups":[],"at":11430.5},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"played","tts_backend":"continuous_batch","tts_weight_instances":1,"tts_temperature":0.8,"informed_bridge_status":"replaced_by_informed_main","main_informed_by_reaction":true,"main_launch_ms":382.4,"llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","quick_reply_generation_ms":376.98,"quick_reply_model":"nemotron-bolt","quick_reply_memory_entries":0,"quick_reply_memory_person_matched":false,"quick_reply_words":4,"first_audio_ms":569.2,"tts_first_audio_ms":184.1,"llm_actual_provider":"Cerebras","llm_ttft_ms":388.72,"llm_first_sentence_ms":411.87,"duplicate_first_ack_skipped":true,"finish_reason":"stop","usage":{"prompt_tokens":2360,"completion_tokens":54,"total_tokens":2414,"cost":0.0023926914,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00241686,"upstream_inference_prompt_cost":0.0023364,"upstream_inference_completions_cost":8.046e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":470.39,"response_complete":true,"tts_server_stages":[{"generation_ms":2003.17,"alignment_ms":0.01,"alignment":null,"frames":34,"live_alignment":{"enabled":false,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":0}}],"quick_reply_audio_s":2.72,"quick_reply_tts_ms":2005.44,"sentence_tts":[{"sentence":0,"role":"reaction","audio_s":2.72,"total_ms":2005.44,"prefetched":true}],"answer_first_audio_ms":2392.69,"answer_audio_ready_ms":1964.1},"at":11431},{"type":"reply","text":"Das klingt nach Musik.\nEin schneller elektronischer Track mit einer sehr treibenden Basslinie.\nKlingt, als würde jemand gerade auf „Party-Modus“ drücken!","reply_id":1,"performance":{"sentences":[{"emotion":"contentment","secondary":"affection","blend":0.25,"intensity":0.55,"direction":"contentment, soft, warm","text":"Das klingt nach Musik.","start_s":0,"end_s":2.72},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"playful, rapid","text":"Ein schneller elektronischer Track mit einer sehr treibenden Basslinie.","start_s":2.72,"end_s":6.51},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.55,"direction":"amusement, soft","text":"Klingt, als würde jemand gerade auf „Party-Modus“ drücken!","start_s":6.51,"end_s":9.23}],"source":"Contextual reaction + informed Qwen + shared TTS batch","timing":"Sentence starts anchored to preceding PCM; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":2,"sentence_text":"Klingt, als würde jemand gerade auf „Party-Modus“ drücken!","sentence_role":"answer","speaker_id":null,"speaker_name":null,"speaker_update":null,"timing":{"quick_reply_status":"played","tts_backend":"continuous_batch","tts_weight_instances":1,"tts_temperature":0.8,"informed_bridge_status":"replaced_by_informed_main","main_informed_by_reaction":true,"main_launch_ms":382.4,"llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","quick_reply_generation_ms":376.98,"quick_reply_model":"nemotron-bolt","quick_reply_memory_entries":0,"quick_reply_memory_person_matched":false,"quick_reply_words":4,"first_audio_ms":569.2,"tts_first_audio_ms":184.1,"llm_actual_provider":"Cerebras","llm_ttft_ms":388.72,"llm_first_sentence_ms":411.87,"duplicate_first_ack_skipped":true,"finish_reason":"stop","usage":{"prompt_tokens":2360,"completion_tokens":54,"total_tokens":2414,"cost":0.0023926914,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00241686,"upstream_inference_prompt_cost":0.0023364,"upstream_inference_completions_cost":8.046e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":470.39,"response_complete":true,"tts_server_stages":[{"generation_ms":2003.17,"alignment_ms":0.01,"alignment":null,"frames":34,"live_alignment":{"enabled":false,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":0}},{"generation_ms":3190.69,"alignment_ms":37.27,"alignment":{"applied":true,"lead_s":0.16,"cut_s":3.64,"removed_s":0.05,"words":9,"note":"","spans":[["Ein",0.16,0.32,1],["schneller",0.32,0.72,1],["elektronischer",0.72,1.36,1],["Track",1.36,1.76,1],["mit",1.76,1.84,1],["einer",1.92,2.16,1],["sehr",2.16,2.4,1],["treibenden",2.4,2.96,1],["Basslinie",2.96,3.52,1]],"enabled":true,"available":true,"raw_seconds":3.84,"seconds":3.79,"lead_ramp_s":0.1,"tail_ramp_s":0.15,"score_kind":"structural plausibility, not model confidence"},"frames":48,"live_alignment":{"enabled":true,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":3.79}}],"quick_reply_audio_s":2.72,"quick_reply_tts_ms":2005.44,"sentence_tts":[{"sentence":0,"role":"reaction","audio_s":2.72,"total_ms":2005.44,"prefetched":true},{"sentence":1,"role":"answer","audio_s":3.79,"total_ms":3201.42,"prefetched":true}],"answer_first_audio_ms":2392.69,"answer_audio_ready_ms":1964.1,"sentence_ready_ms":4019.63,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"amusement.soft"},"sources":[],"lookups":[],"at":13061.199996948242},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"played","tts_backend":"continuous_batch","tts_weight_instances":1,"tts_temperature":0.8,"informed_bridge_status":"replaced_by_informed_main","main_informed_by_reaction":true,"main_launch_ms":382.4,"llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","quick_reply_generation_ms":376.98,"quick_reply_model":"nemotron-bolt","quick_reply_memory_entries":0,"quick_reply_memory_person_matched":false,"quick_reply_words":4,"first_audio_ms":569.2,"tts_first_audio_ms":184.1,"llm_actual_provider":"Cerebras","llm_ttft_ms":388.72,"llm_first_sentence_ms":411.87,"duplicate_first_ack_skipped":true,"finish_reason":"stop","usage":{"prompt_tokens":2360,"completion_tokens":54,"total_tokens":2414,"cost":0.0023926914,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00241686,"upstream_inference_prompt_cost":0.0023364,"upstream_inference_completions_cost":8.046e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":470.39,"response_complete":true,"tts_server_stages":[{"generation_ms":2003.17,"alignment_ms":0.01,"alignment":null,"frames":34,"live_alignment":{"enabled":false,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":0}},{"generation_ms":3190.69,"alignment_ms":37.27,"alignment":{"applied":true,"lead_s":0.16,"cut_s":3.64,"removed_s":0.05,"words":9,"note":"","spans":[["Ein",0.16,0.32,1],["schneller",0.32,0.72,1],["elektronischer",0.72,1.36,1],["Track",1.36,1.76,1],["mit",1.76,1.84,1],["einer",1.92,2.16,1],["sehr",2.16,2.4,1],["treibenden",2.4,2.96,1],["Basslinie",2.96,3.52,1]],"enabled":true,"available":true,"raw_seconds":3.84,"seconds":3.79,"lead_ramp_s":0.1,"tail_ramp_s":0.15,"score_kind":"structural plausibility, not model confidence"},"frames":48,"live_alignment":{"enabled":true,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":3.79}}],"quick_reply_audio_s":2.72,"quick_reply_tts_ms":2005.44,"sentence_tts":[{"sentence":0,"role":"reaction","audio_s":2.72,"total_ms":2005.44,"prefetched":true},{"sentence":1,"role":"answer","audio_s":3.79,"total_ms":3201.42,"prefetched":true}],"answer_first_audio_ms":2392.69,"answer_audio_ready_ms":1964.1},"at":13061.699996948242},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"played","tts_backend":"continuous_batch","tts_weight_instances":1,"tts_temperature":0.8,"informed_bridge_status":"replaced_by_informed_main","main_informed_by_reaction":true,"main_launch_ms":382.4,"llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","quick_reply_generation_ms":376.98,"quick_reply_model":"nemotron-bolt","quick_reply_memory_entries":0,"quick_reply_memory_person_matched":false,"quick_reply_words":4,"first_audio_ms":569.2,"tts_first_audio_ms":184.1,"llm_actual_provider":"Cerebras","llm_ttft_ms":388.72,"llm_first_sentence_ms":411.87,"duplicate_first_ack_skipped":true,"finish_reason":"stop","usage":{"prompt_tokens":2360,"completion_tokens":54,"total_tokens":2414,"cost":0.0023926914,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00241686,"upstream_inference_prompt_cost":0.0023364,"upstream_inference_completions_cost":8.046e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":470.39,"response_complete":true,"tts_server_stages":[{"generation_ms":2003.17,"alignment_ms":0.01,"alignment":null,"frames":34,"live_alignment":{"enabled":false,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":0}},{"generation_ms":3190.69,"alignment_ms":37.27,"alignment":{"applied":true,"lead_s":0.16,"cut_s":3.64,"removed_s":0.05,"words":9,"note":"","spans":[["Ein",0.16,0.32,1],["schneller",0.32,0.72,1],["elektronischer",0.72,1.36,1],["Track",1.36,1.76,1],["mit",1.76,1.84,1],["einer",1.92,2.16,1],["sehr",2.16,2.4,1],["treibenden",2.4,2.96,1],["Basslinie",2.96,3.52,1]],"enabled":true,"available":true,"raw_seconds":3.84,"seconds":3.79,"lead_ramp_s":0.1,"tail_ramp_s":0.15,"score_kind":"structural plausibility, not model confidence"},"frames":48,"live_alignment":{"enabled":true,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":3.79}},{"generation_ms":2521.53,"alignment_ms":37.11,"alignment":{"applied":true,"lead_s":0.32,"cut_s":3.4,"removed_s":0.13,"words":9,"note":"","spans":[["Klingt",0.32,0.8,1],["als",0.88,1.12,1],["würde",1.12,1.28,1],["jemand",1.28,1.52,1],["gerade",1.52,1.76,1],["auf",1.76,1.92,1],["Party",2,2.48,1],["Modus",2.48,2.88,1],["drücken",2.88,3.28,1]],"enabled":true,"available":true,"raw_seconds":3.68,"seconds":3.55,"lead_ramp_s":0.1,"tail_ramp_s":0.15,"score_kind":"structural plausibility, not model confidence"},"frames":46,"live_alignment":{"enabled":true,"backend":"Parakeet v3 word timestamps + final Qwen forced alignment","lead_s":null,"cut_s":null,"stopped":false,"lookahead_seconds":1,"emitted_seconds":3.55}}],"quick_reply_audio_s":2.72,"quick_reply_tts_ms":2005.44,"sentence_tts":[{"sentence":0,"role":"reaction","audio_s":2.72,"total_ms":2005.44,"prefetched":true},{"sentence":1,"role":"answer","audio_s":3.79,"total_ms":3201.42,"prefetched":true},{"sentence":2,"role":"answer","audio_s":3.55,"total_ms":4062.69,"prefetched":true}],"answer_first_audio_ms":2392.69,"answer_audio_ready_ms":1964.1,"total_ms":4918.03,"audio_s":10.06,"sentences":3,"continuous_tts":{"scheduler":"continuous CFG batch","weight_instances":1,"batch_histogram":{"1":33,"2":49},"prefill_ms":[58.95,43.03,40.59],"frame_ms":[11.01,113,18.28,17.72,17.68,17.67,111.17,18.05,16.93,16.94,17.53,112.64,119.81,20.61,19.72,19.64,114.27,113.71,20.25,19.7,19.7,113.45,112.82,20.35,20.44,20.1,116.32,112.89,20.4,19.85,19.84,115.21,113.93,20.99,112.97,19.56,116.26,114.12,20.5,19.9,20.55,116.3,114.17,20.47,19.92,19.88,112.55,114.04,21.19,20.34,19.98,113.05,113.82,20.54,20.07,19.98,113.22,117.28,20.98,151.09,17.83,112.31,18.42,17.5,17.12,17.1,109.56,17.69,17.15,17.13,17.82,110.42,17.7,17.14,17.1,17.09,110.01,17.73,17.89,17.45,17.12,151.22],"max_active":2,"jobs":3,"cancelled_jobs":0,"total_ms":4895.85,"gpu_allocated_mb":10982.2},"generation_rtf":0.4889},"at":13949.699996948242},{"type":"turn_done","barged":false,"at":20026.79999923706}],"reply":"Das klingt nach Musik.\nEin schneller elektronischer Track mit einer sehr treibenden Basslinie.\nKlingt, als würde jemand gerade auf „Party-Modus“ drücken!","captions":[{"status":"available","caption":"The audio features a fast-paced electronic track with a prominent, repetitive synth melody. A driving beat is present, with a strong bassline and a consistent rhythmic pulse.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.4,"tokens":35,"preprocess_ms":9.53,"encoder_gpu_ms":9.09,"decoder_gpu_ms":333.23,"inference_ms":342.55,"total_ms":352.08,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":28.05,"window_s":2.4,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1397},{"label":"Synthesizer","score":0.0709},{"label":"Drum machine","score":0.0705},{"label":"Electric piano","score":0.0644},{"label":"Keyboard (musical)","score":0.0612}],"music_score":0.9159,"speech_score":0.0159,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8257},{"label":"Ringtone","score":0.0555},{"label":"Musical instrument","score":0.0549}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.9159},{"label":"Musical instrument","score":0.1042},{"label":"Synthesizer","score":0.0667}]},{"start_s":1.12,"end_s":2.4,"candidates":[{"label":"Music","score":0.8883},{"label":"Musical instrument","score":0.1397},{"label":"Synthesizer","score":0.0709}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.42},"type":"sound_caption","request_ms":368.33,"capture_ms":2400,"event_age_ms":2768.34,"kind":"music","eligible":true,"suppression_reason":null,"requested":true,"at":9025.599998474121}],"pcm_chunks":28,"page_errors":[],"passed":true},{"case":"spoken_question_then_music","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":4335.200000762939},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":4863.200000762939},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":5301.399997711182},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":5822.299999237061},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":5976.099998474121},{"type":"asr_preview","text":"Hopy","status":"sentence_stream","source":"local_sentence_asr","at":6041.5},{"status":"available","caption":"The audio contains a single, short, and somewhat muffled vocalization. It sounds like a soft, breathy sigh or a quiet utterance of a word.