{"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":"20261027T133159Z","group":"speakers","source_sha256":"6332c5bf777de223431c0796be08868e870cccd7e2fefe45b7ca3c517e3eaf1a","injections":[{"browser_start_ms":4859.5,"browser_end_ms":9299,"packets":222,"max_schedule_lag_ms":7.099998474121094,"mean_schedule_lag_ms":0.6144139573380754,"stream_start_sample":38400,"samples":70880,"sha256":"5d665c310e221c7e3af7ac9ed35c4762517c8a14d93056ea055ebafca2b766b7"},{"browser_start_ms":20638.699996948242,"browser_end_ms":26998.199996948242,"packets":318,"max_schedule_lag_ms":5.400001525878906,"mean_schedule_lag_ms":0.5943399945145134,"stream_start_sample":285280,"samples":101600,"sha256":"422772ed90db71170764837bcd8dbc42cf8ed947e0cd93d788cb37ca1b2fd150"},{"browser_start_ms":31086.699996948242,"browser_end_ms":35046.5,"packets":198,"max_schedule_lag_ms":20.5,"mean_schedule_lag_ms":2.612121890289615,"stream_start_sample":450880,"samples":63200,"sha256":"dd2ca90a85488b86997a05c74e6fe3017ea091f6832a627d073e141fc3738249"},{"browser_start_ms":54562.69999694824,"browser_end_ms":59562.79999923706,"packets":250,"max_schedule_lag_ms":6.3000030517578125,"mean_schedule_lag_ms":0.7772003631591797,"stream_start_sample":818080,"samples":79840,"sha256":"4dcb7fd02b14b2e29cbc064502468df8c18d7b862839701a500f21f6821559fd"},{"browser_start_ms":77911.09999847412,"browser_end_ms":81874.19999694824,"packets":198,"max_schedule_lag_ms":3.299999237060547,"mean_schedule_lag_ms":0.38838374975955847,"stream_start_sample":1182720,"samples":63200,"sha256":"dd2ca90a85488b86997a05c74e6fe3017ea091f6832a627d073e141fc3738249"}],"cases":[{"case":"christoph-intro","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":5409.099998474121},{"type":"asr_preview","text":"Hey Rao.","status":"sentence_stream","source":"local_sentence_asr","at":5880.699996948242},{"type":"asr_preview","text":"Hey Rob, ich","status":"sentence_stream","source":"local_sentence_asr","at":6432.299999237061},{"type":"asr_preview","text":"Hey Robbie, bin.","status":"sentence_stream","source":"native_asr","at":7014.099998474121},{"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":11.78,"encoder_gpu_ms":15.41,"decoder_gpu_ms":231.25,"inference_ms":246.92,"total_ms":258.7,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","type":"sound_caption","request_ms":271.5,"capture_ms":1920,"event_age_ms":2191.51,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":7475.799999237061},{"type":"asr_preview","text":"Hey Rabbi, ich bin Christoph.","status":"sentence_stream","source":"local_sentence_asr","at":7761},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":7815.599998474121},{"type":"asr_preview","text":"Hey Robbie, ich bin Christoph.","status":"sentence_stream","source":"local_sentence_asr","at":7876.299999237061},{"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":52.42,"speech_pause_ms":700,"observed_quiet_ms":16},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":16,"at":7876.899997711182},{"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":7877.199996948242},{"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":36.44,"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":43.71,"added_wait_ms":0,"at":7877.899997711182},{"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":52.42,"speech_pause_ms":700,"observed_quiet_ms":16},"speaker_name":"Christoph","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":0.538,"speaker_embedding_wait_ms":0.001,"speaker_lookup_ms":0.537,"speaker_prefetch_used":true,"speaker_prefetch_compute_ms":67.97,"sentence_queue_ms":0.66,"sentence_audio_s":5.12},"at":7878.099998474121},{"type":"latency_start","reply_id":1,"input":{"kind":"voice","asr":{"text":"Hey Robbie, ich bin Christoph.","model":"parakeet_v3","chunks":1,"transcribe_ms":52.42,"speech_pause_ms":700,"observed_quiet_ms":16},"voice_caption_ms":43.71,"voice_caption_status":"available","speaker_assignment_ms":0.538,"speaker_prefetch_compute_ms":67.97,"speaker_prefetch_used":true,"sentence_queue_ms":0.66},"at":7883},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":8352.399997711182},{"type":"reply","text":"Hallo Christoph!","reply_id":1,"performance":{"sentences":[{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","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":549.88,"llm_first_sentence_ms":556.04,"sentence_ready_ms":565.06,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"elation.soft"},"sources":[],"lookups":[],"at":8447.79999923706},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":8856.099998474121},{"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":549.88,"llm_first_sentence_ms":556.04,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":22,"total_tokens":1767,"cost":0.0017427267,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00176033,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":3.278e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":584.21,"response_complete":true,"answer_first_audio_ms":1309.32,"tts_first_audio_ms":744.1,"first_audio_ms":1309.32},"at":9189.099998474121},{"type":"asr_preview","text":"schön dich kennenzulernen.","status":"sentence_stream","source":"local_sentence_asr","at":9364.099998474121},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":9726.199996948242},{"type":"asr_preview","text":"Schön, dich kennenzulernen.","status":"sentence_stream","source":"local_sentence_asr","at":9801.099998474121},{"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":57.02,"speech_pause_ms":700,"observed_quiet_ms":618},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":618,"at":9802.199996948242},{"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":9802.399997711182},{"type":"reply","text":"Hallo Christoph! Schön, dass du da bist!","reply_id":1,"performance":{"sentences":[{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Hallo Christoph!","start_s":0,"end_s":2.51},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Schön, dass du da bist!","start_s":2.51,"end_s":4.21}],"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!","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":549.88,"llm_first_sentence_ms":556.04,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":22,"total_tokens":1767,"cost":0.0017427267,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00176033,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":3.278e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":584.21,"response_complete":true,"answer_first_audio_ms":1309.32,"tts_first_audio_ms":744.1,"first_audio_ms":1309.32,"tts_server_stages":[{"generation_ms":1265,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1425,"performance":{"device_seconds":1.0150613689422607,"raw_audio_seconds":3.36,"aggregate_rtf":0.3021015978994824,"audio_seconds_per_second":3.3101446895780007,"batches":[{"batch_size":1,"device_seconds":1.0150613689422607,"audio_seconds":[3.36],"model_wall_seconds":1.1512103229761124,"first_audio_seconds":0.45372006225585937,"graph_capture_seconds":0,"live_callback_seconds":0.13556092977523804}],"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":1435.7,"audio_s":2.51}],"sentence_ready_ms":2001.9,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"elation.soft"},"sources":[],"lookups":[],"at":9893},{"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":549.88,"llm_first_sentence_ms":556.04,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":22,"total_tokens":1767,"cost":0.0017427267,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00176033,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":3.278e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":584.21,"response_complete":true,"answer_first_audio_ms":1309.32,"tts_first_audio_ms":744.1,"first_audio_ms":1309.32,"tts_server_stages":[{"generation_ms":1265,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1425,"performance":{"device_seconds":1.0150613689422607,"raw_audio_seconds":3.36,"aggregate_rtf":0.3021015978994824,"audio_seconds_per_second":3.3101446895780007,"batches":[{"batch_size":1,"device_seconds":1.0150613689422607,"audio_seconds":[3.36],"model_wall_seconds":1.1512103229761124,"first_audio_seconds":0.45372006225585937,"graph_capture_seconds":0,"live_callback_seconds":0.13556092977523804}],"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":1435.7,"audio_s":2.51}]},"at":10626.199996948242},{"type":"reply","text":"Hallo Christoph! Schön, dass du da bist! Wie geht es dir heute?","reply_id":1,"performance":{"sentences":[{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Hallo Christoph!","start_s":0,"end_s":2.51},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Schön, dass du da bist!","start_s":2.51,"end_s":4.54},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, playful","text":"Wie