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.84,"tokens":34,"preprocess_ms":9.78,"encoder_gpu_ms":8.11,"decoder_gpu_ms":258.65,"inference_ms":266.97,"total_ms":276.76,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":23.42,"window_s":1.84,"candidates":[{"label":"Speech","score":0.4056},{"label":"Sigh","score":0.3286},{"label":"Gasp","score":0.2831},{"label":"Snort","score":0.0881},{"label":"Breathing","score":0.0834},{"label":"Wheeze","score":0.0803}],"music_score":0.0464,"speech_score":0.4056,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Sigh","score":0.3286},{"label":"Speech","score":0.2143},{"label":"Breathing","score":0.0834}]},{"start_s":0.56,"end_s":1.84,"candidates":[{"label":"Gasp","score":0.2831},{"label":"Speech","score":0.2012},{"label":"Sigh","score":0.0964}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":30.55},"type":"sound_caption","request_ms":289.35,"capture_ms":1640,"event_age_ms":1929.36,"kind":"sound","eligible":true,"suppression_reason":null,"at":6124.899997711182},{"type":"asr_preview","text":"Hoppi, was ist","status":"sentence_stream","source":"local_sentence_asr","at":6513.799999237061},{"type":"asr_preview","text":"Hoppi, das ist das","status":"sentence_stream","source":"native_asr","at":7002},{"type":"asr_preview","text":"Hoppi, was ist denn das hier","status":"sentence_stream","source":"local_sentence_asr","at":7828.5},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":7930.299999237061},{"type":"asr_preview","text":"Hobby, was ist denn das hier?","status":"sentence_stream","source":"local_sentence_asr","at":8021.599998474121},{"type":"asr_sentence","text":"Hobby, was ist denn das hier?","audio_s":6.08,"queued":false,"asr":{"text":"Hobby, was ist denn das hier?","model":"parakeet_v3","chunks":1,"transcribe_ms":76.85,"speech_pause_ms":700,"observed_quiet_ms":524},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":524,"at":8022.200000762939},{"type":"identity","enabled":false,"mode":"open-sentences","sentence_mode":true,"open_enrollment":true,"speakers":[{"id":"speaker-1","name":"Speaker 1","named":false,"reference_ready":true,"voice_variants":1,"admitted":true}],"users":["Speaker 1"],"threshold":0.38,"conversation_focus":{"enabled":false,"threshold":0.28,"timeout_s":180,"speaker_id":null,"name":null,"primary_id":"speaker-1","participants":[{"id":"speaker-1","name":"Speaker 1"}]},"handover_pending":false,"storage":"Session RAM","prediction":{"name":"Speaker 1","similarity":-1},"at":8090.700000762939},{"type":"voice_caption","status":"available","caption":"Possibly very breathy, very choppy, slightly slow, very flabbergasted","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":2.72,"inference_ms":26.06,"model":"laion/humaneness-ears-base-medium","model_revision":"818506970d93c809a3295a0a02c40dee8ff2bfc3","rules_revision":"7ebd4a9b7da96df989af8fc6ad0ce34f4e8aa3ed","device":"cuda:1","precision":"bfloat16","selected":{"voicenet":[{"dim":"RESP","name":"Respiration","value":4.075,"z":3.32,"direction":"above","intensity":"Extremely","always_on":false,"phrase":"extremely audibly breathing and gasping","tag":"very breathy","reliability":1,"rank_score":3.319},{"dim":"CHNK","name":"Chunking","value":1.623,"z":-2.19,"direction":"below","intensity":"Extremely","always_on":false,"phrase":"extremely choppy and fragmented","tag":"very choppy","reliability":1,"rank_score":2.194},{"dim":"TEMP","name":"Tempo","value":2.528,"z":-0.83,"direction":"below","intensity":"Somewhat","always_on":true,"phrase":"somewhat slow and deliberate in tempo","tag":"slightly slow","reliability":1,"rank_score":0.829}],"emotions":[{"emotion":"Confusion","value":2.018,"z":4.32,"direction":"above","intensity":"Extremely","synonym":"flabbergasted","phrase":"extremely carrying flabbergasted","tag":"very flabbergasted"}]},"normalization":"speech reference z-scores; not probabilities","scope":"this turn","request_ms":35.65,"added_wait_ms":0,"at":8094},{"type":"endpoint","transcript":"Hobby, was ist denn das hier?","speaker_id":"speaker-1","asr":{"text":"Hobby, was ist denn das hier?","model":"parakeet_v3","chunks":1,"transcribe_ms":76.85,"speech_pause_ms":700,"observed_quiet_ms":524},"speaker_name":"Speaker 1","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":73.48,"speaker_embedding_wait_ms":72.488,"speaker_lookup_ms":0.992,"speaker_prefetch_used":false,"speaker_prefetch_compute_ms":null,"sentence_queue_ms":73.66,"sentence_audio_s":6.08},"at":8094.099998474121},{"type":"environment_observation","kind":"requested_sound","status":"listening_after_question","at":8094.700000762939},{"status":"available","caption":"The audio features a musical piece with a moderate tempo. The instrumentation includes a prominent electric guitar playing a melodic riff, accompanied by a steady drum beat and bassline.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.4,"tokens":36,"preprocess_ms":7.86,"encoder_gpu_ms":8.59,"decoder_gpu_ms":262.75,"inference_ms":271.59,"total_ms":279.44,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":26.87,"window_s":2.4,"candidates":[{"label":"Music","score":0.8706},{"label":"Drum machine","score":0.2128},{"label":"Musical instrument","score":0.1262},{"label":"Keyboard (musical)","score":0.0835},{"label":"Electric piano","score":0.0816},{"label":"Synthesizer","score":0.0763}],"music_score":0.8706,"speech_score":0.0177,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8706},{"label":"Musical instrument","score":0.1008},{"label":"Keyboard (musical)","score":0.0835}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.8612},{"label":"Musical instrument","score":0.1262},{"label":"Drum machine","score":0.089}]},{"start_s":1.12,"end_s":2.4,"candidates":[{"label":"Music","score":0.801},{"label":"Drum machine","score":0.2128},{"label":"Musical instrument","score":0.1128}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.83},"type":"sound_caption","request_ms":296.54,"capture_ms":2400,"event_age_ms":2696.54,"kind":"music","eligible":true,"suppression_reason":null,"requested":true,"at":10773.099998474121},{"type":"latency_start","reply_id":1,"input":{"kind":"voice","asr":{"text":"Hobby, was ist denn das hier?","model":"parakeet_v3","chunks":1,"transcribe_ms":76.85,"speech_pause_ms":700,"observed_quiet_ms":524},"voice_caption_ms":35.65,"voice_caption_status":"available","speaker_assignment_ms":73.48,"speaker_prefetch_compute_ms":null,"speaker_prefetch_used":false,"sentence_queue_ms":73.66},"at":10773.20000076294},{"type":"reply","text":"Ich höre gerade etwas Musik im Hintergrund.","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Ich höre gerade etwas Musik im Hintergrund.","start_s":0,"end_s":2.38}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Ich höre gerade etwas Musik im Hintergrund.","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Speaker 1","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"sentence_ready_ms":787.19,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":11563.5},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75},"at":12320.5},{"type":"reply","text":"Ich höre gerade etwas Musik im Hintergrund.\nEs klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Ich höre gerade etwas Musik im Hintergrund.","start_s":0,"end_s":2.83},{"emotion":"contentment","secondary":null,"blend":0,"intensity":0.72,"direction":"natural, calm","text":"Es klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.","start_s":2.83,"end_s":6.91}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Es klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Speaker 1","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75,"tts_server_stages":[{"generation_ms":1094,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1248,"performance":{"device_seconds":0.8858450121879577,"raw_audio_seconds":3.12,"aggregate_rtf":0.2839246833935762,"audio_seconds_per_second":3.5220608086891856,"batches":[{"batch_size":1,"device_seconds":0.8858450121879577,"audio_seconds":[3.12],"model_wall_seconds":0.9795265421271324,"first_audio_seconds":0.46653338623046875,"graph_capture_seconds":0,"live_callback_seconds":0.09096918255090714}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1260.87,"audio_s":2.83}],"sentence_ready_ms":2049.02,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.natural.soft"},"sources":[],"lookups":[],"at":12949.20000076294},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75,"tts_server_stages":[{"generation_ms":1094,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1248,"performance":{"device_seconds":0.8858450121879577,"raw_audio_seconds":3.12,"aggregate_rtf":0.2839246833935762,"audio_seconds_per_second":3.5220608086891856,"batches":[{"batch_size":1,"device_seconds":0.8858450121879577,"audio_seconds":[3.12],"model_wall_seconds":0.9795265421271324,"first_audio_seconds":0.46653338623046875,"graph_capture_seconds":0,"live_callback_seconds":0.09096918255090714}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1260.87,"audio_s":2.83}]},"at":13587.5},{"type":"reply","text":"Ich höre gerade etwas Musik im Hintergrund.\nEs klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.\nIst das ein Song, der dir besonders gefällt?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Ich höre gerade etwas Musik im Hintergrund.","start_s":0,"end_s":2.83},{"emotion":"contentment","secondary":null,"blend":0,"intensity":0.72,"direction":"natural, calm","text":"Es klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.","start_s":2.83,"end_s":7.82},{"emotion":"amusement","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"playful, warm","text":"Ist das ein Song, der dir besonders gefällt?","start_s":7.82,"end_s":10.540000000000001}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":2,"sentence_text":"Ist das ein Song, der dir besonders gefällt?","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Speaker 1","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75,"tts_server_stages":[{"generation_ms":1094,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1248,"performance":{"device_seconds":0.8858450121879577,"raw_audio_seconds":3.12,"aggregate_rtf":0.2839246833935762,"audio_seconds_per_second":3.5220608086891856,"batches":[{"batch_size":1,"device_seconds":0.8858450121879577,"audio_seconds":[3.12],"model_wall_seconds":0.9795265421271324,"first_audio_seconds":0.46653338623046875,"graph_capture_seconds":0,"live_callback_seconds":0.09096918255090714}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1823,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1980,"performance":{"device_seconds":1.4755313563346864,"raw_audio_seconds":5.04,"aggregate_rtf":0.29276415800291394,"audio_seconds_per_second":3.415718668642651,"batches":[{"batch_size":1,"device_seconds":1.4755313563346864,"audio_seconds":[5.04],"model_wall_seconds":1.7108753323554993,"first_audio_seconds":0.469484619140625,"graph_capture_seconds":0,"live_callback_seconds":0.23069273680448532}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1260.87,"audio_s":2.83},{"sentence":1,"total_ms":1996.62,"audio_s":4.99}],"sentence_ready_ms":4046.66,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.warm.soft"},"sources":[],"lookups":[],"at":14841.099998474121},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75,"tts_server_stages":[{"generation_ms":1094,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1248,"performance":{"device_seconds":0.8858450121879577,"raw_audio_seconds":3.12,"aggregate_rtf":0.2839246833935762,"audio_seconds_per_second":3.5220608086891856,"batches":[{"batch_size":1,"device_seconds":0.8858450121879577,"audio_seconds":[3.12],"model_wall_seconds":0.9795265421271324,"first_audio_seconds":0.46653338623046875,"graph_capture_seconds":0,"live_callback_seconds":0.09096918255090714}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1823,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1980,"performance":{"device_seconds":1.4755313563346864,"raw_audio_seconds":5.04,"aggregate_rtf":0.29276415800291394,"audio_seconds_per_second":3.415718668642651,"batches":[{"batch_size":1,"device_seconds":1.4755313563346864,"audio_seconds":[5.04],"model_wall_seconds":1.7108753323554993,"first_audio_seconds":0.469484619140625,"graph_capture_seconds":0,"live_callback_seconds":0.23069273680448532}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1260.87,"audio_s":2.83},{"sentence":1,"total_ms":1996.62,"audio_s":4.99}]},"at":15572.899997711182},{"type":"reply","text":"Ich höre gerade etwas Musik im Hintergrund.\nEs klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.\nIst das ein Song, der dir besonders gefällt? Und wie darf ich dich nennen?","reply_id":1,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"Ich höre gerade etwas Musik im Hintergrund.","start_s":0,"end_s":2.83},{"emotion":"contentment","secondary":null,"blend":0,"intensity":0.72,"direction":"natural, calm","text":"Es klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.","start_s":2.83,"end_s":7.82},{"emotion":"amusement","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"playful, warm","text":"Ist das ein Song, der dir besonders gefällt?","start_s":7.82,"end_s":11.21},{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.55,"direction":"interest, soft, warm","text":"Und wie darf ich dich nennen?","start_s":11.21,"end_s":13.25}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":3,"sentence_text":"Und wie darf ich dich nennen?","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Speaker 1","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75,"tts_server_stages":[{"generation_ms":1094,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1248,"performance":{"device_seconds":0.8858450121879577,"raw_audio_seconds":3.12,"aggregate_rtf":0.2839246833935762,"audio_seconds_per_second":3.5220608086891856,"batches":[{"batch_size":1,"device_seconds":0.8858450121879577,"audio_seconds":[3.12],"model_wall_seconds":0.9795265421271324,"first_audio_seconds":0.46653338623046875,"graph_capture_seconds":0,"live_callback_seconds":0.09096918255090714}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1823,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1980,"performance":{"device_seconds":1.4755313563346864,"raw_audio_seconds":5.04,"aggregate_rtf":0.29276415800291394,"audio_seconds_per_second":3.415718668642651,"batches":[{"batch_size":1,"device_seconds":1.4755313563346864,"audio_seconds":[5.04],"model_wall_seconds":1.7108753323554993,"first_audio_seconds":0.469484619140625,"graph_capture_seconds":0,"live_callback_seconds":0.23069273680448532}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1336,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1488,"performance":{"device_seconds":1.0763843870162964,"raw_audio_seconds":3.6,"aggregate_rtf":0.2989956630600823,"audio_seconds_per_second":3.3445301171443846,"batches":[{"batch_size":1,"device_seconds":1.0763843870162964,"audio_seconds":[3.6],"model_wall_seconds":1.2261390760540962,"first_audio_seconds":0.4602345886230469,"graph_capture_seconds":0,"live_callback_seconds":0.14763828366994858}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1260.87,"audio_s":2.83},{"sentence":1,"total_ms":1996.62,"audio_s":4.99},{"sentence":2,"total_ms":1495.26,"audio_s":3.39}],"sentence_ready_ms":5543.25,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":16323.20000076294},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75,"tts_server_stages":[{"generation_ms":1094,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1248,"performance":{"device_seconds":0.8858450121879577,"raw_audio_seconds":3.12,"aggregate_rtf":0.2839246833935762,"audio_seconds_per_second":3.5220608086891856,"batches":[{"batch_size":1,"device_seconds":0.8858450121879577,"audio_seconds":[3.12],"model_wall_seconds":0.9795265421271324,"first_audio_seconds":0.46653338623046875,"graph_capture_seconds":0,"live_callback_seconds":0.09096918255090714}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1823,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1980,"performance":{"device_seconds":1.4755313563346864,"raw_audio_seconds":5.04,"aggregate_rtf":0.29276415800291394,"audio_seconds_per_second":3.415718668642651,"batches":[{"batch_size":1,"device_seconds":1.4755313563346864,"audio_seconds":[5.04],"model_wall_seconds":1.7108753323554993,"first_audio_seconds":0.469484619140625,"graph_capture_seconds":0,"live_callback_seconds":0.23069273680448532}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1336,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1488,"performance":{"device_seconds":1.0763843870162964,"raw_audio_seconds":3.6,"aggregate_rtf":0.2989956630600823,"audio_seconds_per_second":3.3445301171443846,"batches":[{"batch_size":1,"device_seconds":1.0763843870162964,"audio_seconds":[3.6],"model_wall_seconds":1.2261390760540962,"first_audio_seconds":0.4602345886230469,"graph_capture_seconds":0,"live_callback_seconds":0.14763828366994858}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1260.87,"audio_s":2.83},{"sentence":1,"total_ms":1996.62,"audio_s":4.99},{"sentence":2,"total_ms":1495.26,"audio_s":3.39}]},"at":17057.89999771118},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":741.24,"llm_first_sentence_ms":786.31,"finish_reason":"stop","usage":{"prompt_tokens":2710,"completion_tokens":58,"total_tokens":2768,"cost":0.0027416268,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00276932,"upstream_inference_prompt_cost":0.0026829,"upstream_inference_completions_cost":8.642e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":849.26,"response_complete":true,"answer_first_audio_ms":1539.74,"tts_first_audio_ms":752.44,"first_audio_ms":1539.75,"tts_server_stages":[{"generation_ms":1094,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1248,"performance":{"device_seconds":0.8858450121879577,"raw_audio_seconds":3.12,"aggregate_rtf":0.2839246833935762,"audio_seconds_per_second":3.5220608086891856,"batches":[{"batch_size":1,"device_seconds":0.8858450121879577,"audio_seconds":[3.12],"model_wall_seconds":0.9795265421271324,"first_audio_seconds":0.46653338623046875,"graph_capture_seconds":0,"live_callback_seconds":0.09096918255090714}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1823,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1980,"performance":{"device_seconds":1.4755313563346864,"raw_audio_seconds":5.04,"aggregate_rtf":0.29276415800291394,"audio_seconds_per_second":3.415718668642651,"batches":[{"batch_size":1,"device_seconds":1.4755313563346864,"audio_seconds":[5.04],"model_wall_seconds":1.7108753323554993,"first_audio_seconds":0.469484619140625,"graph_capture_seconds":0,"live_callback_seconds":0.23069273680448532}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1336,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1488,"performance":{"device_seconds":1.0763843870162964,"raw_audio_seconds":3.6,"aggregate_rtf":0.2989956630600823,"audio_seconds_per_second":3.3445301171443846,"batches":[{"batch_size":1,"device_seconds":1.0763843870162964,"audio_seconds":[3.6],"model_wall_seconds":1.2261390760540962,"first_audio_seconds":0.4602345886230469,"graph_capture_seconds":0,"live_callback_seconds":0.14763828366994858}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1028,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1182,"performance":{"device_seconds":0.8327944688796997,"raw_audio_seconds":2.88,"aggregate_rtf":0.2891647461387846,"audio_seconds_per_second":3.4582362246885032,"batches":[{"batch_size":1,"device_seconds":0.8327944688796997,"audio_seconds":[2.88],"model_wall_seconds":0.9195641130208969,"first_audio_seconds":0.4591247253417969,"graph_capture_seconds":0,"live_callback_seconds":0.08654589205980301}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1260.87,"audio_s":2.83},{"sentence":1,"total_ms":1996.62,"audio_s":4.99},{"sentence":2,"total_ms":1495.26,"audio_s":3.39},{"sentence":3,"total_ms":1189.33,"audio_s":2.43}],"total_ms":6732.84,"audio_s":13.64,"sentences":4,"generation_rtf":0.4936},"at":17507.79999923706},{"type":"turn_done","barged":false,"at":26407.39999771118}],"reply":"Ich höre gerade etwas Musik im Hintergrund.\nEs klingt nach einem elektrischen Gitarrenriff mit einem gleichmäßigen Drumschlag und Bass.\nIst das ein Song, der dir besonders gefällt? Und wie darf ich dich nennen?","captions":[{"status":"available","caption":"The audio contains a single, short, and somewhat muffled vocalization. It sounds like a soft, breathy sigh or a quiet utterance of a word.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.84,"tokens":34,"preprocess_ms":9.78,"encoder_gpu_ms":8.11,"decoder_gpu_ms":258.65,"inference_ms":266.97,"total_ms":276.76,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":23.42,"window_s":1.84,"candidates":[{"label":"Speech","score":0.4056},{"label":"Sigh","score":0.3286},{"label":"Gasp","score":0.2831},{"label":"Snort","score":0.0881},{"label":"Breathing","score":0.0834},{"label":"Wheeze","score":0.0803}],"music_score":0.0464,"speech_score":0.4056,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Sigh","score":0.3286},{"label":"Speech","score":0.2143},{"label":"Breathing","score":0.0834}]},{"start_s":0.56,"end_s":1.84,"candidates":[{"label":"Gasp","score":0.2831},{"label":"Speech","score":0.2012},{"label":"Sigh","score":0.0964}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":30.55},"type":"sound_caption","request_ms":289.35,"capture_ms":1640,"event_age_ms":1929.36,"kind":"sound","eligible":true,"suppression_reason":null,"at":6124.899997711182},{"status":"available","caption":"The audio features a musical piece with a moderate tempo. The instrumentation includes a prominent electric guitar playing a melodic riff, accompanied by a steady drum beat and bassline.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.4,"tokens":36,"preprocess_ms":7.86,"encoder_gpu_ms":8.59,"decoder_gpu_ms":262.75,"inference_ms":271.59,"total_ms":279.44,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":26.87,"window_s":2.4,"candidates":[{"label":"Music","score":0.8706},{"label":"Drum machine","score":0.2128},{"label":"Musical instrument","score":0.1262},{"label":"Keyboard (musical)","score":0.0835},{"label":"Electric piano","score":0.0816},{"label":"Synthesizer","score":0.0763}],"music_score":0.8706,"speech_score":0.0177,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Music","score":0.8706},{"label":"Musical instrument","score":0.1008},{"label":"Keyboard (musical)","score":0.0835}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Music","score":0.8612},{"label":"Musical instrument","score":0.1262},{"label":"Drum machine","score":0.089}]},{"start_s":1.12,"end_s":2.4,"candidates":[{"label":"Music","score":0.801},{"label":"Drum machine","score":0.2128},{"label":"Musical instrument","score":0.1128}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.83},"type":"sound_caption","request_ms":296.54,"capture_ms":2400,"event_age_ms":2696.54,"kind":"music","eligible":true,"suppression_reason":null,"requested":true,"at":10773.099998474121}],"pcm_chunks":17,"page_errors":[],"passed":true,"transcripts":["Hobby, was ist denn das hier?"]},{"case":"christoph-intro","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":4989.099998474121},{"type":"asr_preview","text":"Hey Rao.","status":"sentence_stream","source":"local_sentence_asr","at":5475.400001525879},{"type":"asr_preview","text":"Hey Rob, ich","status":"sentence_stream","source":"local_sentence_asr","at":5997.099998474121},{"type":"asr_preview","text":"Hey Robbie, bin.","status":"sentence_stream","source":"native_asr","at":6607.900001525879},{"status":"available","caption":"The audio contains a single male voice speaking in German. The speaker's tone is conversational and relatively neutral, with a moderate pace.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.12,"tokens":28,"preprocess_ms":9.42,"encoder_gpu_ms":16.44,"decoder_gpu_ms":257.39,"inference_ms":274.04,"total_ms":283.46,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":29.37,"window_s":2.12,"candidates":[{"label":"Speech","score":0.8537},{"label":"Speech synthesizer","score":0.4532},{"label":"Narration, monologue","score":0.1053},{"label":"Music","score":0.0936},{"label":"Clicking","score":0.0243},{"label":"Animal","score":0.0238}],"music_score":0.0936,"speech_score":0.8537,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.7348},{"label":"Speech synthesizer","score":0.1177},{"label":"Music","score":0.0936}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.8537},{"label":"Speech synthesizer","score":0.3911},{"label":"Narration, monologue","score":0.1053}]},{"start_s":0.84,"end_s":2.12,"candidates":[{"label":"Speech","score":0.7871},{"label":"Speech synthesizer","score":0.3879},{"label":"Narration, monologue","score":0.0928}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":38.18},"type":"sound_caption","request_ms":301.34,"capture_ms":1920,"event_age_ms":2221.34,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":7096},{"type":"asr_preview","text":"Hey Rabbi, ich bin Christoph.","status":"sentence_stream","source":"local_sentence_asr","at":7359.5},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":7407.299999237061},{"type":"asr_preview","text":"Hey Robbie, ich bin Christoph.","status":"sentence_stream","source":"local_sentence_asr","at":7469.5},{"type":"asr_sentence","text":"Hey Robbie, ich bin Christoph.","audio_s":5.12,"queued":false,"asr":{"text":"Hey Robbie, ich bin Christoph.","model":"parakeet_v3","chunks":1,"transcribe_ms":53.9,"speech_pause_ms":700,"observed_quiet_ms":16},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":16,"at":7469.799999237061},{"type":"identity","enabled":false,"mode":"open-sentences","sentence_mode":true,"open_enrollment":true,"speakers":[{"id":"speaker-1","name":"Christoph","named":true,"reference_ready":true,"voice_variants":1,"admitted":true}],"users":["Christoph"],"threshold":0.38,"conversation_focus":{"enabled":false,"threshold":0.28,"timeout_s":180,"speaker_id":null,"name":null,"primary_id":"speaker-1","participants":[{"id":"speaker-1","name":"Christoph"}]},"handover_pending":false,"storage":"Session RAM","prediction":{"name":"Christoph","similarity":-1},"at":7471.799999237061},{"type":"voice_caption","status":"available","caption":"Possibly slightly smooth, slightly steady, very liberation","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":2.12,"inference_ms":44.99,"model":"laion/humaneness-ears-base-medium","model_revision":"818506970d93c809a3295a0a02c40dee8ff2bfc3","rules_revision":"7ebd4a9b7da96df989af8fc6ad0ce34f4e8aa3ed","device":"cuda:1","precision":"bfloat16","selected":{"voicenet":[{"dim":"DFLU","name":"Disfluency","value":0.397,"z":-0.9,"direction":"below","intensity":"Somewhat","always_on":false,"phrase":"somewhat smooth and flawless","tag":"slightly smooth","reliability":1,"rank_score":0.898},{"dim":"VOLT","name":"Volatility","value":1.314,"z":-0.76,"direction":"below","intensity":"Somewhat","always_on":false,"phrase":"somewhat emotionally steady and static","tag":"slightly steady","reliability":1,"rank_score":0.763},{"dim":"TEMP","name":"Tempo","value":2.828,"z":-0.25,"direction":"below","intensity":"About-average","always_on":true,"phrase":"moderate in tempo","tag":null,"reliability":1,"rank_score":0.254}],"emotions":[{"emotion":"Relief","value":1.095,"z":1.85,"direction":"above","intensity":"Very","synonym":"liberation","phrase":"very carrying liberation","tag":"very liberation"}]},"normalization":"speech reference z-scores; not probabilities","scope":"this turn","request_ms":52.73,"added_wait_ms":0,"at":7475.799999237061},{"type":"endpoint","transcript":"Hey Robbie, ich bin Christoph.","speaker_id":"speaker-1","asr":{"text":"Hey Robbie, ich bin Christoph.","model":"parakeet_v3","chunks":1,"transcribe_ms":53.9,"speech_pause_ms":700,"observed_quiet_ms":16},"speaker_name":"Christoph","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":0.75,"speaker_embedding_wait_ms":0.002,"speaker_lookup_ms":0.748,"speaker_prefetch_used":true,"speaker_prefetch_compute_ms":61.76,"sentence_queue_ms":1.15,"sentence_audio_s":5.12},"at":7476},{"type":"latency_start","reply_id":1,"input":{"kind":"voice","asr":{"text":"Hey Robbie, ich bin Christoph.","model":"parakeet_v3","chunks":1,"transcribe_ms":53.9,"speech_pause_ms":700,"observed_quiet_ms":16},"voice_caption_ms":52.73,"voice_caption_status":"available","speaker_assignment_ms":0.75,"speaker_prefetch_compute_ms":61.76,"speaker_prefetch_used":true,"sentence_queue_ms":1.14},"at":7476.700000762939},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":7947.299999237061},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":8447.20000076294},{"type":"reply","text":"Hallo Christoph!","reply_id":1,"performance":{"sentences":[{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"friendly, warm","text":"Hallo Christoph!","start_s":0,"end_s":0.68}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Hallo Christoph!","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Christoph","speaker_update":{"speaker_id":"speaker-1","display_name":"Christoph","is_self_introduction":true,"source":"validated_acoustic_transcript"},"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1384.69,"llm_first_sentence_ms":1385.21,"sentence_ready_ms":1386.37,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.warm.soft"},"sources":[],"lookups":[],"at":8859.400001525879},{"type":"asr_preview","text":"schön dich kennenzulernen.","status":"sentence_stream","source":"local_sentence_asr","at":8952.599998474121},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":9253.099998474121},{"type":"asr_preview","text":"Schön, dich kennenzulernen.","status":"sentence_stream","source":"local_sentence_asr","at":9314.599998474121},{"type":"asr_sentence","text":"Schön, dich kennenzulernen.","audio_s":7.04,"queued":true,"asr":{"text":"Schön, dich kennenzulernen.","model":"parakeet_v3","chunks":1,"transcribe_ms":50.67,"speech_pause_ms":700,"observed_quiet_ms":598},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":598,"at":9314.900001525879},{"type":"identity","enabled":false,"mode":"open-sentences","sentence_mode":true,"open_enrollment":true,"speakers":[{"id":"speaker-1","name":"Christoph","named":true,"reference_ready":true,"voice_variants":1,"admitted":true}],"users":["Christoph"],"threshold":0.38,"conversation_focus":{"enabled":false,"threshold":0.28,"timeout_s":180,"speaker_id":null,"name":null,"primary_id":"speaker-1","participants":[{"id":"speaker-1","name":"Christoph"}]},"handover_pending":false,"storage":"Session