geht es dir heute?","start_s":4.54,"end_s":6.24}],"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":"Wie geht es dir heute?","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":549.88,"llm_first_sentence_ms":556.04,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":22,"total_tokens":1767,"cost":0.0017427267,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00176033,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":3.278e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":584.21,"response_complete":true,"answer_first_audio_ms":1309.32,"tts_first_audio_ms":744.1,"first_audio_ms":1309.32,"tts_server_stages":[{"generation_ms":1265,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1425,"performance":{"device_seconds":1.0150613689422607,"raw_audio_seconds":3.36,"aggregate_rtf":0.3021015978994824,"audio_seconds_per_second":3.3101446895780007,"batches":[{"batch_size":1,"device_seconds":1.0150613689422607,"audio_seconds":[3.36],"model_wall_seconds":1.1512103229761124,"first_audio_seconds":0.45372006225585937,"graph_capture_seconds":0,"live_callback_seconds":0.13556092977523804}],"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":1008,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1159,"performance":{"device_seconds":0.8145895433425903,"raw_audio_seconds":2.8,"aggregate_rtf":0.290924836908068,"audio_seconds_per_second":3.4373139489496363,"batches":[{"batch_size":1,"device_seconds":0.8145895433425903,"audio_seconds":[2.8],"model_wall_seconds":0.8980023264884949,"first_audio_seconds":0.4565985412597656,"graph_capture_seconds":0,"live_callback_seconds":0.08320546895265579}],"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":1435.7,"audio_s":2.51},{"sentence":1,"total_ms":1168.89,"audio_s":2.03}],"sentence_ready_ms":3171.88,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"elation.soft"},"sources":[],"lookups":[],"at":11062.099998474121},{"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":549.88,"llm_first_sentence_ms":556.04,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":22,"total_tokens":1767,"cost":0.0017427267,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00176033,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":3.278e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":584.21,"response_complete":true,"answer_first_audio_ms":1309.32,"tts_first_audio_ms":744.1,"first_audio_ms":1309.32,"tts_server_stages":[{"generation_ms":1265,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1425,"performance":{"device_seconds":1.0150613689422607,"raw_audio_seconds":3.36,"aggregate_rtf":0.3021015978994824,"audio_seconds_per_second":3.3101446895780007,"batches":[{"batch_size":1,"device_seconds":1.0150613689422607,"audio_seconds":[3.36],"model_wall_seconds":1.1512103229761124,"first_audio_seconds":0.45372006225585937,"graph_capture_seconds":0,"live_callback_seconds":0.13556092977523804}],"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":1008,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1159,"performance":{"device_seconds":0.8145895433425903,"raw_audio_seconds":2.8,"aggregate_rtf":0.290924836908068,"audio_seconds_per_second":3.4373139489496363,"batches":[{"batch_size":1,"device_seconds":0.8145895433425903,"audio_seconds":[2.8],"model_wall_seconds":0.8980023264884949,"first_audio_seconds":0.4565985412597656,"graph_capture_seconds":0,"live_callback_seconds":0.08320546895265579}],"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":1435.7,"audio_s":2.51},{"sentence":1,"total_ms":1168.89,"audio_s":2.03}]},"at":11794.099998474121},{"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":549.88,"llm_first_sentence_ms":556.04,"finish_reason":"stop","usage":{"prompt_tokens":1745,"completion_tokens":22,"total_tokens":1767,"cost":0.0017427267,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00176033,"upstream_inference_prompt_cost":0.00172755,"upstream_inference_completions_cost":3.278e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":584.21,"response_complete":true,"answer_first_audio_ms":1309.32,"tts_first_audio_ms":744.1,"first_audio_ms":1309.32,"tts_server_stages":[{"generation_ms":1265,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1425,"performance":{"device_seconds":1.0150613689422607,"raw_audio_seconds":3.36,"aggregate_rtf":0.3021015978994824,"audio_seconds_per_second":3.3101446895780007,"batches":[{"batch_size":1,"device_seconds":1.0150613689422607,"audio_seconds":[3.36],"model_wall_seconds":1.1512103229761124,"first_audio_seconds":0.45372006225585937,"graph_capture_seconds":0,"live_callback_seconds":0.13556092977523804}],"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":1008,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1159,"performance":{"device_seconds":0.8145895433425903,"raw_audio_seconds":2.8,"aggregate_rtf":0.290924836908068,"audio_seconds_per_second":3.4373139489496363,"batches":[{"batch_size":1,"device_seconds":0.8145895433425903,"audio_seconds":[2.8],"model_wall_seconds":0.8980023264884949,"first_audio_seconds":0.4565985412597656,"graph_capture_seconds":0,"live_callback_seconds":0.08320546895265579}],"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":938,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1087,"performance":{"device_seconds":0.7441845273971558,"raw_audio_seconds":2.48,"aggregate_rtf":0.30007440620853054,"audio_seconds_per_second":3.332506802679701,"batches":[{"batch_size":1,"device_seconds":0.7441845273971558,"audio_seconds":[2.48],"model_wall_seconds":0.8291971608996391,"first_audio_seconds":0.4533749694824219,"graph_capture_seconds":0,"live_callback_seconds":0.08356590569019318}],"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":1435.7,"audio_s":2.51},{"sentence":1,"total_ms":1168.89,"audio_s":2.03},{"sentence":2,"total_ms":1097.13,"audio_s":1.71}],"total_ms":4269.22,"audio_s":6.25,"sentences":3,"generation_rtf":0.6831},"at":12149.29999923706},{"type":"turn_done","barged":false,"at":15892.099998474121},{"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":26.99,"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":39.02,"added_wait_ms":0,"at":15892.79999923706},{"type":"endpoint","transcript":"Schön, dich kennenzulernen.","speaker_id":"speaker-1","asr":{"text":"Schön, dich kennenzulernen.","model":"parakeet_v3","chunks":1,"transcribe_ms":57.02,"speech_pause_ms":700,"observed_quiet_ms":618},"speaker_name":"Christoph","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":1.058,"speaker_embedding_wait_ms":0.002,"speaker_lookup_ms":1.056,"speaker_prefetch_used":true,"speaker_prefetch_compute_ms":45.83,"sentence_queue_ms":6103.19,"sentence_audio_s":7.04},"at":15893},{"type":"latency_start","reply_id":2,"input":{"kind":"voice","asr":{"text":"Schön, dich kennenzulernen.","model":"parakeet_v3","chunks":1,"transcribe_ms":57.02,"speech_pause_ms":700,"observed_quiet_ms":618},"voice_caption_ms":39.02,"voice_caption_status":"available","speaker_assignment_ms":1.058,"speaker_prefetch_compute_ms":45.83,"speaker_prefetch_used":true,"sentence_queue_ms":6103.19},"at":15893.399997711182},{"type":"reply","text":"Das freut mich sehr!","reply_id":2,"performance":{"sentences":[{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"interest, warm","text":"Das freut mich sehr!","start_s":0,"end_s":1.36}],"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 freut mich sehr!","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":433.62,"finish_reason":"stop","usage":{"prompt_tokens":1813,"completion_tokens":11,"total_tokens":1824,"cost":0.0017931474,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00181126,"upstream_inference_prompt_cost":0.00179487,"upstream_inference_completions_cost":1.639e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":469.13,"generation_calls":1,"llm_total_ms":469.16,"response_complete":true,"sentence_ready_ms":470.26,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":16363.5},{"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":433.62,"finish_reason":"stop","usage":{"prompt_tokens":1813,"completion_tokens":11,"total_tokens":1824,"cost":0.0017931474,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00181126,"upstream_inference_prompt_cost":0.00179487,"upstream_inference_completions_cost":1.639e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":469.13,"generation_calls":1,"llm_total_ms":469.16,"response_complete":true,"answer_first_audio_ms":1214.48,"tts_first_audio_ms":744.03,"first_audio_ms":1214.49},"at":17110.199996948242},{"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":433.62,"finish_reason":"stop","usage":{"prompt_tokens":1813,"completion_tokens":11,"total_tokens":1824,"cost":0.0017931474,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00181126,"upstream_inference_prompt_cost":0.00179487,"upstream_inference_completions_cost":1.639e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":469.13,"generation_calls":1,"llm_total_ms":469.16,"response_complete":true,"answer_first_audio_ms":1214.48,"tts_first_audio_ms":744.03,"first_audio_ms":1214.49,"tts_server_stages":[{"generation_ms":934,"alignment_ms":35,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1085,"performance":{"device_seconds":0.7384938068389892,"raw_audio_seconds":2.4,"aggregate_rtf":0.30770575284957885,"audio_seconds_per_second":3.249857991731625,"batches":[{"batch_size":1,"device_seconds":0.7384938068389892,"audio_seconds":[2.4],"model_wall_seconds":0.8184135034680367,"first_audio_seconds":0.46523495483398436,"graph_capture_seconds":0,"live_callback_seconds":0.07969994843006134}],"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":1094.92,"audio_s":1.87}],"total_ms":1565.39,"audio_s":1.87,"sentences":1,"generation_rtf":0.8371},"at":17464.199996948242},{"type":"turn_done","barged":false,"at":19423.199996948242}],"reply":"Das