RAM","prediction":{"name":"Christoph","similarity":0.3914},"at":9315.20000076294},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1384.69,"llm_first_sentence_ms":1385.21,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":30,"total_tokens":1775,"cost":0.0017545275,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00177225,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":4.47e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":1504.49,"response_complete":true,"answer_first_audio_ms":2146.35,"tts_first_audio_ms":759.83,"first_audio_ms":2146.35},"at":9624.599998474121},{"type":"reply","text":"Hallo Christoph! Schön, dass du dich vorstellst.","reply_id":1,"performance":{"sentences":[{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"friendly, warm","text":"Hallo Christoph!","start_s":0,"end_s":2.19},{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"friendly, warm","text":"Schön, dass du dich vorstellst.","start_s":2.19,"end_s":3.8899999999999997}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Schön, dass du dich vorstellst.","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Christoph","speaker_update":{"speaker_id":"speaker-1","display_name":"Christoph","is_self_introduction":true,"source":"validated_acoustic_transcript"},"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1384.69,"llm_first_sentence_ms":1385.21,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":30,"total_tokens":1775,"cost":0.0017545275,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00177225,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":4.47e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":1504.49,"response_complete":true,"answer_first_audio_ms":2146.35,"tts_first_audio_ms":759.83,"first_audio_ms":2146.35,"tts_server_stages":[{"generation_ms":1076,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1242,"performance":{"device_seconds":0.86781423664093,"raw_audio_seconds":3.04,"aggregate_rtf":0.28546520942135856,"audio_seconds_per_second":3.5030538468313255,"batches":[{"batch_size":1,"device_seconds":0.86781423664093,"audio_seconds":[3.04],"model_wall_seconds":0.960164837539196,"first_audio_seconds":0.46063922119140627,"graph_capture_seconds":0,"live_callback_seconds":0.09010444581508636}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1251.04,"audio_s":2.19}],"sentence_ready_ms":2638.03,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.warm.soft"},"sources":[],"lookups":[],"at":10154.599998474121},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1384.69,"llm_first_sentence_ms":1385.21,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":30,"total_tokens":1775,"cost":0.0017545275,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00177225,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":4.47e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":1504.49,"response_complete":true,"answer_first_audio_ms":2146.35,"tts_first_audio_ms":759.83,"first_audio_ms":2146.35,"tts_server_stages":[{"generation_ms":1076,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1242,"performance":{"device_seconds":0.86781423664093,"raw_audio_seconds":3.04,"aggregate_rtf":0.28546520942135856,"audio_seconds_per_second":3.5030538468313255,"batches":[{"batch_size":1,"device_seconds":0.86781423664093,"audio_seconds":[3.04],"model_wall_seconds":0.960164837539196,"first_audio_seconds":0.46063922119140627,"graph_capture_seconds":0,"live_callback_seconds":0.09010444581508636}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1251.04,"audio_s":2.19}]},"at":10848.099998474121},{"type":"reply","text":"Hallo Christoph! Schön, dass du dich vorstellst. Ich bin Robby, dein kleiner Roboter-Buddy.","reply_id":1,"performance":{"sentences":[{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"friendly, warm","text":"Hallo Christoph!","start_s":0,"end_s":2.19},{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"friendly, warm","text":"Schön, dass du dich vorstellst.","start_s":2.19,"end_s":4.7},{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"friendly, warm","text":"Ich bin Robby, dein kleiner Roboter-Buddy.","start_s":4.7,"end_s":6.74}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":2,"sentence_text":"Ich bin Robby, dein kleiner Roboter-Buddy.","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Christoph","speaker_update":{"speaker_id":"speaker-1","display_name":"Christoph","is_self_introduction":true,"source":"validated_acoustic_transcript"},"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1384.69,"llm_first_sentence_ms":1385.21,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":30,"total_tokens":1775,"cost":0.0017545275,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00177225,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":4.47e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":1504.49,"response_complete":true,"answer_first_audio_ms":2146.35,"tts_first_audio_ms":759.83,"first_audio_ms":2146.35,"tts_server_stages":[{"generation_ms":1076,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1242,"performance":{"device_seconds":0.86781423664093,"raw_audio_seconds":3.04,"aggregate_rtf":0.28546520942135856,"audio_seconds_per_second":3.5030538468313255,"batches":[{"batch_size":1,"device_seconds":0.86781423664093,"audio_seconds":[3.04],"model_wall_seconds":0.960164837539196,"first_audio_seconds":0.46063922119140627,"graph_capture_seconds":0,"live_callback_seconds":0.09010444581508636}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":976,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1131,"performance":{"device_seconds":0.7708185715675354,"raw_audio_seconds":2.56,"aggregate_rtf":0.3011010045185685,"audio_seconds_per_second":3.3211446823264628,"batches":[{"batch_size":1,"device_seconds":0.7708185715675354,"audio_seconds":[2.56],"model_wall_seconds":0.8650579005479813,"first_audio_seconds":0.4558837890625,"graph_capture_seconds":0,"live_callback_seconds":0.09119796007871628}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1251.04,"audio_s":2.19},{"sentence":1,"total_ms":1136.85,"audio_s":2.51}],"sentence_ready_ms":3775.98,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.warm.soft"},"sources":[],"lookups":[],"at":11253.29999923706},{"type":"latency_update","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1384.69,"llm_first_sentence_ms":1385.21,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":30,"total_tokens":1775,"cost":0.0017545275,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00177225,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":4.47e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":1504.49,"response_complete":true,"answer_first_audio_ms":2146.35,"tts_first_audio_ms":759.83,"first_audio_ms":2146.35,"tts_server_stages":[{"generation_ms":1076,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1242,"performance":{"device_seconds":0.86781423664093,"raw_audio_seconds":3.04,"aggregate_rtf":0.28546520942135856,"audio_seconds_per_second":3.5030538468313255,"batches":[{"batch_size":1,"device_seconds":0.86781423664093,"audio_seconds":[3.04],"model_wall_seconds":0.960164837539196,"first_audio_seconds":0.46063922119140627,"graph_capture_seconds":0,"live_callback_seconds":0.09010444581508636}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":976,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1131,"performance":{"device_seconds":0.7708185715675354,"raw_audio_seconds":2.56,"aggregate_rtf":0.3011010045185685,"audio_seconds_per_second":3.3211446823264628,"batches":[{"batch_size":1,"device_seconds":0.7708185715675354,"audio_seconds":[2.56],"model_wall_seconds":0.8650579005479813,"first_audio_seconds":0.4558837890625,"graph_capture_seconds":0,"live_callback_seconds":0.09119796007871628}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1251.04,"audio_s":2.19},{"sentence":1,"total_ms":1136.85,"audio_s":2.51}]},"at":12008.20000076294},{"type":"pipeline_metrics","reply_id":1,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":1384.69,"llm_first_sentence_ms":1385.21,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":30,"total_tokens":1775,"cost":0.0017545275,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00177225,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":4.47e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":1504.49,"response_complete":true,"answer_first_audio_ms":2146.35,"tts_first_audio_ms":759.83,"first_audio_ms":2146.35,"tts_server_stages":[{"generation_ms":1076,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1242,"performance":{"device_seconds":0.86781423664093,"raw_audio_seconds":3.04,"aggregate_rtf":0.28546520942135856,"audio_seconds_per_second":3.5030538468313255,"batches":[{"batch_size":1,"device_seconds":0.86781423664093,"audio_seconds":[3.04],"model_wall_seconds":0.960164837539196,"first_audio_seconds":0.46063922119140627,"graph_capture_seconds":0,"live_callback_seconds":0.09010444581508636}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":976,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1131,"performance":{"device_seconds":0.7708185715675354,"raw_audio_seconds":2.56,"aggregate_rtf":0.3011010045185685,"audio_seconds_per_second":3.3211446823264628,"batches":[{"batch_size":1,"device_seconds":0.7708185715675354,"audio_seconds":[2.56],"model_wall_seconds":0.8650579005479813,"first_audio_seconds":0.4558837890625,"graph_capture_seconds":0,"live_callback_seconds":0.09119796007871628}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1046,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1199,"performance":{"device_seconds":0.8393501958847046,"raw_audio_seconds":2.88,"aggregate_rtf":0.29144104023774464,"audio_seconds_per_second":3.43122574358177,"batches":[{"batch_size":1,"device_seconds":0.8393501958847046,"audio_seconds":[2.88],"model_wall_seconds":0.938080184161663,"first_audio_seconds":0.46228173828125,"graph_capture_seconds":0,"live_callback_seconds":0.0966586172580719}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1251.04,"audio_s":2.19},{"sentence":1,"total_ms":1136.85,"audio_s":2.51},{"sentence":2,"total_ms":1211.72,"audio_s":2.83}],"total_ms":4987.92,"audio_s":7.53,"sentences":3,"generation_rtf":0.6624},"at":12469.20000076294},{"type":"turn_done","barged":false,"at":17611.70000076294},{"type":"voice_caption","status":"available","caption":"Possibly very fading, very smooth, slow, very liberation","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":1.62,"inference_ms":37.76,"model":"laion/humaneness-ears-base-medium","model_revision":"818506970d93c809a3295a0a02c40dee8ff2bfc3","rules_revision":"7ebd4a9b7da96df989af8fc6ad0ce34f4e8aa3ed","device":"cuda:1","precision":"bfloat16","selected":{"voicenet":[{"dim":"DARC","name":"Dynamic Arc","value":2.357,"z":-1.67,"direction":"below","intensity":"Very","always_on":false,"phrase":"very fading in loudness","tag":"very fading","reliability":1,"rank_score":1.671},{"dim":"DFLU","name":"Disfluency","value":-0.258,"z":-1.56,"direction":"below","intensity":"Very","always_on":false,"phrase":"very smooth and flawless","tag":"very smooth","reliability":1,"rank_score":1.563},{"dim":"TEMP","name":"Tempo","value":2.38,"z":-1.11,"direction":"below","intensity":"Notably","always_on":true,"phrase":"notably slow and deliberate in tempo","tag":"slow","reliability":1,"rank_score":1.112}],"emotions":[{"emotion":"Relief","value":1.929,"z":3.7,"direction":"above","intensity":"Extremely","synonym":"liberation","phrase":"extremely carrying liberation","tag":"very liberation"}]},"normalization":"speech reference z-scores; not probabilities","scope":"this turn","request_ms":50.39,"added_wait_ms":0,"at":17612.5},{"type":"endpoint","transcript":"Schön, dich kennenzulernen.","speaker_id":"speaker-1","asr":{"text":"Schön, dich kennenzulernen.","model":"parakeet_v3","chunks":1,"transcribe_ms":50.67,"speech_pause_ms":700,"observed_quiet_ms":598},"speaker_name":"Christoph","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":0.553,"speaker_embedding_wait_ms":0.002,"speaker_lookup_ms":0.55,"speaker_prefetch_used":true,"speaker_prefetch_compute_ms":47.62,"sentence_queue_ms":8284.39,"sentence_audio_s":7.04},"at":17612.59999847412},{"type":"latency_start","reply_id":2,"input":{"kind":"voice","asr":{"text":"Schön, dich kennenzulernen.","model":"parakeet_v3","chunks":1,"transcribe_ms":50.67,"speech_pause_ms":700,"observed_quiet_ms":598},"voice_caption_ms":50.39,"voice_caption_status":"available","speaker_assignment_ms":0.553,"speaker_prefetch_compute_ms":47.62,"speaker_prefetch_used":true,"sentence_queue_ms":8284.39},"at":17613.20000076294},{"type":"reply","text":"Mir auch, Christoph!","reply_id":2,"performance":{"sentences":[{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, warm","text":"Mir auch, Christoph!","start_s":0,"end_s":1.02}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Mir auch, Christoph!","sentence_role":"answer","speaker_id":"speaker-1","speaker_name":"Christoph","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":299.56,"finish_reason":"stop","usage":{"prompt_tokens":1821,"completion_tokens":11,"total_tokens":1832,"cost":0.0018009882,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00181918,"upstream_inference_prompt_cost":0.00180279,"upstream_inference_completions_cost":1.639e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":349.7,"generation_calls":1,"llm_total_ms":349.72,"response_complete":true,"sentence_ready_ms":350.78,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"elation.soft"},"sources":[],"lookups":[],"at":17972.09999847412},{"type":"latency_update","reply_id":2,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":299.56,"finish_reason":"stop","usage":{"prompt_tokens":1821,"completion_tokens":11,"total_tokens":1832,"cost":0.0018009882,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00181918,"upstream_inference_prompt_cost":0.00180279,"upstream_inference_completions_cost":1.639e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":349.7,"generation_calls":1,"llm_total_ms":349.72,"response_complete":true,"answer_first_audio_ms":1107.63,"tts_first_audio_ms":756.71,"first_audio_ms":1107.64},"at":18707.09999847412},{"type":"pipeline_metrics","reply_id":2,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":299.56,"finish_reason":"stop","usage":{"prompt_tokens":1821,"completion_tokens":11,"total_tokens":1832,"cost":0.0018009882,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00181918,"upstream_inference_prompt_cost":0.00180279,"upstream_inference_completions_cost":1.639e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":349.7,"generation_calls":1,"llm_total_ms":349.72,"response_complete":true,"answer_first_audio_ms":1107.63,"tts_first_audio_ms":756.71,"first_audio_ms":1107.64,"tts_server_stages":[{"generation_ms":1058,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1226,"performance":{"device_seconds":0.8582469816207886,"raw_audio_seconds":3.04,"aggregate_rtf":0.28231808605946995,"audio_seconds_per_second":3.5421039224151984,"batches":[{"batch_size":1,"device_seconds":0.8582469816207886,"audio_seconds":[3.04],"model_wall_seconds":0.9435448423027992,"first_audio_seconds":0.4578456115722656,"graph_capture_seconds":0,"live_callback_seconds":0.08505809307098389}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1234.26,"audio_s":2.51}],"total_ms":1585.2,"audio_s":2.51,"sentences":1,"generation_rtf":0.6316},"at":19204.40000152588},{"type":"turn_done","barged":false,"at":21646.59999847412}],"reply":"Mir auch, Christoph!","captions":[{"status":"available","caption":"The audio contains a single male voice speaking in German. The speaker's tone is conversational and relatively neutral, with a moderate pace.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.12,"tokens":28,"preprocess_ms":9.42,"encoder_gpu_ms":16.44,"decoder_gpu_ms":257.39,"inference_ms":274.04,"total_ms":283.46,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":29.37,"window_s":2.12,"candidates":[{"label":"Speech","score":0.8537},{"label":"Speech synthesizer","score":0.4532},{"label":"Narration, monologue","score":0.1053},{"label":"Music","score":0.0936},{"label":"Clicking","score":0.0243},{"label":"Animal","score":0.0238}],"music_score":0.0936,"speech_score":0.8537,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.7348},{"label":"Speech synthesizer","score":0.1177},{"label":"Music","score":0.0936}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.8537},{"label":"Speech synthesizer","score":0.3911},{"label":"Narration, monologue","score":0.1053}]},{"start_s":0.84,"end_s":2.12,"candidates":[{"label":"Speech","score":0.7871},{"label":"Speech synthesizer","score":0.3879},{"label":"Narration, monologue","score":0.0928}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":38.18},"type":"sound_caption","request_ms":301.34,"capture_ms":1920,"event_age_ms":2221.34,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":7096}],"pcm_chunks":14,"page_errors":[],"passed":true,"speakers":["Christoph","Christoph"]},{"case":"selma-followup-clear","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":35121.70000076294},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":35612.09999847412},{"type":"asr_preview","text":"Was können wir heute","status":"sentence_stream","source":"local_sentence_asr","at":36120.79999923706},{"type":"asr_preview","text":"Was können wir heute zusammen?","status":"sentence_stream","source":"native_asr","at":36686.40000152588},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":37444.29999923706},{"status":"available","caption":"A female voice speaks in German. The speaker is asking a question, and the tone is neutral and conversational.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.42,"tokens":24,"preprocess_ms":7.15,"encoder_gpu_ms":8.69,"decoder_gpu_ms":212.35,"inference_ms":221.31,"total_ms":228.46,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":27.75,"window_s":2.42,"candidates":[{"label":"Speech","score":0.714},{"label":"Speech synthesizer","score":0.6872},{"label":"Narration, monologue","score":0.3675},{"label":"Conversation","score":0.0382},{"label":"Inside, small room","score":0.0299},{"label":"Music","score":0.0272}],"music_score":0.0272,"speech_score":0.714,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.6551},{"label":"Speech synthesizer","score":0.302},{"label":"Narration, monologue","score":0.2678}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.714},{"label":"Narration, monologue","score":0.3675},{"label":"Speech synthesizer","score":0.2021}]},{"start_s":1.14,"end_s":2.42,"candidates":[{"label":"Speech","score":0.6996},{"label":"Speech synthesizer","score":0.3903},{"label":"Narration, monologue","score":0.1741}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.75},"type":"sound_caption","request_ms":246.52,"capture_ms":2220,"event_age_ms":2466.53,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":37465.09999847412},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"native_asr","at":37657.79999923706},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":37687.29999923706},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":37742.70000076294},{"type":"asr_sentence","text":"Was können wir heute zusammen machen?","audio_s":34.88,"queued":false,"asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":46.6,"speech_pause_ms":700,"observed_quiet_ms":654},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":654,"at":37742.90000152588},{"type":"identity","enabled":false,"mode":"open-sentences","sentence_mode":true,"open_enrollment":true,"speakers":[{"id":"speaker-1","name":"Christoph","named":true,"reference_ready":true,"voice_variants":1,"admitted":true},{"id":"speaker-2","name":"Speaker 2","named":false,"reference_ready":true,"voice_variants":1,"admitted":true}],"users":["Christoph","Speaker 2"],"threshold":0.38,"conversation_focus":{"enabled":false,"threshold":0.28,"timeout_s":180,"speaker_id":null,"name":null,"primary_id":"speaker-1","participants":[{"id":"speaker-1","name":"Christoph"},{"id":"speaker-2","name":"Speaker 2"}]},"handover_pending":false,"storage":"Session RAM","prediction":{"name":"Speaker 2","similarity":0.049},"at":37802.09999847412},{"type":"voice_caption","status":"available","caption":"Possibly smooth, slightly steady, very hope","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":2.62,"inference_ms":23.3,"model":"laion/humaneness-ears-base-medium","model_revision":"818506970d93c809a3295a0a02c40dee8ff2bfc3","rules_revision":"7ebd4a9b7da96df989af8fc6ad0ce34f4e8aa3ed","device":"cuda:1","precision":"bfloat16","selected":{"voicenet":[{"dim":"DFLU","name":"Disfluency","value":0.259,"z":-1.04,"direction":"below","intensity":"Notably","always_on":false,"phrase":"notably smooth and flawless","tag":"smooth","reliability":1,"rank_score":1.039},{"dim":"VOLT","name":"Volatility","value":1.3,"z":-0.79,"direction":"below","intensity":"Somewhat","always_on":false,"phrase":"somewhat emotionally steady and static","tag":"slightly steady","reliability":1,"rank_score":0.79},{"dim":"TEMP","name":"Tempo","value":2.847,"z":-0.22,"direction":"below","intensity":"About-average","always_on":true,"phrase":"moderate in tempo","tag":null,"reliability":1,"rank_score":0.218}],"emotions":[{"emotion":"Hope/Optimism","value":2.198,"z":2.06,"direction":"above","intensity":"Extremely","synonym":"hope","phrase":"extremely carrying hope","tag":"very hope"}]},"normalization":"speech reference z-scores; not probabilities","scope":"this turn","request_ms":35.58,"added_wait_ms":0,"at":37804.09999847412},{"type":"endpoint","transcript":"Was können wir heute zusammen machen?","speaker_id":"speaker-2","asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":46.6,"speech_pause_ms":700,"observed_quiet_ms":654},"speaker_name":"Speaker 2","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":57.362,"speaker_embedding_wait_ms":56.694,"speaker_lookup_ms":0.668,"speaker_prefetch_used":false,"speaker_prefetch_compute_ms":null,"sentence_queue_ms":57.66,"sentence_audio_s":34.88},"at":37804.20000076294},{"type":"latency_start","reply_id":3,"input":{"kind":"voice","asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":46.6,"speech_pause_ms":700,"observed_quiet_ms":654},"voice_caption_ms":35.58,"voice_caption_status":"available","speaker_assignment_ms":57.362,"speaker_prefetch_compute_ms":null,"speaker_prefetch_used":false,"sentence_queue_ms":57.65},"at":37804.5},{"type":"reply","text":"Wie wär’s mit einer Mini-Geschichten-Challenge: Du gibst mir ein Wort, ich erfinde eine verrückte Kurzgeschichte, und du drehst das Ende um?","reply_id":3,"performance":{"sentences":[{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Wie wär’s mit einer Mini-Geschichten-Challenge: Du gibst mir ein Wort, ich erfinde eine verrückte Kurzgeschichte, und du drehst das Ende um?","start_s":0,"end_s":7.14}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Wie wär’s mit einer Mini-Geschichten-Challenge: Du gibst mir ein Wort, ich erfinde eine verrückte Kurzgeschichte, und du drehst das Ende um?","sentence_role":"answer","speaker_id":"speaker-2","speaker_name":"Speaker 2","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":518.64,"finish_reason":"stop","usage":{"prompt_tokens":1953,"completion_tokens":46,"total_tokens":1999,"cost":0.0019819899,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200201,"upstream_inference_prompt_cost":0.00193347,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":624.59,"generation_calls":1,"name_question_added":true,"llm_total_ms":624.65,"response_complete":true,"sentence_ready_ms":625.77,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":38433.5},{"type":"latency_update","reply_id":3,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":518.64,"finish_reason":"stop","usage":{"prompt_tokens":1953,"completion_tokens":46,"total_tokens":1999,"cost":0.0019819899,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200201,"upstream_inference_prompt_cost":0.00193347,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":624.59,"generation_calls":1,"name_question_added":true,"llm_total_ms":624.65,"response_complete":true,"answer_first_audio_ms":1388.95,"tts_first_audio_ms":763.03,"first_audio_ms":1388.95},"at":39191.79999923706},{"type":"reply","text":"Wie wär’s mit einer Mini-Geschichten-Challenge: Du gibst mir ein Wort, ich erfinde eine verrückte Kurzgeschichte, und du drehst das Ende um? Und wie darf ich dich nennen?","reply_id":3,"performance":{"sentences":[{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Wie wär’s mit einer Mini-Geschichten-Challenge: Du gibst mir ein Wort, ich erfinde eine verrückte Kurzgeschichte, und du drehst das Ende um?","start_s":0,"end_s":8.43},{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.55,"direction":"interest, soft, warm","text":"Und wie darf ich dich nennen?","start_s":8.43,"end_s":10.469999999999999}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Und wie darf ich dich nennen?","sentence_role":"answer","speaker_id":"speaker-2","speaker_name":"Speaker 2","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":518.64,"finish_reason":"stop","usage":{"prompt_tokens":1953,"completion_tokens":46,"total_tokens":1999,"cost":0.0019819899,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200201,"upstream_inference_prompt_cost":0.00193347,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":624.59,"generation_calls":1,"name_question_added":true,"llm_total_ms":624.65,"response_complete":true,"answer_first_audio_ms":1388.95,"tts_first_audio_ms":763.03,"first_audio_ms":1388.95,"tts_server_stages":[{"generation_ms":2992,"alignment_ms":40,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":3148,"performance":{"device_seconds":2.420494098186493,"raw_audio_seconds":8.48,"aggregate_rtf":0.28543562478614304,"audio_seconds_per_second":3.503416928945818,"batches":[{"batch_size":1,"device_seconds":2.420494098186493,"audio_seconds":[8.48],"model_wall_seconds":2.879118360579014,"first_audio_seconds":0.4776857604980469,"graph_capture_seconds":0,"live_callback_seconds":0.4525199830532074}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":3163.5,"audio_s":8.43}],"sentence_ready_ms":3790.24,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":41607.59999847412},{"type":"latency_update","reply_id":3,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":518.64,"finish_reason":"stop","usage":{"prompt_tokens":1953,"completion_tokens":46,"total_tokens":1999,"cost":0.0019819899,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200201,"upstream_inference_prompt_cost":0.00193347,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":624.59,"generation_calls":1,"name_question_added":true,"llm_total_ms":624.65,"response_complete":true,"answer_first_audio_ms":1388.95,"tts_first_audio_ms":763.03,"first_audio_ms":1388.95,"tts_server_stages":[{"generation_ms":2992,"alignment_ms":40,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":3148,"performance":{"device_seconds":2.420494098186493,"raw_audio_seconds":8.48,"aggregate_rtf":0.28543562478614304,"audio_seconds_per_second":3.503416928945818,"batches":[{"batch_size":1,"device_seconds":2.420494098186493,"audio_seconds":[8.48],"model_wall_seconds":2.879118360579014,"first_audio_seconds":0.4776857604980469,"graph_capture_seconds":0,"live_callback_seconds":0.4525199830532074}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":3163.5,"audio_s":8.43}]},"at":42339.70000076294},{"type":"pipeline_metrics","reply_id":3,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":518.64,"finish_reason":"stop","usage":{"prompt_tokens":1953,"completion_tokens":46,"total_tokens":1999,"cost":0.0019819899,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200201,"upstream_inference_prompt_cost":0.00193347,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":624.59,"generation_calls":1,"name_question_added":true,"llm_total_ms":624.65,"response_complete":true,"answer_first_audio_ms":1388.95,"tts_first_audio_ms":763.03,"first_audio_ms":1388.95,"tts_server_stages":[{"generation_ms":2992,"alignment_ms":40,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":3148,"performance":{"device_seconds":2.420494098186493,"raw_audio_seconds":8.48,"aggregate_rtf":0.28543562478614304,"audio_seconds_per_second":3.503416928945818,"batches":[{"batch_size":1,"device_seconds":2.420494098186493,"audio_seconds":[8.48],"model_wall_seconds":2.879118360579014,"first_audio_seconds":0.4776857604980469,"graph_capture_seconds":0,"live_callback_seconds":0.4525199830532074}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1013,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1168,"performance":{"device_seconds":0.8175622749328613,"raw_audio_seconds":2.8,"aggregate_rtf":0.2919865267617362,"audio_seconds_per_second":3.4248155594384997,"batches":[{"batch_size":1,"device_seconds":0.8175622749328613,"audio_seconds":[2.8],"model_wall_seconds":0.9032638818025589,"first_audio_seconds":0.4623276672363281,"graph_capture_seconds":0,"live_callback_seconds":0.08545474708080292}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":3163.5,"audio_s":8.43},{"sentence":1,"total_ms":1176.27,"audio_s":2.11}],"total_ms":4966.66,"audio_s":10.54,"sentences":2,"generation_rtf":0.4712},"at":42769.90000152588},{"type":"turn_done","barged":false,"at":50172.90000152588}],"reply":"Wie wär’s mit einer Mini-Geschichten-Challenge: Du gibst mir ein Wort, ich erfinde eine verrückte Kurzgeschichte, und du drehst das Ende um? Und wie darf ich dich nennen?","captions":[{"status":"available","caption":"A female voice speaks in German. The speaker is asking a question, and the tone is neutral and conversational.