freut mich sehr!","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":11.78,"encoder_gpu_ms":15.41,"decoder_gpu_ms":231.25,"inference_ms":246.92,"total_ms":258.7,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","type":"sound_caption","request_ms":271.5,"capture_ms":1920,"event_age_ms":2191.51,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":7475.799999237061}],"pcm_chunks":17,"page_errors":[],"passed":true,"speakers":["Christoph","Christoph"]},{"case":"selma-followup-clear","events":[{"type":"turn_done","barged":false,"at":32358.199996948242},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":32940},{"type":"asr_preview","text":"Was können","status":"sentence_stream","source":"local_sentence_asr","at":33446.69999694824},{"type":"asr_preview","text":"Was können","status":"sentence_stream","source":"native_asr","at":33693.89999771118},{"type":"asr_preview","text":"Was können wir heute zusammen?","status":"sentence_stream","source":"local_sentence_asr","at":34446.59999847412},{"type":"asr_preview","text":"Was können wir heute zusammen","status":"sentence_stream","source":"native_asr","at":34652.69999694824},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":35446.29999923706},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":35531.29999923706},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":35589.29999923706},{"type":"asr_sentence","text":"Was können wir heute zusammen machen?","audio_s":32.32,"queued":false,"asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":44.89,"speech_pause_ms":700,"observed_quiet_ms":654},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":654,"at":35589.69999694824},{"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":0.065},"at":35591},{"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":26.89,"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":39.16,"added_wait_ms":0,"at":35593.79999923706},{"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":44.89,"speech_pause_ms":700,"observed_quiet_ms":654},"speaker_name":"Speaker 2","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":0.671,"speaker_embedding_wait_ms":0.002,"speaker_lookup_ms":0.669,"speaker_prefetch_used":true,"speaker_prefetch_compute_ms":53.17,"sentence_queue_ms":0.91,"sentence_audio_s":32.32},"at":35593.89999771118},{"type":"latency_start","reply_id":4,"input":{"kind":"voice","asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":44.89,"speech_pause_ms":700,"observed_quiet_ms":654},"voice_caption_ms":39.16,"voice_caption_status":"available","speaker_assignment_ms":0.671,"speaker_prefetch_compute_ms":53.17,"speaker_prefetch_used":true,"sentence_queue_ms":0.91},"at":35594.59999847412},{"type":"reply","text":"Wie wäre es mit einer kleinen Reisederby?","reply_id":4,"performance":{"sentences":[{"emotion":"affection","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"warm, playful","text":"Wie wäre es mit einer kleinen Reisederby?","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 Reisederby?","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":319.8,"llm_first_sentence_ms":353.86,"sentence_ready_ms":355,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.warm.soft"},"sources":[],"lookups":[],"at":35949.79999923706},{"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":319.8,"llm_first_sentence_ms":353.86,"finish_reason":"stop","usage":{"prompt_tokens":1962,"completion_tokens":67,"total_tokens":2029,"cost":0.0020217879,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00204221,"upstream_inference_prompt_cost":0.00194238,"upstream_inference_completions_cost":9.983e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":435.65,"response_complete":true,"answer_first_audio_ms":1110.82,"tts_first_audio_ms":755.69,"first_audio_ms":1110.83},"at":36708.5},{"type":"reply","text":"Wie wäre es mit einer kleinen Reisederby? Du nennst mir ein Ziel, und ich erfinde dir eine verrückte, kurze Geschichte, wie wir beide dorthin kommen – vielleicht per fliegendem Koffer oder einem selbstgebauten Waffel-Raketenwagen?","reply_id":4,"performance":{"sentences":[{"emotion":"affection","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"warm, playful","text":"Wie wäre es mit einer kleinen Reisederby?","start_s":0,"end_s":2.83},{"emotion":"affection","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"warm, playful","text":"Du nennst mir ein Ziel, und ich erfinde dir eine verrückte, kurze Geschichte, wie wir beide dorthin kommen – vielleicht per fliegendem Koffer oder einem selbstgebauten Waffel-Raketenwagen?","start_s":2.83,"end_s":12.01}],"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":"Du nennst mir ein Ziel, und ich erfinde dir eine verrückte, kurze Geschichte, wie wir beide dorthin kommen – vielleicht per fliegendem Koffer oder einem selbstgebauten Waffel-Raketenwagen?","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":319.8,"llm_first_sentence_ms":353.86,"finish_reason":"stop","usage":{"prompt_tokens":1962,"completion_tokens":67,"total_tokens":2029,"cost":0.0020217879,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00204221,"upstream_inference_prompt_cost":0.00194238,"upstream_inference_completions_cost":9.983e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":435.65,"response_complete":true,"answer_first_audio_ms":1110.82,"tts_first_audio_ms":755.69,"first_audio_ms":1110.83,"tts_server_stages":[{"generation_ms":1067,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1225,"performance":{"device_seconds":0.8556111192703247,"raw_audio_seconds":2.96,"aggregate_rtf":0.2890578105642989,"audio_seconds_per_second":3.45951558287873,"batches":[{"batch_size":1,"device_seconds":0.8556111192703247,"audio_seconds":[2.96],"model_wall_seconds":0.9526316523551941,"first_audio_seconds":0.464721923828125,"graph_capture_seconds":0,"live_callback_seconds":0.09679634124040604}],"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":1241.03,"audio_s":2.83}],"sentence_ready_ms":1597.11,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.warm.soft"},"sources":[],"lookups":[],"at":37195.29999923706},{"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":319.8,"llm_first_sentence_ms":353.86,"finish_reason":"stop","usage":{"prompt_tokens":1962,"completion_tokens":67,"total_tokens":2029,"cost":0.0020217879,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00204221,"upstream_inference_prompt_cost":0.00194238,"upstream_inference_completions_cost":9.983e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":435.65,"response_complete":true,"answer_first_audio_ms":1110.82,"tts_first_audio_ms":755.69,"first_audio_ms":1110.83,"tts_server_stages":[{"generation_ms":1067,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1225,"performance":{"device_seconds":0.8556111192703247,"raw_audio_seconds":2.96,"aggregate_rtf":0.2890578105642989,"audio_seconds_per_second":3.45951558287873,"batches":[{"batch_size":1,"device_seconds":0.8556111192703247,"audio_seconds":[2.96],"model_wall_seconds":0.9526316523551941,"first_audio_seconds":0.464721923828125,"graph_capture_seconds":0,"live_callback_seconds":0.09679634124040604}],"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":1241.03,"audio_s":2.83}]},"at":37957.39999771118},{"type":"reply","text":"Wie wäre es mit einer kleinen Reisederby? Du nennst mir ein Ziel, und ich erfinde dir eine verrückte, kurze Geschichte, wie wir beide dorthin kommen – vielleicht per fliegendem Koffer oder einem selbstgebauten Waffel-Raketenwagen? Und wie darf ich dich nennen?","reply_id":4,"performance":{"sentences":[{"emotion":"affection","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"warm, playful","text":"Wie wäre es mit einer kleinen