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.42,"tokens":24,"preprocess_ms":7.15,"encoder_gpu_ms":8.69,"decoder_gpu_ms":212.35,"inference_ms":221.31,"total_ms":228.46,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":27.75,"window_s":2.42,"candidates":[{"label":"Speech","score":0.714},{"label":"Speech synthesizer","score":0.6872},{"label":"Narration, monologue","score":0.3675},{"label":"Conversation","score":0.0382},{"label":"Inside, small room","score":0.0299},{"label":"Music","score":0.0272}],"music_score":0.0272,"speech_score":0.714,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.6551},{"label":"Speech synthesizer","score":0.302},{"label":"Narration, monologue","score":0.2678}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.714},{"label":"Narration, monologue","score":0.3675},{"label":"Speech synthesizer","score":0.2021}]},{"start_s":1.14,"end_s":2.42,"candidates":[{"label":"Speech","score":0.6996},{"label":"Speech synthesizer","score":0.3903},{"label":"Narration, monologue","score":0.1741}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":36.75},"type":"sound_caption","request_ms":246.52,"capture_ms":2220,"event_age_ms":2466.53,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":37465.09999847412}],"pcm_chunks":11,"page_errors":[],"passed":true,"speakers":["Speaker 2"]},{"case":"selma-intro","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":52392.59999847412},{"type":"asr_preview","text":"Hey Ra.","status":"sentence_stream","source":"local_sentence_asr","at":52887.09999847412},{"type":"asr_preview","text":"Hey Robbie.","status":"sentence_stream","source":"local_sentence_asr","at":53498.5},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":53772},{"type":"asr_preview","text":"Hey Robbie, you","status":"sentence_stream","source":"local_sentence_asr","at":53840.40000152588},{"type":"asr_preview","text":"Hey Robi.","status":"sentence_stream","source":"native_asr","at":54058},{"type":"asr_preview","text":"Hey Robbie, ich bin Selma.","status":"sentence_stream","source":"local_sentence_asr","at":54788.70000076294},{"type":"asr_preview","text":"Hey Robbie, ich bin Selma.","status":"sentence_stream","source":"native_asr","at":55046.40000152588},{"status":"available","caption":"A female voice, perceived as adult, speaks in a calm and clear tone. The speech is delivered at a moderate pace, with distinct pronunciation.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.86,"tokens":30,"preprocess_ms":10.22,"encoder_gpu_ms":14.06,"decoder_gpu_ms":290.41,"inference_ms":304.71,"total_ms":314.94,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":28.93,"window_s":2.86,"candidates":[{"label":"Speech","score":0.8348},{"label":"Whispering","score":0.2783},{"label":"Narration, monologue","score":0.1524},{"label":"Inside, small room","score":0.0547},{"label":"Clicking","score":0.0541},{"label":"Mantra","score":0.0539}],"music_score":0.0407,"speech_score":0.8348,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.3071},{"label":"Mantra","score":0.0539},{"label":"Music","score":0.0407}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.6407},{"label":"Narration, monologue","score":0.1257},{"label":"Whispering","score":0.0563}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Speech","score":0.8109},{"label":"Whispering","score":0.2783},{"label":"Inside, small room","score":0.0547}]},{"start_s":1.58,"end_s":2.86,"candidates":[{"label":"Speech","score":0.8348},{"label":"Narration, monologue","score":0.1524},{"label":"Music","score":0.0294}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":37.49},"type":"sound_caption","request_ms":332.73,"capture_ms":2660,"event_age_ms":2992.74,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":55226.29999923706},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":55293.79999923706},{"type":"asr_preview","text":"Hey Robbie, ich bin Selma.","status":"sentence_stream","source":"local_sentence_asr","at":55354.09999847412},{"type":"asr_sentence","text":"Hey Robbie, ich bin Selma.","audio_s":51.84,"queued":false,"asr":{"text":"Hey Robbie, ich bin Selma.","model":"parakeet_v3","chunks":1,"transcribe_ms":45.82,"speech_pause_ms":700,"observed_quiet_ms":594},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":594,"at":55354.09999847412},{"type":"identity","enabled":false,"mode":"open-sentences","sentence_mode":true,"open_enrollment":true,"speakers":[{"id":"speaker-1","name":"Christoph","named":true,"reference_ready":true,"voice_variants":1,"admitted":true},{"id":"speaker-2","name":"Selma","named":true,"reference_ready":true,"voice_variants":2,"admitted":true}],"users":["Christoph","Selma"],"threshold":0.38,"conversation_focus":{"enabled":false,"threshold":0.28,"timeout_s":180,"speaker_id":null,"name":null,"primary_id":"speaker-1","participants":[{"id":"speaker-1","name":"Christoph"},{"id":"speaker-2","name":"Selma"}]},"handover_pending":false,"storage":"Session RAM","prediction":{"name":"Selma","similarity":0.3162},"at":55423},{"type":"voice_caption","status":"available","caption":"Possibly very soft-onset, very fading, slow, liberation","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":2.98,"inference_ms":21.97,"model":"laion/humaneness-ears-base-medium","model_revision":"818506970d93c809a3295a0a02c40dee8ff2bfc3","rules_revision":"7ebd4a9b7da96df989af8fc6ad0ce34f4e8aa3ed","device":"cuda:1","precision":"bfloat16","selected":{"voicenet":[{"dim":"ATCK","name":"Attack","value":1.535,"z":-2.52,"direction":"below","intensity":"Extremely","always_on":false,"phrase":"extremely soft and gentle in onset","tag":"very soft-onset","reliability":1,"rank_score":2.523},{"dim":"DARC","name":"Dynamic Arc","value":2.137,"z":-2.28,"direction":"below","intensity":"Extremely","always_on":false,"phrase":"extremely fading in loudness","tag":"very fading","reliability":1,"rank_score":2.281},{"dim":"TEMP","name":"Tempo","value":2.341,"z":-1.19,"direction":"below","intensity":"Notably","always_on":true,"phrase":"notably slow and deliberate in tempo","tag":"slow","reliability":1,"rank_score":1.186}],"emotions":[{"emotion":"Relief","value":0.904,"z":1.43,"direction":"above","intensity":"Notably","synonym":"liberation","phrase":"notably carrying liberation","tag":"liberation"}]},"normalization":"speech reference z-scores; not probabilities","scope":"this turn","request_ms":35.95,"added_wait_ms":0,"at":55424.79999923706},{"type":"endpoint","transcript":"Hey Robbie, ich bin Selma.","speaker_id":"speaker-2","asr":{"text":"Hey Robbie, ich bin Selma.","model":"parakeet_v3","chunks":1,"transcribe_ms":45.82,"speech_pause_ms":700,"observed_quiet_ms":594},"speaker_name":"Selma","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":63.784,"speaker_embedding_wait_ms":63.305,"speaker_lookup_ms":0.479,"speaker_prefetch_used":false,"speaker_prefetch_compute_ms":null,"sentence_queue_ms":63.96,"sentence_audio_s":51.84},"at":55424.90000152588},{"type":"latency_start","reply_id":4,"input":{"kind":"voice","asr":{"text":"Hey Robbie, ich bin Selma.","model":"parakeet_v3","chunks":1,"transcribe_ms":45.82,"speech_pause_ms":700,"observed_quiet_ms":594},"voice_caption_ms":35.95,"voice_caption_status":"available","speaker_assignment_ms":63.784,"speaker_prefetch_compute_ms":null,"speaker_prefetch_used":false,"sentence_queue_ms":63.95},"at":55425.29999923706},{"type":"reply","text":"Hallo Selma!","reply_id":4,"performance":{"sentences":[{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Hallo Selma!","start_s":0,"end_s":0.68}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Hallo Selma!","sentence_role":"answer","speaker_id":"speaker-2","speaker_name":"Selma","speaker_update":{"speaker_id":"speaker-2","display_name":"Selma","is_self_introduction":true,"source":"validated_acoustic_transcript"},"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":297.3,"llm_first_sentence_ms":320.99,"sentence_ready_ms":322.11,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"elation.soft"},"sources":[],"lookups":[],"at":55744.70000076294},{"type":"asr_preview","text":"Schön.","status":"sentence_stream","source":"local_sentence_asr","at":55838.79999923706},{"type":"asr_preview","text":"schön, dass du dabei","status":"sentence_stream","source":"local_sentence_asr","at":56342.40000152588},{"type":"latency_update","reply_id":4,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":297.3,"llm_first_sentence_ms":320.99,"finish_reason":"stop","usage":{"prompt_tokens":1992,"completion_tokens":25,"total_tokens":2017,"cost":0.0019892367,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200933,"upstream_inference_prompt_cost":0.00197208,"upstream_inference_completions_cost":3.725e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":347.07,"response_complete":true,"answer_first_audio_ms":1077.75,"tts_first_audio_ms":755.5,"first_audio_ms":1077.75},"at":56498.20000076294},{"type":"reply","text":"Hallo Selma! Schön, dass du da bist – jetzt sind wir ein richtiges Trio!","reply_id":4,"performance":{"sentences":[{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Hallo Selma!","start_s":0,"end_s":1.23},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Schön, dass du da bist – jetzt sind wir ein richtiges Trio!","start_s":1.23,"end_s":5.3100000000000005}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Schön, dass du da bist – jetzt sind wir ein richtiges Trio!","sentence_role":"answer","speaker_id":"speaker-2","speaker_name":"Selma","speaker_update":{"speaker_id":"speaker-2","display_name":"Selma","is_self_introduction":true,"source":"validated_acoustic_transcript"},"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":297.3,"llm_first_sentence_ms":320.99,"finish_reason":"stop","usage":{"prompt_tokens":1992,"completion_tokens":25,"total_tokens":2017,"cost":0.0019892367,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200933,"upstream_inference_prompt_cost":0.00197208,"upstream_inference_completions_cost":3.725e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":347.07,"response_complete":true,"answer_first_audio_ms":1077.75,"tts_first_audio_ms":755.5,"first_audio_ms":1077.75,"tts_server_stages":[{"generation_ms":622,"alignment_ms":35,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":785,"performance":{"device_seconds":0.4651031546592712,"raw_audio_seconds":1.6,"aggregate_rtf":0.2906894716620445,"audio_seconds_per_second":3.4400970708791263,"batches":[{"batch_size":1,"device_seconds":0.4651031546592712,"audio_seconds":[1.6],"model_wall_seconds":0.509982518851757,"first_audio_seconds":0.45872967529296876,"graph_capture_seconds":0,"live_callback_seconds":0.04469561576843262}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":793.6,"audio_s":1.23}],"sentence_ready_ms":1116.37,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"elation.soft"},"sources":[],"lookups":[],"at":56542},{"type":"asr_preview","text":"schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":56835.90000152588},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":57105.59999847412},{"type":"asr_preview","text":"Schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":57172.40000152588},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":57271.70000076294},{"type":"latency_update","reply_id":4,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":297.3,"llm_first_sentence_ms":320.99,"finish_reason":"stop","usage":{"prompt_tokens":1992,"completion_tokens":25,"total_tokens":2017,"cost":0.0019892367,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200933,"upstream_inference_prompt_cost":0.00197208,"upstream_inference_completions_cost":3.725e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":347.07,"response_complete":true,"answer_first_audio_ms":1077.75,"tts_first_audio_ms":755.5,"first_audio_ms":1077.75,"tts_server_stages":[{"generation_ms":622,"alignment_ms":35,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":785,"performance":{"device_seconds":0.4651031546592712,"raw_audio_seconds":1.6,"aggregate_rtf":0.2906894716620445,"audio_seconds_per_second":3.4400970708791263,"batches":[{"batch_size":1,"device_seconds":0.4651031546592712,"audio_seconds":[1.6],"model_wall_seconds":0.509982518851757,"first_audio_seconds":0.45872967529296876,"graph_capture_seconds":0,"live_callback_seconds":0.04469561576843262}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":793.6,"audio_s":1.23}]},"at":57308},{"type":"asr_preview","text":"Schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":57339.09999847412},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":57469.79999923706},{"type":"asr_preview","text":"Schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":57544},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":57630.79999923706},{"type":"asr_preview","text":"Schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":57705.5},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":57841.5},{"type":"asr_preview","text":"Schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":57886.70000076294},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":58004.90000152588},{"type":"asr_preview","text":"Schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":58039.09999847412},{"type":"pipeline_metrics","reply_id":4,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":297.3,"llm_first_sentence_ms":320.99,"finish_reason":"stop","usage":{"prompt_tokens":1992,"completion_tokens":25,"total_tokens":2017,"cost":0.0019892367,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00200933,"upstream_inference_prompt_cost":0.00197208,"upstream_inference_completions_cost":3.725e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":347.07,"response_complete":true,"answer_first_audio_ms":1077.75,"tts_first_audio_ms":755.5,"first_audio_ms":1077.75,"tts_server_stages":[{"generation_ms":622,"alignment_ms":35,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":785,"performance":{"device_seconds":0.4651031546592712,"raw_audio_seconds":1.6,"aggregate_rtf":0.2906894716620445,"audio_seconds_per_second":3.4400970708791263,"batches":[{"batch_size":1,"device_seconds":0.4651031546592712,"audio_seconds":[1.6],"model_wall_seconds":0.509982518851757,"first_audio_seconds":0.45872967529296876,"graph_capture_seconds":0,"live_callback_seconds":0.04469561576843262}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":1628,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1780,"performance":{"device_seconds":1.3603325991630555,"raw_audio_seconds":4.72,"aggregate_rtf":0.28820605914471514,"audio_seconds_per_second":3.4697396819748194,"batches":[{"batch_size":1,"device_seconds":1.3603325991630555,"audio_seconds":[4.72],"model_wall_seconds":1.5200840756297112,"first_audio_seconds":0.4727408447265625,"graph_capture_seconds":0,"live_callback_seconds":0.15416038036346436}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":793.6,"audio_s":1.23},{"sentence":1,"total_ms":1793.05,"audio_s":4.72}],"total_ms":2909.53,"audio_s":5.95,"sentences":2,"generation_rtf":0.489},"at":58338.09999847412},{"type":"turn_done","barged":false,"at":62802.59999847412}],"reply":"Hallo Selma! Schön, dass du da bist – jetzt sind wir ein richtiges Trio!","captions":[{"status":"available","caption":"A female voice, perceived as adult, speaks in a calm and clear tone. The speech is delivered at a moderate pace, with distinct pronunciation.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.86,"tokens":30,"preprocess_ms":10.22,"encoder_gpu_ms":14.06,"decoder_gpu_ms":290.41,"inference_ms":304.71,"total_ms":314.94,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":28.93,"window_s":2.86,"candidates":[{"label":"Speech","score":0.8348},{"label":"Whispering","score":0.2783},{"label":"Narration, monologue","score":0.1524},{"label":"Inside, small room","score":0.0547},{"label":"Clicking","score":0.0541},{"label":"Mantra","score":0.0539}],"music_score":0.0407,"speech_score":0.8348,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.3071},{"label":"Mantra","score":0.0539},{"label":"Music","score":0.0407}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.6407},{"label":"Narration, monologue","score":0.1257},{"label":"Whispering","score":0.0563}]},{"start_s":1.28,"end_s":2.56,"candidates":[{"label":"Speech","score":0.8109},{"label":"Whispering","score":0.2783},{"label":"Inside, small room","score":0.0547}]},{"start_s":1.58,"end_s":2.86,"candidates":[{"label":"Speech","score":0.8348},{"label":"Narration, monologue","score":0.1524},{"label":"Music","score":0.0294}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":37.49},"type":"sound_caption","request_ms":332.73,"capture_ms":2660,"event_age_ms":2992.74,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":55226.29999923706}],"pcm_chunks":7,"page_errors":[],"passed":true,"speakers":["Selma"]},{"case":"selma-followup-clear","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":66357.90000152588},{"type":"asr_preview","text":"Was können wir","status":"sentence_stream","source":"local_sentence_asr","at":66904.29999923706},{"type":"asr_preview","text":"Was können können","status":"sentence_stream","source":"native_asr","at":67121.70000076294},{"type":"asr_preview","text":"Was können wir heute zusammen?","status":"sentence_stream","source":"local_sentence_asr","at":67843.20000076294},{"type":"asr_preview","text":"Was können wir heute zusammen?","status":"sentence_stream","source":"native_asr","at":68074.90000152588},{"status":"available","caption":"A female voice speaks in German. The speaker is asking a question, and the tone is neutral and conversational.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.42,"tokens":24,"preprocess_ms":7.95,"encoder_gpu_ms":8.98,"decoder_gpu_ms":172.68,"inference_ms":181.9,"total_ms":189.85,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":27.09,"window_s":2.42,"candidates":[{"label":"Speech","score":0.714},{"label":"Speech synthesizer","score":0.6872},{"label":"Narration, monologue","score":0.3675},{"label":"Conversation","score":0.0382},{"label":"Inside, small room","score":0.0299},{"label":"Music","score":0.0272}],"music_score":0.0272,"speech_score":0.714,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.6551},{"label":"Speech synthesizer","score":0.302},{"label":"Narration, monologue","score":0.2678}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.714},{"label":"Narration, monologue","score":0.3675},{"label":"Speech synthesizer","score":0.2021}]},{"start_s":1.14,"end_s":2.42,"candidates":[{"label":"Speech","score":0.6996},{"label":"Speech synthesizer","score":0.3903},{"label":"Narration, monologue","score":0.1741}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":34.66},"type":"sound_caption","request_ms":206.92,"capture_ms":2220,"event_age_ms":2426.93,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":68661.40000152588},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":68861.29999923706},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":68932.09999847412},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":68991.59999847412},{"type":"asr_sentence","text":"Was können wir heute zusammen machen?","audio_s":65.28,"queued":false,"asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":49.34,"speech_pause_ms":700,"observed_quiet_ms":686},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":686,"at":68992.09999847412},{"type":"identity","enabled":false,"mode":"open-sentences","sentence_mode":true,"open_enrollment":true,"speakers":[{"id":"speaker-1","name":"Christoph","named":true,"reference_ready":true,"voice_variants":1,"admitted":true},{"id":"speaker-2","name":"Selma","named":true,"reference_ready":true,"voice_variants":2,"admitted":true}],"users":["Christoph","Selma"],"threshold":0.38,"conversation_focus":{"enabled":false,"threshold":0.28,"timeout_s":180,"speaker_id":null,"name":null,"primary_id":"speaker-1","participants":[{"id":"speaker-1","name":"Christoph"},{"id":"speaker-2","name":"Selma"}]},"handover_pending":false,"storage":"Session RAM","prediction":{"name":"Selma","similarity":0.9999},"at":69053.59999847412},{"type":"voice_caption","status":"available","caption":"Possibly smooth, slightly steady, very hope","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":2.6,"inference_ms":26.32,"model":"laion/humaneness-ears-base-medium","model_revision":"818506970d93c809a3295a0a02c40dee8ff2bfc3","rules_revision":"7ebd4a9b7da96df989af8fc6ad0ce34f4e8aa3ed","device":"cuda:1","precision":"bfloat16","selected":{"voicenet":[{"dim":"DFLU","name":"Disfluency","value":0.24,"z":-1.06,"direction":"below","intensity":"Notably","always_on":false,"phrase":"notably smooth and flawless","tag":"smooth","reliability":1,"rank_score":1.057},{"dim":"VOLT","name":"Volatility","value":1.287,"z":-0.82,"direction":"below","intensity":"Somewhat","always_on":false,"phrase":"somewhat emotionally steady and static","tag":"slightly steady","reliability":1,"rank_score":0.817},{"dim":"TEMP","name":"Tempo","value":2.912,"z":-0.09,"direction":"below","intensity":"About-average","always_on":true,"phrase":"moderate in tempo","tag":null,"reliability":1,"rank_score":0.092}],"emotions":[{"emotion":"Hope/Optimism","value":2.27,"z":2.18,"direction":"above","intensity":"Extremely","synonym":"hope","phrase":"extremely carrying hope","tag":"very hope"}]},"normalization":"speech reference z-scores; not probabilities","scope":"this turn","request_ms":37.46,"added_wait_ms":0,"at":69056.70000076294},{"type":"endpoint","transcript":"Was können wir heute zusammen machen?","speaker_id":"speaker-2","asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":49.34,"speech_pause_ms":700,"observed_quiet_ms":686},"speaker_name":"Selma","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":54.876,"speaker_embedding_wait_ms":54.2,"speaker_lookup_ms":0.675,"speaker_prefetch_used":false,"speaker_prefetch_compute_ms":null,"sentence_queue_ms":55.17,"sentence_audio_s":65.28},"at":69056.90000152588},{"type":"latency_start","reply_id":5,"input":{"kind":"voice","asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":49.34,"speech_pause_ms":700,"observed_quiet_ms":686},"voice_caption_ms":37.46,"voice_caption_status":"available","speaker_assignment_ms":54.876,"speaker_prefetch_compute_ms":null,"speaker_prefetch_used":false,"sentence_queue_ms":55.16},"at":69057.5},{"type":"reply","text":"Wie wäre es mit einer kleinen Wort-Challenge?","reply_id":5,"performance":{"sentences":[{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Wie wäre es mit einer kleinen Wort-Challenge?","start_s":0,"end_s":2.38}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":0,"sentence_text":"Wie wäre es mit einer kleinen Wort-Challenge?","sentence_role":"answer","speaker_id":"speaker-2","speaker_name":"Selma","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":313.48,"llm_first_sentence_ms":346.47,"sentence_ready_ms":347.64,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":69402.59999847412},{"type":"latency_update","reply_id":5,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":313.48,"llm_first_sentence_ms":346.47,"finish_reason":"stop","usage":{"prompt_tokens":2049,"completion_tokens":56,"total_tokens":2105,"cost":0.0020908305,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00211195,"upstream_inference_prompt_cost":0.00202851,"upstream_inference_completions_cost":8.344e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":392.87,"response_complete":true,"answer_first_audio_ms":1099.26,"tts_first_audio_ms":751.5,"first_audio_ms":1099.27},"at":70153.09999847412},{"type":"reply","text":"Wie wäre es mit einer kleinen Wort-Challenge? Ihr beide denkt euch je ein verrücktes Wort aus, und ich baue daraus eine minikurze, absurde Geschichte – mit Robotern, Katzen und mindestens einem schiefen Kuchen.","reply_id":5,"performance":{"sentences":[{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Wie wäre es mit einer kleinen Wort-Challenge?","start_s":0,"end_s":2.67},{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Ihr beide denkt euch je ein verrücktes Wort aus, und ich baue daraus eine minikurze, absurde Geschichte – mit Robotern, Katzen und mindestens einem schiefen Kuchen.","start_s":2.67,"end_s":11.51}],"source":"qwen/qwen3.8-27b sentence cues + fixed GENERAL catalogue","timing":"Sentence starts anchored to actual preceding PCM duration; within-sentence estimates"},"avatar_actions":[],"avatar_action_id":null,"avatar_event":null,"sentence_index":1,"sentence_text":"Ihr beide denkt euch je ein verrücktes Wort aus, und ich baue daraus eine minikurze, absurde Geschichte – mit Robotern, Katzen und mindestens einem schiefen Kuchen.","sentence_role":"answer","speaker_id":"speaker-2","speaker_name":"Selma","speaker_update":null,"timing":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":313.48,"llm_first_sentence_ms":346.47,"finish_reason":"stop","usage":{"prompt_tokens":2049,"completion_tokens":56,"total_tokens":2105,"cost":0.0020908305,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00211195,"upstream_inference_prompt_cost":0.00202851,"upstream_inference_completions_cost":8.344e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":392.87,"response_complete":true,"answer_first_audio_ms":1099.26,"tts_first_audio_ms":751.5,"first_audio_ms":1099.27,"tts_server_stages":[{"generation_ms":1117,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1272,"performance":{"device_seconds":0.9066310291290283,"raw_audio_seconds":3.2,"aggregate_rtf":0.2833221966028213,"audio_seconds_per_second":3.5295504975978362,"batches":[{"batch_size":1,"device_seconds":0.9066310291290283,"audio_seconds":[3.2],"model_wall_seconds":1.0037527978420258,"first_audio_seconds":0.46231756591796874,"graph_capture_seconds":0,"live_callback_seconds":0.09690460562705994}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1284.95,"audio_s":2.67}],"sentence_ready_ms":1633.48,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":70704},{"type":"latency_update","reply_id":5,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":313.48,"llm_first_sentence_ms":346.47,"finish_reason":"stop","usage":{"prompt_tokens":2049,"completion_tokens":56,"total_tokens":2105,"cost":0.0020908305,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00211195,"upstream_inference_prompt_cost":0.00202851,"upstream_inference_completions_cost":8.344e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":392.87,"response_complete":true,"answer_first_audio_ms":1099.26,"tts_first_audio_ms":751.5,"first_audio_ms":1099.27,"tts_server_stages":[{"generation_ms":1117,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1272,"performance":{"device_seconds":0.9066310291290283,"raw_audio_seconds":3.2,"aggregate_rtf":0.2833221966028213,"audio_seconds_per_second":3.5295504975978362,"batches":[{"batch_size":1,"device_seconds":0.9066310291290283,"audio_seconds":[3.2],"model_wall_seconds":1.0037527978420258,"first_audio_seconds":0.46231756591796874,"graph_capture_seconds":0,"live_callback_seconds":0.09690460562705994}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1284.95,"audio_s":2.67}]},"at":71451.29999923706},{"type":"pipeline_metrics","reply_id":5,"metrics":{"quick_reply_status":"disabled_or_ineligible","llm_model":"qwen/qwen3.8-27b","llm_provider":"cerebras/fp16","llm_backend":"OpenRouter","llm_actual_provider":"Cerebras","llm_ttft_ms":313.48,"llm_first_sentence_ms":346.47,"finish_reason":"stop","usage":{"prompt_tokens":2049,"completion_tokens":56,"total_tokens":2105,"cost":0.0020908305,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00211195,"upstream_inference_prompt_cost":0.00202851,"upstream_inference_completions_cost":8.344e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":392.87,"response_complete":true,"answer_first_audio_ms":1099.26,"tts_first_audio_ms":751.5,"first_audio_ms":1099.27,"tts_server_stages":[{"generation_ms":1117,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1272,"performance":{"device_seconds":0.9066310291290283,"raw_audio_seconds":3.2,"aggregate_rtf":0.2833221966028213,"audio_seconds_per_second":3.5295504975978362,"batches":[{"batch_size":1,"device_seconds":0.9066310291290283,"audio_seconds":[3.2],"model_wall_seconds":1.0037527978420258,"first_audio_seconds":0.46231756591796874,"graph_capture_seconds":0,"live_callback_seconds":0.09690460562705994}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}},{"generation_ms":3649,"alignment_ms":41,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":3802,"performance":{"device_seconds":2.9898438029289243,"raw_audio_seconds":10.48,"aggregate_rtf":0.2852904392107752,"audio_seconds_per_second":3.5051998334272634,"batches":[{"batch_size":1,"device_seconds":2.9898438029289243,"audio_seconds":[10.48],"model_wall_seconds":3.5347136929631233,"first_audio_seconds":0.4815165405273438,"graph_capture_seconds":0,"live_callback_seconds":0.5374252274632454}],"streaming":true,"chunk_frames":20,"backend":"fused CFG + local CUDA graphs + compiled global","includes":"TTS prefill, autoregressive decode and PCM codec; GPU event span including launch gaps","excludes":"reference encoding, prompt packing, Luna, alignment, scoring, voice conversion, CPU audio copy, MP3, network","graph_capture_seconds":0,"oom_batch_retries":[]}}],"sentence_tts":[{"sentence":0,"total_ms":1284.95,"audio_s":2.67},{"sentence":1,"total_ms":3818.01,"audio_s":10.48}],"total_ms":5451.68,"audio_s":13.15,"sentences":2,"generation_rtf":0.4146},"at":74505.5},{"type":"turn_done","barged":false,"at":83733.79999923706}],"reply":"Wie wäre es mit einer kleinen Wort-Challenge? Ihr beide denkt euch je ein verrücktes Wort aus, und ich baue daraus eine minikurze, absurde Geschichte – mit Robotern, Katzen und mindestens einem schiefen Kuchen.","captions":[{"status":"available","caption":"A female voice speaks in German. The speaker is asking a question, and the tone is neutral and conversational.