Reisederby?","start_s":0,"end_s":2.83},{"emotion":"affection","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"warm, playful","text":"Du nennst mir ein Ziel, und ich erfinde dir eine verrückte, kurze Geschichte, wie wir beide dorthin kommen – vielleicht per fliegendem Koffer oder einem selbstgebauten Waffel-Raketenwagen?","start_s":2.83,"end_s":13.63},{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.55,"direction":"interest, soft, warm","text":"Und wie darf ich dich nennen?","start_s":13.63,"end_s":15.670000000000002}],"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":"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":319.8,"llm_first_sentence_ms":353.86,"finish_reason":"stop","usage":{"prompt_tokens":1962,"completion_tokens":67,"total_tokens":2029,"cost":0.0020217879,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00204221,"upstream_inference_prompt_cost":0.00194238,"upstream_inference_completions_cost":9.983e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":435.65,"response_complete":true,"answer_first_audio_ms":1110.82,"tts_first_audio_ms":755.69,"first_audio_ms":1110.83,"tts_server_stages":[{"generation_ms":1067,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1225,"performance":{"device_seconds":0.8556111192703247,"raw_audio_seconds":2.96,"aggregate_rtf":0.2890578105642989,"audio_seconds_per_second":3.45951558287873,"batches":[{"batch_size":1,"device_seconds":0.8556111192703247,"audio_seconds":[2.96],"model_wall_seconds":0.9526316523551941,"first_audio_seconds":0.464721923828125,"graph_capture_seconds":0,"live_callback_seconds":0.09679634124040604}],"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":3720,"alignment_ms":44,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":3886,"performance":{"device_seconds":3.048328489303589,"raw_audio_seconds":10.8,"aggregate_rtf":0.28225263789848043,"audio_seconds_per_second":3.5429252581854564,"batches":[{"batch_size":1,"device_seconds":3.048328489303589,"audio_seconds":[10.8],"model_wall_seconds":3.6077364087104797,"first_audio_seconds":0.48071578979492186,"graph_capture_seconds":0,"live_callback_seconds":0.555124931037426}],"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":1241.03,"audio_s":2.83},{"sentence":1,"total_ms":3901.28,"audio_s":10.8}],"sentence_ready_ms":5499.71,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":41095.59999847412},{"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":319.8,"llm_first_sentence_ms":353.86,"finish_reason":"stop","usage":{"prompt_tokens":1962,"completion_tokens":67,"total_tokens":2029,"cost":0.0020217879,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00204221,"upstream_inference_prompt_cost":0.00194238,"upstream_inference_completions_cost":9.983e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":435.65,"response_complete":true,"answer_first_audio_ms":1110.82,"tts_first_audio_ms":755.69,"first_audio_ms":1110.83,"tts_server_stages":[{"generation_ms":1067,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1225,"performance":{"device_seconds":0.8556111192703247,"raw_audio_seconds":2.96,"aggregate_rtf":0.2890578105642989,"audio_seconds_per_second":3.45951558287873,"batches":[{"batch_size":1,"device_seconds":0.8556111192703247,"audio_seconds":[2.96],"model_wall_seconds":0.9526316523551941,"first_audio_seconds":0.464721923828125,"graph_capture_seconds":0,"live_callback_seconds":0.09679634124040604}],"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":3720,"alignment_ms":44,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":3886,"performance":{"device_seconds":3.048328489303589,"raw_audio_seconds":10.8,"aggregate_rtf":0.28225263789848043,"audio_seconds_per_second":3.5429252581854564,"batches":[{"batch_size":1,"device_seconds":3.048328489303589,"audio_seconds":[10.8],"model_wall_seconds":3.6077364087104797,"first_audio_seconds":0.48071578979492186,"graph_capture_seconds":0,"live_callback_seconds":0.555124931037426}],"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":1241.03,"audio_s":2.83},{"sentence":1,"total_ms":3901.28,"audio_s":10.8}]},"at":41840},{"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":319.8,"llm_first_sentence_ms":353.86,"finish_reason":"stop","usage":{"prompt_tokens":1962,"completion_tokens":67,"total_tokens":2029,"cost":0.0020217879,"is_byok":false,"prompt_tokens_details":{"cached_tokens":0,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00204221,"upstream_inference_prompt_cost":0.00194238,"upstream_inference_completions_cost":9.983e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"name_question_added":true,"llm_total_ms":435.65,"response_complete":true,"answer_first_audio_ms":1110.82,"tts_first_audio_ms":755.69,"first_audio_ms":1110.83,"tts_server_stages":[{"generation_ms":1067,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1225,"performance":{"device_seconds":0.8556111192703247,"raw_audio_seconds":2.96,"aggregate_rtf":0.2890578105642989,"audio_seconds_per_second":3.45951558287873,"batches":[{"batch_size":1,"device_seconds":0.8556111192703247,"audio_seconds":[2.96],"model_wall_seconds":0.9526316523551941,"first_audio_seconds":0.464721923828125,"graph_capture_seconds":0,"live_callback_seconds":0.09679634124040604}],"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":3720,"alignment_ms":44,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":3886,"performance":{"device_seconds":3.048328489303589,"raw_audio_seconds":10.8,"aggregate_rtf":0.28225263789848043,"audio_seconds_per_second":3.5429252581854564,"batches":[{"batch_size":1,"device_seconds":3.048328489303589,"audio_seconds":[10.8],"model_wall_seconds":3.6077364087104797,"first_audio_seconds":0.48071578979492186,"graph_capture_seconds":0,"live_callback_seconds":0.555124931037426}],"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":1066,"alignment_ms":41,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1226,"performance":{"device_seconds":0.8695385689735413,"raw_audio_seconds":3.04,"aggregate_rtf":0.2860324240044544,"audio_seconds_per_second":3.4961071405821706,"batches":[{"batch_size":1,"device_seconds":0.8695385689735413,"audio_seconds":[3.04],"model_wall_seconds":0.9588442742824554,"first_audio_seconds":0.4649798278808594,"graph_capture_seconds":0,"live_callback_seconds":0.08588071912527084}],"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":1241.03,"audio_s":2.83},{"sentence":1,"total_ms":3901.28,"audio_s":10.8},{"sentence":2,"total_ms":1233.76,"audio_s":2.59}],"total_ms":6733.72,"audio_s":16.22,"sentences":3,"generation_rtf":0.4151},"at":42326.59999847412},{"type":"turn_done","barged":false,"at":53388.69999694824}],"reply":"Wie wäre es mit einer kleinen Reisederby? Du nennst mir ein Ziel, und ich erfinde dir eine verrückte, kurze Geschichte, wie wir beide dorthin kommen – vielleicht per fliegendem Koffer oder einem selbstgebauten Waffel-Raketenwagen? Und wie darf ich dich nennen?","captions":[],"pcm_chunks":16,"page_errors":[],"passed":true,"speakers":["Speaker 2"]},{"case":"selma-intro","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":55141.69999694824},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":55640.79999923706},{"type":"asr_preview","text":"Hey Robbie.","status":"sentence_stream","source":"local_sentence_asr","at":56140.59999847412},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":56520.09999847412},{"type":"asr_preview","text":"Hey Robbie.","status":"sentence_stream","source":"local_sentence_asr","at":56604.19999694824},{"type":"asr_preview","text":"Hey Robby.","status":"sentence_stream","source":"native_asr","at":56633.39999771118},{"type":"asr_preview","text":"Hey Robbie, ich bin Selma.","status":"sentence_stream","source":"local_sentence_asr","at":57366.09999847412},{"type":"asr_preview","text":"Hey Robbie, ich bin Selma.","status":"sentence_stream","source":"native_asr","at":57590},{"status":"available","caption":"A female voice speaks in German. The speech is clear and articulate, with a moderate pace.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.46,"tokens":20,"preprocess_ms":10.92,"encoder_gpu_ms":8.8,"decoder_gpu_ms":141.63,"inference_ms":150.67,"total_ms":161.59,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","type":"sound_caption","request_ms":172.64,"capture_ms":2460,"event_age_ms":2632.65,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":57813.09999847412},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":58040.5},{"type":"asr_preview","text":"Hey Robbie, ich bin Selma.","status":"sentence_stream","source":"local_sentence_asr","at":58092.69999694824},{"type":"asr_sentence","text":"Hey Robbie, ich bin Selma.","audio_s":54.4,"queued":false,"asr":{"text":"Hey Robbie, ich bin Selma.","model":"parakeet_v3","chunks":1,"transcribe_ms":48.93,"speech_pause_ms":700,"observed_quiet_ms":594},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":594,"at":58092.89999771118},{"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":0.2929},"at":58152.79999923706},{"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":21.86,"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":28.9,"added_wait_ms":0,"at":58155.09999847412},{"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":48.93,"speech_pause_ms":700,"observed_quiet_ms":594},"speaker_name":"Selma","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":59.879,"speaker_embedding_wait_ms":59.445,"speaker_lookup_ms":0.434,"speaker_prefetch_used":false,"speaker_prefetch_compute_ms":null,"sentence_queue_ms":60.06,"sentence_audio_s":54.4},"at":58155.19999694824},{"type":"latency_start","reply_id":5,"input":{"kind":"voice","asr":{"text":"Hey Robbie, ich bin Selma.","model":"parakeet_v3","chunks":1,"transcribe_ms":48.93,"speech_pause_ms":700,"observed_quiet_ms":594},"voice_caption_ms":28.9,"voice_caption_status":"available","speaker_assignment_ms":59.879,"speaker_prefetch_compute_ms":null,"speaker_prefetch_used":false,"sentence_queue_ms":60.05},"at":58155.89999771118},{"type":"asr_preview","text":"Sh.","status":"sentence_stream","source":"local_sentence_asr","at":58582.19999694824},{"type":"reply","text":"Aha, Selma!","reply_id":5,"performance":{"sentences":[{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"interest, warm","text":"Aha, 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":"Aha, 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":505.04,"llm_first_sentence_ms":534.87,"sentence_ready_ms":536.06,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":58692.09999847412},{"type":"asr_preview","text":"schön, dass du, dass du, dass du,","status":"sentence_stream","source":"local_sentence_asr","at":59076.09999847412},{"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":505.04,"llm_first_sentence_ms":534.87,"finish_reason":"stop","usage":{"prompt_tokens":2022,"completion_tokens":51,"total_tokens":2073,"cost":0.0020569923,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00207777,"upstream_inference_prompt_cost":0.00200178,"upstream_inference_completions_cost":7.599e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":586.08,"response_complete":true,"answer_first_audio_ms":1253.23,"tts_first_audio_ms":717.02,"first_audio_ms":1253.24},"at":59408.89999771118},{"type":"asr_preview","text":"schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":59596.79999923706},{"type":"reply","text":"Aha, Selma! Schön, dass du dich meldest.","reply_id":5,"performance":{"sentences":[{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"interest, warm","text":"Aha, Selma!","start_s":0,"end_s":1.79},{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"interest, warm","text":"Schön, dass du dich meldest.","start_s":1.79,"end_s":3.49}],"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 meldest.","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":505.04,"llm_first_sentence_ms":534.87,"finish_reason":"stop","usage":{"prompt_tokens":2022,"completion_tokens":51,"total_tokens":2073,"cost":0.0020569923,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00207777,"upstream_inference_prompt_cost":0.00200178,"upstream_inference_completions_cost":7.599e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":586.08,"response_complete":true,"answer_first_audio_ms":1253.23,"tts_first_audio_ms":717.02,"first_audio_ms":1253.24,"tts_server_stages":[{"generation_ms":953,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1093,"performance":{"device_seconds":0.7582847023010255,"raw_audio_seconds":2.56,"aggregate_rtf":0.2962049618363381,"audio_seconds_per_second":3.3760406773757197,"batches":[{"batch_size":1,"device_seconds":0.7582847023010255,"audio_seconds":[2.56],"model_wall_seconds":0.8416454792022705,"first_audio_seconds":0.451051513671875,"graph_capture_seconds":0,"live_callback_seconds":0.08312194794416428}],"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":1102.91,"audio_s":1.79}],"sentence_ready_ms":1640.12,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":59795.69999694824},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":59796.59999847412},{"type":"asr_preview","text":"Schön, dass du da bist.","status":"sentence_stream","source":"local_sentence_asr","at":59856.79999923706},{"type":"asr_sentence","text":"Schön, dass du da bist.","audio_s":56,"queued":true,"asr":{"text":"Schön, dass du da bist.","model":"parakeet_v3","chunks":1,"transcribe_ms":48.38,"speech_pause_ms":700,"observed_quiet_ms":468},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":468,"at":59857.89999771118},{"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":0.397},"at":59858.39999771118},{"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":505.04,"llm_first_sentence_ms":534.87,"finish_reason":"stop","usage":{"prompt_tokens":2022,"completion_tokens":51,"total_tokens":2073,"cost":0.0020569923,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00207777,"upstream_inference_prompt_cost":0.00200178,"upstream_inference_completions_cost":7.599e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":586.08,"response_complete":true,"answer_first_audio_ms":1253.23,"tts_first_audio_ms":717.02,"first_audio_ms":1253.24,"tts_server_stages":[{"generation_ms":953,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1093,"performance":{"device_seconds":0.7582847023010255,"raw_audio_seconds":2.56,"aggregate_rtf":0.2962049618363381,"audio_seconds_per_second":3.3760406773757197,"batches":[{"batch_size":1,"device_seconds":0.7582847023010255,"audio_seconds":[2.56],"model_wall_seconds":0.8416454792022705,"first_audio_seconds":0.451051513671875,"graph_capture_seconds":0,"live_callback_seconds":0.08312194794416428}],"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":1102.91,"audio_s":1.79}]},"at":60528.09999847412},{"type":"reply","text":"Aha, Selma! Schön, dass du dich meldest.\nDie kleine Reisederby von eben – passt sie auch für dich, oder hast du eine ganz andere Idee im Kopf?","reply_id":5,"performance":{"sentences":[{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"interest, warm","text":"Aha, Selma!","start_s":0,"end_s":1.79},{"emotion":"interest","secondary":"affection","blend":0.25,"intensity":0.72,"direction":"interest, warm","text":"Schön, dass du dich meldest.","start_s":1.79,"end_s":4.38},{"emotion":"amusement","secondary":null,"blend":0,"intensity":0.55,"direction":"playful, soft","text":"Die kleine Reisederby von eben – passt sie auch für dich, oder hast du eine ganz andere Idee im Kopf?","start_s":4.38,"end_s":11.18}],"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":"Die kleine Reisederby von eben – passt sie auch für dich, oder hast du eine ganz andere Idee im Kopf?","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":505.04,"llm_first_sentence_ms":534.87,"finish_reason":"stop","usage":{"prompt_tokens":2022,"completion_tokens":51,"total_tokens":2073,"cost":0.0020569923,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00207777,"upstream_inference_prompt_cost":0.00200178,"upstream_inference_completions_cost":7.599e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":586.08,"response_complete":true,"answer_first_audio_ms":1253.23,"tts_first_audio_ms":717.02,"first_audio_ms":1253.24,"tts_server_stages":[{"generation_ms":953,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1093,"performance":{"device_seconds":0.7582847023010255,"raw_audio_seconds":2.56,"aggregate_rtf":0.2962049618363381,"audio_seconds_per_second":3.3760406773757197,"batches":[{"batch_size":1,"device_seconds":0.7582847023010255,"audio_seconds":[2.56],"model_wall_seconds":0.8416454792022705,"first_audio_seconds":0.451051513671875,"graph_capture_seconds":0,"live_callback_seconds":0.08312194794416428}],"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":989,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1140,"performance":{"device_seconds":0.7915084390640259,"raw_audio_seconds":2.72,"aggregate_rtf":0.29099574965589187,"audio_seconds_per_second":3.4364763099891302,"batches":[{"batch_size":1,"device_seconds":0.7915084390640259,"audio_seconds":[2.72],"model_wall_seconds":0.8808960765600204,"first_audio_seconds":0.45344857788085935,"graph_capture_seconds":0,"live_callback_seconds":0.08622851222753525}],"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":1102.91,"audio_s":1.79},{"sentence":1,"total_ms":1152.08,"audio_s":2.59}],"sentence_ready_ms":2793.27,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"delivery.playful.soft"},"sources":[],"lookups":[],"at":60958.79999923706},{"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":505.04,"llm_first