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.42,"tokens":24,"preprocess_ms":7.95,"encoder_gpu_ms":8.98,"decoder_gpu_ms":172.68,"inference_ms":181.9,"total_ms":189.85,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":27.09,"window_s":2.42,"candidates":[{"label":"Speech","score":0.714},{"label":"Speech synthesizer","score":0.6872},{"label":"Narration, monologue","score":0.3675},{"label":"Conversation","score":0.0382},{"label":"Inside, small room","score":0.0299},{"label":"Music","score":0.0272}],"music_score":0.0272,"speech_score":0.714,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.6551},{"label":"Speech synthesizer","score":0.302},{"label":"Narration, monologue","score":0.2678}]},{"start_s":0.64,"end_s":1.92,"candidates":[{"label":"Speech","score":0.714},{"label":"Narration, monologue","score":0.3675},{"label":"Speech synthesizer","score":0.2021}]},{"start_s":1.14,"end_s":2.42,"candidates":[{"label":"Speech","score":0.6996},{"label":"Speech synthesizer","score":0.3903},{"label":"Narration, monologue","score":0.1741}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":34.66},"type":"sound_caption","request_ms":206.92,"capture_ms":2220,"event_age_ms":2426.93,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":68661.40000152588}],"pcm_chunks":13,"page_errors":[],"passed":true,"speakers":["Selma"]}],"sound_provenance":{"dog-1":{"filename":"1-100032-A-0.wav","fold":"1","target":"0","category":"dog","esc10":"True","src_file":"100032","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-100032-A-0.wav","sha256":"f40a849a2375c8c63312a73dd2dd6c74007301fcc21b4be2ece29a642831e3d8","credit":"- [1-100032-A.ogg]: clip derived from rose_bark.wav (http://www.freesound.org/people/nfrae/sounds/100032/) by nfrae [CC0]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"dog-2":{"filename":"1-110389-A-0.wav","fold":"1","target":"0","category":"dog","esc10":"True","src_file":"110389","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-110389-A-0.wav","sha256":"4a18f3f834eb184b9f02595ac74c7de37ef8048a3da92f58f27018fec580b36a","credit":"- [1-110389-A.ogg]: clip derived from animals_dog_bark_springer_spaniel_001.wav (http://www.freesound.org/people/soundscalpel.com/sounds/110389/) by soundscalpel.com [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"cat-1":{"filename":"1-34094-A-5.wav","fold":"1","target":"5","category":"cat","esc10":"False","src_file":"34094","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-34094-A-5.wav","sha256":"ceddf296bcdda3996d35d3f4c292c338050bcb4c5a05a96ca22a69d33e33e4de","credit":"- [1-34094-A.ogg]: clip derived from 070422-cats-sample.wav (http://www.freesound.org/people/sazman/sounds/34094/) by sazman [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"cat-2":{"filename":"1-34094-B-5.wav","fold":"1","target":"5","category":"cat","esc10":"False","src_file":"34094","take":"B","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-34094-B-5.wav","sha256":"43814c96e62a5c2fbf14b84828cc929e381e53c0add7c1e8211d212e1da97734","credit":"- [1-34094-B.ogg]: clip derived from 070422-cats-sample.wav (http://www.freesound.org/people/sazman/sounds/34094/) by sazman [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"glass_breaking-1":{"filename":"1-20133-A-39.wav","fold":"1","target":"39","category":"glass_breaking","esc10":"False","src_file":"20133","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-20133-A-39.wav","sha256":"db6946133c37fbd4513fe932dbb71e0b077bf7d240beaed339eb6c88e48c903b","credit":"- [1-20133-A.ogg]: clip derived from breaking_glass.wav (http://www.freesound.org/people/gezortenplotz/sounds/20133/) by gezortenplotz [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"glass_breaking-2":{"filename":"1-84536-A-39.wav","fold":"1","target":"39","category":"glass_breaking","esc10":"False","src_file":"84536","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-84536-A-39.wav","sha256":"79efba4c0c54acadc88783e83b428fc9c430db94515b25339a6b5598e2ade0c0","credit":"- [1-84536-A.ogg]: clip derived from glass.wav (http://www.freesound.org/people/cmusounddesign/sounds/84536/) by cmusounddesign [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"car_horn-1":{"filename":"1-17124-A-43.wav","fold":"1","target":"43","category":"car_horn","esc10":"False","src_file":"17124","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-17124-A-43.wav","sha256":"9d5e01e07a386ee36c2e73f40817d132a54d91c936eccc01414280abca5a4209","credit":"- [1-17124-A.ogg]: clip derived from car sharp horn.wav (http://www.freesound.org/people/cognito%20perceptu/sounds/17124/) by cognito perceptu [CC0]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"car_horn-2":{"filename":"1-19026-A-43.wav","fold":"1","target":"43","category":"car_horn","esc10":"False","src_file":"19026","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-19026-A-43.wav","sha256":"67bb2fe5cf138930eb6a6a84bedeb7284093076738cd47ef73cc8b3b8393ff40","credit":"- [1-19026-A.ogg]: clip derived from claxon.wav (http://www.freesound.org/people/han1/sounds/19026/) by han1 [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"engine-1":{"filename":"1-18527-A-44.wav","fold":"1","target":"44","category":"engine","esc10":"False","src_file":"18527","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-18527-A-44.wav","sha256":"83eb3e6065246fd87d1acbbe414aa5bfb7c2081a57c188713cdc017819ffc38f","credit":"- [1-18527-A.ogg]: clip derived from garbage_truck.wav (http://www.freesound.org/people/Corsica_S/sounds/18527/) by Corsica_S [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"engine-2":{"filename":"1-18527-B-44.wav","fold":"1","target":"44","category":"engine","esc10":"False","src_file":"18527","take":"B","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-18527-B-44.wav","sha256":"40f0ab8a2fa68cbdf814f10fc8b2c26916b3dc54c1b0e719d4697ba58c0bdb81","credit":"- [1-18527-B.ogg]: clip derived from garbage_truck.wav (http://www.freesound.org/people/Corsica_S/sounds/18527/) by Corsica_S [CC-BY]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"rain-1":{"filename":"1-17367-A-10.wav","fold":"1","target":"10","category":"rain","esc10":"True","src_file":"17367","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-17367-A-10.wav","sha256":"d732acd2e0c7c40405968010f49f250740c86998fc416eff2ed5330fc8c89dea","credit":"- [1-17367-A.ogg]: clip derived from unlocki door to mod rain.wav (http://www.freesound.org/people/cognito%20perceptu/sounds/17367/) by cognito perceptu [CC0]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"rain-2":{"filename":"1-21189-A-10.wav","fold":"1","target":"10","category":"rain","esc10":"True","src_file":"21189","take":"A","url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/audio/1-21189-A-10.wav","sha256":"5f15ab2d8f6ef9932909b1b75c8ac254cbea46c93b4b2f0f3542abdf09b9abc0","credit":"- [1-21189-A.ogg]: clip derived from Louisiana Rain 1.wav (http://www.freesound.org/people/uberhuberman/sounds/21189/) by uberhuberman [CC0]","source_license_url":"https://raw.githubusercontent.com/karolpiczak/ESC-50/33c8ce9eb2cf0b1c2f8bcf322eb349b6be34dbb6/LICENSE","audio_redistributed":false},"authored_music":{"license":"CC0 authored deterministic melody, environment_fixtures.py"},"authored_impacts":{"license":"CC0 procedural impacts"}},"quiet_background":{"events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":23447.40000152588},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":23935.09999847412},{"type":"asr_preview","text":"Morgen wird es","status":"sentence_stream","source":"local_sentence_asr","at":24423.40000152588},{"type":"asr_preview","text":"Morgen wird es sonnig","status":"sentence_stream","source":"local_sentence_asr","at":24897.70000076294},{"type":"asr_preview","text":"Morgen wird es sonnig.","status":"sentence_stream","source":"native_asr","at":25056.79999923706},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":25177},{"type":"asr_preview","text":"Morgen wird es sonnig.","status":"sentence_stream","source":"local_sentence_asr","at":25240.70000076294},{"type":"asr_preview","text":"Morgen wird es sonnig im Hintergrund","status":"sentence_stream","source":"local_sentence_asr","at":26070.20000076294},{"type":"asr_preview","text":"Hintergrund läuft","status":"sentence_stream","source":"native_asr","at":26657.70000076294},{"type":"asr_preview","text":"Morgen wird es sonnig im Hintergrund läuft ein","status":"sentence_stream","source":"local_sentence_asr","at":26696.70000076294},{"status":"available","caption":"The audio contains a brief, high-pitched electronic tone, similar to a short beep or a digital alert sound. This is a simple electronic sound, likely a notification, alert, or a short electronic signal.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":1.84,"tokens":44,"preprocess_ms":10.26,"encoder_gpu_ms":9.42,"decoder_gpu_ms":357.04,"inference_ms":366.73,"total_ms":376.99,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","sound_events":{"status":"available","model":"EfficientAT mn04_as","revision":"a425fdce92572e602a1d5634799bd9f1f2efa806","weight_sha256":"899a8c6217063f6941793102a2780b3dd788ab11fce468d9cf5f0577c453321c","device":"cpu","parameters":983599,"total_ms":23.78,"window_s":1.84,"candidates":[{"label":"Speech","score":0.8552},{"label":"Narration, monologue","score":0.2129},{"label":"Silence","score":0.0726},{"label":"Inside, small room","score":0.0523},{"label":"Music","score":0.0384},{"label":"Speech synthesizer","score":0.0331}],"music_score":0.0384,"speech_score":0.8552,"frames":[{"start_s":0,"end_s":1.28,"candidates":[{"label":"Speech","score":0.7761},{"label":"Silence","score":0.0726},{"label":"Inside, small room","score":0.0523}]},{"start_s":0.56,"end_s":1.84,"candidates":[{"label":"Speech","score":0.8552},{"label":"Narration, monologue","score":0.2129},{"label":"Music","score":0.0384}]}],"temporal_resolution_note":"Overlapping 1.28-second classification windows, 0.64-second stride; not precise event boundaries or spatial direction.","uncertainty":"Multilabel model scores, not calibrated event probabilities.","request_ms":34.09},"type":"sound_caption","request_ms":397.17,"capture_ms":1640,"event_age_ms":2037.17,"kind":"voice","routing_source":"dominant_speech_event_hint","eligible":false,"suppression_reason":"speech_or_quiet","at":27256.70000076294},{"type":"asr_preview","text":"Morgen wird es sonnig im Hintergrund läuft ein Bericht über das Welt.","status":"sentence_stream","source":"local_sentence_asr","at":27371.5},{"type":"asr_preview","text":"Hintergrund läuft ein Bericht über das Wert.","status":"sentence_stream","source":"native_asr","at":27603.79999923706},{"type":"asr_preview","text":"Morgen wird es sonnig im Hintergrund läuft ein Bericht über das Wetter.","status":"sentence_stream","source":"local_sentence_asr","at":28619},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":28956.40000152588},{"type":"asr_preview","text":"Morgen wird es sonnig. Im Hintergrund läuft ein Bericht über das Wetter. Und die Nachrichten.","status":"sentence_stream","source":"local_sentence_asr","at":29122},{"type":"asr_preview","text":"und die Nachrichten","status":"sentence_stream","source":"native_asr","at":29211.90000152588},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":29893.40000152588},{"type":"asr_preview","text":"Morgen wird es sonnig. Im Hintergrund läuft ein Bericht über das Wetter. Und die Nachrichten.","status":"sentence_stream","source":"local_sentence_asr","at":30032.29999923706},{"type":"asr_sentence","text":"Morgen wird es sonnig. Im Hintergrund läuft ein Bericht über das Wetter. Und die Nachrichten.","audio_s":26.88,"queued":false,"asr":{"text":"Morgen wird es sonnig. Im Hintergrund läuft ein Bericht über das Wetter. Und die Nachrichten.","model":"parakeet_v3","chunks":1,"transcribe_ms":130.6,"speech_pause_ms":700,"observed_quiet_ms":852},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":852,"at":30034.70000076294},{"type":"speaker_focus","operation":"turn","accepted":false,"source":"background_salience","relative_level":0.124,"known_voice_similarity":0.1542,"explicit_address":false,"estimated_level_above_floor_db":99.8,"reason":"Very quiet, unmatched speech without an explicit address","confidence_is_calibrated":false,"at":30035.20000076294},{"type":"ignored","accepted":false,"source":"background_salience","relative_level":0.124,"known_voice_similarity":0.1542,"explicit_address":false,"estimated_level_above_floor_db":99.8,"reason":"Very quiet, unmatched speech without an explicit address","confidence_is_calibrated":false,"at":30035.20000076294}],"passed":true},"final_identity":{"type":"identity","enabled":false,"mode":"open-sentences","sentence_mode":true,"open_enrollment":true,"speakers":[{"id":"speaker-1","name":"Christoph","named":true,"reference_ready":true,"voice_variants":1,"admitted":true},{"id":"speaker-2","name":"Selma","named":true,"reference_ready":true,"voice_variants":2,"admitted":true}],"users":["Christoph","Selma"],"threshold":0.38,"conversation_focus":{"enabled":false,"threshold":0.28,"timeout_s":180,"speaker_id":null,"name":null,"primary_id":"speaker-1","participants":[{"id":"speaker-1","name":"Christoph"},{"id":"speaker-2","name":"Selma"}]},"handover_pending":false,"storage":"Session RAM","prediction":{"name":"Selma","similarity":0.9999},"at":69053.59999847412},"named_speakers_present":true,"source_unchanged":true}