_sentence_ms":534.87,"finish_reason":"stop","usage":{"prompt_tokens":2022,"completion_tokens":51,"total_tokens":2073,"cost":0.0020569923,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00207777,"upstream_inference_prompt_cost":0.00200178,"upstream_inference_completions_cost":7.599e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":586.08,"response_complete":true,"answer_first_audio_ms":1253.23,"tts_first_audio_ms":717.02,"first_audio_ms":1253.24,"tts_server_stages":[{"generation_ms":953,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1093,"performance":{"device_seconds":0.7582847023010255,"raw_audio_seconds":2.56,"aggregate_rtf":0.2962049618363381,"audio_seconds_per_second":3.3760406773757197,"batches":[{"batch_size":1,"device_seconds":0.7582847023010255,"audio_seconds":[2.56],"model_wall_seconds":0.8416454792022705,"first_audio_seconds":0.451051513671875,"graph_capture_seconds":0,"live_callback_seconds":0.08312194794416428}],"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":989,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1140,"performance":{"device_seconds":0.7915084390640259,"raw_audio_seconds":2.72,"aggregate_rtf":0.29099574965589187,"audio_seconds_per_second":3.4364763099891302,"batches":[{"batch_size":1,"device_seconds":0.7915084390640259,"audio_seconds":[2.72],"model_wall_seconds":0.8808960765600204,"first_audio_seconds":0.45344857788085935,"graph_capture_seconds":0,"live_callback_seconds":0.08622851222753525}],"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":1102.91,"audio_s":1.79},{"sentence":1,"total_ms":1152.08,"audio_s":2.59}]},"at":61691.09999847412},{"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":505.04,"llm_first_sentence_ms":534.87,"finish_reason":"stop","usage":{"prompt_tokens":2022,"completion_tokens":51,"total_tokens":2073,"cost":0.0020569923,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00207777,"upstream_inference_prompt_cost":0.00200178,"upstream_inference_completions_cost":7.599e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":586.08,"response_complete":true,"answer_first_audio_ms":1253.23,"tts_first_audio_ms":717.02,"first_audio_ms":1253.24,"tts_server_stages":[{"generation_ms":953,"alignment_ms":36,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1093,"performance":{"device_seconds":0.7582847023010255,"raw_audio_seconds":2.56,"aggregate_rtf":0.2962049618363381,"audio_seconds_per_second":3.3760406773757197,"batches":[{"batch_size":1,"device_seconds":0.7582847023010255,"audio_seconds":[2.56],"model_wall_seconds":0.8416454792022705,"first_audio_seconds":0.451051513671875,"graph_capture_seconds":0,"live_callback_seconds":0.08312194794416428}],"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":989,"alignment_ms":37,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1140,"performance":{"device_seconds":0.7915084390640259,"raw_audio_seconds":2.72,"aggregate_rtf":0.29099574965589187,"audio_seconds_per_second":3.4364763099891302,"batches":[{"batch_size":1,"device_seconds":0.7915084390640259,"audio_seconds":[2.72],"model_wall_seconds":0.8808960765600204,"first_audio_seconds":0.45344857788085935,"graph_capture_seconds":0,"live_callback_seconds":0.08622851222753525}],"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":2815,"alignment_ms":40,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":2965,"performance":{"device_seconds":2.3488244409561156,"raw_audio_seconds":8.4,"aggregate_rtf":0.27962195725668043,"audio_seconds_per_second":3.576257064398004,"batches":[{"batch_size":1,"device_seconds":2.3488244409561156,"audio_seconds":[8.4],"model_wall_seconds":2.706190451979637,"first_audio_seconds":0.4632595520019531,"graph_capture_seconds":0,"live_callback_seconds":0.3568967282772064}],"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":1102.91,"audio_s":1.79},{"sentence":1,"total_ms":1152.08,"audio_s":2.59},{"sentence":2,"total_ms":2981.27,"audio_s":8.4}],"total_ms":5774.77,"audio_s":12.78,"sentences":3,"generation_rtf":0.4519},"at":63932.79999923706},{"type":"turn_done","barged":false,"at":72656.39999771118},{"type":"voice_caption","status":"available","caption":"Possibly very smooth, steady, slow, very liberation","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":1.49,"inference_ms":19.97,"model":"laion/humaneness-ears-base-medium","model_revision":"818506970d93c809a3295a0a02c40dee8ff2bfc3","rules_revision":"7ebd4a9b7da96df989af8fc6ad0ce34f4e8aa3ed","device":"cuda:1","precision":"bfloat16","selected":{"voicenet":[{"dim":"DFLU","name":"Disfluency","value":-0.701,"z":-2.01,"direction":"below","intensity":"Extremely","always_on":false,"phrase":"extremely smooth and flawless","tag":"very smooth","reliability":1,"rank_score":2.012},{"dim":"VOLT","name":"Volatility","value":1.015,"z":-1.36,"direction":"below","intensity":"Notably","always_on":false,"phrase":"notably emotionally steady and static","tag":"steady","reliability":1,"rank_score":1.356},{"dim":"TEMP","name":"Tempo","value":2.259,"z":-1.34,"direction":"below","intensity":"Notably","always_on":true,"phrase":"notably slow and deliberate in tempo","tag":"slow","reliability":1,"rank_score":1.344}],"emotions":[{"emotion":"Relief","value":2.496,"z":4.95,"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":28.18,"added_wait_ms":0,"at":72659.29999923706},{"type":"endpoint","transcript":"Schön, dass du da bist.","speaker_id":"speaker-2","asr":{"text":"Schön, dass du da bist.","model":"parakeet_v3","chunks":1,"transcribe_ms":48.38,"speech_pause_ms":700,"observed_quiet_ms":468},"speaker_name":"Selma","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":0.657,"speaker_embedding_wait_ms":0.002,"speaker_lookup_ms":0.655,"speaker_prefetch_used":true,"speaker_prefetch_compute_ms":44.49,"sentence_queue_ms":12806.38,"sentence_audio_s":56},"at":72659.39999771118},{"type":"latency_start","reply_id":6,"input":{"kind":"voice","asr":{"text":"Schön, dass du da bist.","model":"parakeet_v3","chunks":1,"transcribe_ms":48.38,"speech_pause_ms":700,"observed_quiet_ms":468},"voice_caption_ms":28.18,"voice_caption_status":"available","speaker_assignment_ms":0.657,"speaker_prefetch_compute_ms":44.49,"speaker_prefetch_used":true,"sentence_queue_ms":12806.37},"at":72659.89999771118},{"type":"reply","text":"Auch schön, dass du da bist, Selma!","reply_id":6,"performance":{"sentences":[{"emotion":"affection","secondary":null,"blend":0,"intensity":0.72,"direction":"happy, warm","text":"Auch schön, dass du da bist, Selma!","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":"Auch schön, dass du da bist, Selma!","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":401.31,"finish_reason":"stop","usage":{"prompt_tokens":2103,"completion_tokens":17,"total_tokens":2120,"cost":0.002086227,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0021073,"upstream_inference_prompt_cost":0.00208197,"upstream_inference_completions_cost":2.533e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":428.34,"generation_calls":1,"llm_total_ms":428.37,"response_complete":true,"sentence_ready_ms":429.47,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"elation.soft"},"sources":[],"lookups":[],"at":73091.89999771118},{"type":"latency_update","reply_id":6,"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":401.31,"finish_reason":"stop","usage":{"prompt_tokens":2103,"completion_tokens":17,"total_tokens":2120,"cost":0.002086227,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0021073,"upstream_inference_prompt_cost":0.00208197,"upstream_inference_completions_cost":2.533e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":428.34,"generation_calls":1,"llm_total_ms":428.37,"response_complete":true,"answer_first_audio_ms":1184.42,"tts_first_audio_ms":754.78,"first_audio_ms":1184.43},"at":73846.09999847412},{"type":"pipeline_metrics","reply_id":6,"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":401.31,"finish_reason":"stop","usage":{"prompt_tokens":2103,"completion_tokens":17,"total_tokens":2120,"cost":0.002086227,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.0021073,"upstream_inference_prompt_cost":0.00208197,"upstream_inference_completions_cost":2.533e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":428.34,"generation_calls":1,"llm_total_ms":428.37,"response_complete":true,"answer_first_audio_ms":1184.42,"tts_first_audio_ms":754.78,"first_audio_ms":1184.43,"tts_server_stages":[{"generation_ms":1115,"alignment_ms":39,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1277,"performance":{"device_seconds":0.9033443279266358,"raw_audio_seconds":3.2,"aggregate_rtf":0.2822951024770737,"audio_seconds_per_second":3.5423923094139194,"batches":[{"batch_size":1,"device_seconds":0.9033443279266358,"audio_seconds":[3.2],"model_wall_seconds":1.002142608165741,"first_audio_seconds":0.46129867553710935,"graph_capture_seconds":0,"live_callback_seconds":0.0983913466334343}],"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":1287.16,"audio_s":2.35}],"total_ms":1716.81,"audio_s":2.35,"sentences":1,"generation_rtf":0.7306},"at":74391},{"type":"turn_done","barged":false,"at":76655}],"reply":"Auch schön, dass du da bist, Selma!","captions":[{"status":"available","caption":"A female voice speaks in German. The speech is clear and articulate, with a moderate pace.","uncertainty":"Fallible sound hypothesis, not a transcript or confirmed event.","window_s":2.46,"tokens":20,"preprocess_ms":10.92,"encoder_gpu_ms":8.8,"decoder_gpu_ms":141.63,"inference_ms":150.67,"total_ms":161.59,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","type":"sound_caption","request_ms":172.64,"capture_ms":2460,"event_age_ms":2632.65,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":57813.09999847412}],"pcm_chunks":19,"page_errors":[],"passed":true,"speakers":["Selma","Selma"]},{"case":"selma-followup-clear","events":[{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":79742.19999694824},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":80247},{"type":"asr_preview","text":"Was können können","status":"sentence_stream","source":"native_asr","at":80536.79999923706},{"type":"asr_preview","text":"Was können wir heute zusammen?","status":"sentence_stream","source":"local_sentence_asr","at":81271.69999694824},{"type":"asr_preview","text":"Was können wir heute zusammen","status":"sentence_stream","source":"native_asr","at":81497.59999847412},{"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":11.01,"encoder_gpu_ms":9.13,"decoder_gpu_ms":177.89,"inference_ms":187.29,"total_ms":198.3,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","type":"sound_caption","request_ms":210.39,"capture_ms":2220,"event_age_ms":2430.39,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":82098.79999923706},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":82279.29999923706},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":82359.29999923706},{"type":"asr_preview","text":"Was können wir heute zusammen machen?","status":"sentence_stream","source":"local_sentence_asr","at":82423.5},{"type":"asr_sentence","text":"Was können wir heute zusammen machen?","audio_s":78.08,"queued":false,"asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":49.54,"speech_pause_ms":700,"observed_quiet_ms":686},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":686,"at":82423.5},{"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":0.994},"at":82423.79999923706},{"type":"voice_caption","status":"available","caption":"Possibly smooth, slightly fading, very hope","uncertainty":"Uncalibrated acoustic estimates; noise, other speakers and acting may be misleading.","voice_impression_only":false,"window_s":2.64,"inference_ms":21.57,"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":"DARC","name":"Dynamic Arc","value":2.682,"z":-0.78,"direction":"below","intensity":"Somewhat","always_on":false,"phrase":"somewhat fading in loudness","tag":"slightly fading","reliability":1,"rank_score":0.775},{"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":2.139,"z":1.97,"direction":"above","intensity":"Very","synonym":"hope","phrase":"very carrying hope","tag":"very hope"}]},"normalization":"speech reference z-scores; not probabilities","scope":"this turn","request_ms":33.11,"added_wait_ms":0,"at":82424},{"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":49.54,"speech_pause_ms":700,"observed_quiet_ms":686},"speaker_name":"Selma","by":"asr_complete_sentence","timing":{"speaker_assignment_ms":0.347,"speaker_embedding_wait_ms":0.001,"speaker_lookup_ms":0.346,"speaker_prefetch_used":true,"speaker_prefetch_compute_ms":92.61,"sentence_queue_ms":0.47,"sentence_audio_s":78.08},"at":82424.09999847412},{"type":"latency_start","reply_id":7,"input":{"kind":"voice","asr":{"text":"Was können wir heute zusammen machen?","model":"parakeet_v3","chunks":1,"transcribe_ms":49.54,"speech_pause_ms":700,"observed_quiet_ms":686},"voice_caption_ms":33.11,"voice_caption_status":"available","speaker_assignment_ms":0.347,"speaker_prefetch_compute_ms":92.61,"speaker_prefetch_used":true,"sentence_queue_ms":0.47},"at":82424.59999847412},{"type":"reply","text":"Wie wäre es mit einer kreativen Wortreise?","reply_id":7,"performance":{"sentences":[{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Wie wäre es mit einer kreativen Wortreise?","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 kreativen Wortreise?","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":418.69,"llm_first_sentence_ms":444.61,"sentence_ready_ms":445.9,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":82872.69999694824},{"type":"latency_update","reply_id":7,"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":418.69,"llm_first_sentence_ms":444.61,"finish_reason":"stop","usage":{"prompt_tokens":2156,"completion_tokens":46,"total_tokens":2202,"cost":0.0021809502,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00220298,"upstream_inference_prompt_cost":0.00213444,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":499.2,"response_complete":true,"answer_first_audio_ms":1200.14,"tts_first_audio_ms":754.07,"first_audio_ms":1200.15},"at":83623.79999923706},{"type":"reply","text":"Wie wäre es mit einer kreativen Wortreise? Wir erfinden gemeinsam eine kleine Geschichte, in der wir auf einem riesigen, fliegenden Koffer durch eine Wolkenstadt reisen!","reply_id":7,"performance":{"sentences":[{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Wie wäre es mit einer kreativen Wortreise?","start_s":0,"end_s":3.04},{"emotion":"interest","secondary":"amusement","blend":0.25,"intensity":0.72,"direction":"interest, playful","text":"Wir erfinden gemeinsam eine kleine Geschichte, in der wir auf einem riesigen, fliegenden Koffer durch eine Wolkenstadt reisen!","start_s":3.04,"end_s":9.16}],"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":"Wir erfinden gemeinsam eine kleine Geschichte, in der wir auf einem riesigen, fliegenden Koffer durch eine Wolkenstadt reisen!","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":418.69,"llm_first_sentence_ms":444.61,"finish_reason":"stop","usage":{"prompt_tokens":2156,"completion_tokens":46,"total_tokens":2202,"cost":0.0021809502,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00220298,"upstream_inference_prompt_cost":0.00213444,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":499.2,"response_complete":true,"answer_first_audio_ms":1200.14,"tts_first_audio_ms":754.07,"first_audio_ms":1200.15,"tts_server_stages":[{"generation_ms":1102,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1252,"performance":{"device_seconds":0.8890595293045044,"raw_audio_seconds":3.04,"aggregate_rtf":0.2924537925343764,"audio_seconds_per_second":3.4193435870128277,"batches":[{"batch_size":1,"device_seconds":0.8890595293045044,"audio_seconds":[3.04],"model_wall_seconds":0.9888429641723633,"first_audio_seconds":0.4720425109863281,"graph_capture_seconds":0,"live_callback_seconds":0.0975344255566597}],"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":1266.74,"audio_s":3.04}],"sentence_ready_ms":1713.62,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":84148.5},{"type":"latency_update","reply_id":7,"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":418.69,"llm_first_sentence_ms":444.61,"finish_reason":"stop","usage":{"prompt_tokens":2156,"completion_tokens":46,"total_tokens":2202,"cost":0.0021809502,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00220298,"upstream_inference_prompt_cost":0.00213444,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":499.2,"response_complete":true,"answer_first_audio_ms":1200.14,"tts_first_audio_ms":754.07,"first_audio_ms":1200.15,"tts_server_stages":[{"generation_ms":1102,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1252,"performance":{"device_seconds":0.8890595293045044,"raw_audio_seconds":3.04,"aggregate_rtf":0.2924537925343764,"audio_seconds_per_second":3.4193435870128277,"batches":[{"batch_size":1,"device_seconds":0.8890595293045044,"audio_seconds":[3.04],"model_wall_seconds":0.9888429641723633,"first_audio_seconds":0.4720425109863281,"graph_capture_seconds":0,"live_callback_seconds":0.0975344255566597}],"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":1266.74,"audio_s":3.04}]},"at":84905.39999771118},{"type":"pipeline_metrics","reply_id":7,"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":418.69,"llm_first_sentence_ms":444.61,"finish_reason":"stop","usage":{"prompt_tokens":2156,"completion_tokens":46,"total_tokens":2202,"cost":0.0021809502,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00220298,"upstream_inference_prompt_cost":0.00213444,"upstream_inference_completions_cost":6.854e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"generation_calls":1,"llm_total_ms":499.2,"response_complete":true,"answer_first_audio_ms":1200.14,"tts_first_audio_ms":754.07,"first_audio_ms":1200.15,"tts_server_stages":[{"generation_ms":1102,"alignment_ms":38,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1252,"performance":{"device_seconds":0.8890595293045044,"raw_audio_seconds":3.04,"aggregate_rtf":0.2924537925343764,"audio_seconds_per_second":3.4193435870128277,"batches":[{"batch_size":1,"device_seconds":0.8890595293045044,"audio_seconds":[3.04],"model_wall_seconds":0.9888429641723633,"first_audio_seconds":0.4720425109863281,"graph_capture_seconds":0,"live_callback_seconds":0.0975344255566597}],"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":2512,"alignment_ms":40,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":2666,"performance":{"device_seconds":2.049631987094879,"raw_audio_seconds":6.96,"aggregate_rtf":0.294487354467655,"audio_seconds_per_second":3.395731547820451,"batches":[{"batch_size":1,"device_seconds":2.049631987094879,"audio_seconds":[6.96],"model_wall_seconds":2.4010032191872597,"first_audio_seconds":0.4852572021484375,"graph_capture_seconds":0,"live_callback_seconds":0.3475804552435875}],"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":1266.74,"audio_s":3.04},{"sentence":1,"total_ms":2680.3,"audio_s":6.96}],"total_ms":4394.07,"audio_s":10,"sentences":2,"generation_rtf":0.4394},"at":86825.59999847412},{"type":"turn_done","barged":false,"at":94073.39999771118}],"reply":"Wie wäre es mit einer kreativen Wortreise? Wir erfinden gemeinsam eine kleine Geschichte, in der wir auf einem riesigen, fliegenden Koffer durch eine Wolkenstadt reisen!","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":11.01,"encoder_gpu_ms":9.13,"decoder_gpu_ms":177.89,"inference_ms":187.29,"total_ms":198.3,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","type":"sound_caption","request_ms":210.39,"capture_ms":2220,"event_age_ms":2430.39,"kind":"voice","eligible":false,"suppression_reason":"speech_or_quiet","at":82098.79999923706}],"pcm_chunks":11,"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":21254.699996948242},{"type":"asr_preview","text":"","status":"sentence_stream","source":"local_sentence_asr","at":21729.699996948242},{"type":"asr_preview","text":"Morgen wird es sonnig.","status":"sentence_stream","source":"local_sentence_asr","at":22241.89999771118},{"type":"asr_preview","text":"Morgen wird es sonnig.","status":"sentence_stream","source":"native_asr","at":22733.89999771118},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":23011.59999847412},{"type":"asr_preview","text":"Morgen wird es sonnig.","status":"sentence_stream","source":"local_sentence_asr","at":23072.09999847412},{"type":"asr_sentence","text":"Morgen wird es sonnig.","audio_s":19.84,"queued":false,"asr":{"text":"Morgen wird es sonnig.","model":"parakeet_v3","chunks":1,"transcribe_ms":45.84,"speech_pause_ms":700,"observed_quiet_ms":754},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":754,"at":23072.199996948242},{"type":"speaker_focus","operation":"turn","accepted":false,"source":"background_salience","relative_level":0.129,"known_voice_similarity":0.0884,"explicit_address":false,"estimated_level_above_floor_db":100,"reason":"Very quiet, unmatched speech without an explicit address","confidence_is_calibrated":false,"at":23128.29999923706},{"type":"ignored","accepted":false,"source":"background_salience","relative_level":0.129,"known_voice_similarity":0.0884,"explicit_address":false,"estimated_level_above_floor_db":100,"reason":"Very quiet, unmatched speech without an explicit address","confidence_is_calibrated":false,"at":23128.39999771118},{"type":"asr_preview","text":"Im Hintergrund","status":"sentence_stream","source":"local_sentence_asr","at":23734.699996948242},{"type":"asr_preview","text":"Im Hintergrund","status":"sentence_stream","source":"local_sentence_asr","at":24195.79999923706},{"type":"asr_preview","text":"Im Hintergrund","status":"sentence_stream","source":"native_asr","at":24334.29999923706},{"type":"asr_preview","text":"Im Hintergrund läuft ein Bericht über das","status":"sentence_stream","source":"local_sentence_asr","at":25066.699996948242},{"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":16.2,"encoder_gpu_ms":15.62,"decoder_gpu_ms":362.08,"inference_ms":377.96,"total_ms":394.16,"model":"laion/sound-effect-captioning-whisper","revision":"bb36f9c6534037a8d1d78cff97d119dd37da478a","device":"cuda:0","precision":"float16","type":"sound_caption","request_ms":402.38,"capture_ms":1640,"event_age_ms":2042.38,"kind":"sound","eligible":true,"suppression_reason":null,"at":25098.199996948242},{"type":"asr_preview","text":"Im Hintergrund läuft ein Bericht über das","status":"sentence_stream","source":"native_asr","at":25285},{"type":"asr_preview","text":"Im Hintergrund läuft ein Bericht über das Wetter und die","status":"sentence_stream","source":"local_sentence_asr","at":26295.09999847412},{"type":"asr_preview","text":"Im Hintergrund läuft ein Bericht über das Wetter und die Nachrichten","status":"sentence_stream","source":"local_sentence_asr","at":26780.09999847412},{"type":"asr_preview","text":"Und die Nachrichten","status":"sentence_stream","source":"native_asr","at":26908.59999847412},{"type":"asr_preview","text":"Im Hintergrund läuft ein Bericht über das Wetter und die Nachrichten","status":"sentence_stream","source":"local_sentence_asr","at":27661.29999923706},{"type":"asr_finalizing","model":"parakeet_v3","label":"Transcribing the completed speech segment…","at":27742.79999923706},{"type":"asr_preview","text":"Im Hintergrund läuft ein Bericht über das Wetter. Und die Nachrichten, die wichtig wetter.","status":"sentence_stream","source":"local_sentence_asr","at":27824.89999771118},{"type":"asr_sentence","text":"Im Hintergrund läuft ein Bericht über das Wetter. Und die Nachrichten, die wichtig wetter.","audio_s":24.64,"queued":false,"asr":{"text":"Im Hintergrund läuft ein Bericht über das Wetter. Und die Nachrichten, die wichtig wetter.","model":"parakeet_v3","chunks":1,"transcribe_ms":75.57,"speech_pause_ms":700,"observed_quiet_ms":792},"reason":"neural speech pause","speech_pause_ms":700,"observed_quiet_ms":792,"at":27825.199996948242},{"type":"environment_observation","kind":"sound","status":"responding","at":27876.39999771118},{"type":"latency_start","reply_id":3,"input":{"kind":"environment","sound_caption_ms":402.38},"at":27876.59999847412},{"type":"speaker_focus","operation":"turn","accepted":false,"source":"background_salience","relative_level":0.101,"known_voice_similarity":0.1725,"explicit_address":false,"estimated_level_above_floor_db":98,"reason":"Very quiet, unmatched speech without an explicit address","confidence_is_calibrated":false,"at":27881.89999771118},{"type":"ignored","accepted":false,"source":"background_salience","relative_level":0.101,"known_voice_similarity":0.1725,"explicit_address":false,"estimated_level_above_floor_db":98,"reason":"Very quiet, unmatched speech without an explicit address","confidence_is_calibrated":false,"at":27881.89999771118},{"type":"reply","text":"War das gerade ein kleiner Piepton?","reply_id":3,"performance":{"sentences":[{"emotion":"interest","secondary":null,"blend":0,"intensity":0.55,"direction":"interest, soft","text":"War das gerade ein kleiner Piepton?","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":"War das gerade ein kleiner Piepton?","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":314.38,"finish_reason":"stop","usage":{"prompt_tokens":2227,"completion_tokens":14,"total_tokens":2241,"cost":0.0022033341,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00222559,"upstream_inference_prompt_cost":0.00220473,"upstream_inference_completions_cost":2.086e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":370.92,"generation_calls":1,"llm_total_ms":370.94,"response_complete":true,"sentence_ready_ms":371.9,"director_mode":"bolt","director_model":"qwen/qwen3.8-27b","template_id":"interest.soft"},"sources":[],"lookups":[],"at":28244.59999847412},{"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":314.38,"finish_reason":"stop","usage":{"prompt_tokens":2227,"completion_tokens":14,"total_tokens":2241,"cost":0.0022033341,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00222559,"upstream_inference_prompt_cost":0.00220473,"upstream_inference_completions_cost":2.086e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":370.92,"generation_calls":1,"llm_total_ms":370.94,"response_complete":true,"answer_first_audio_ms":1122.77,"tts_first_audio_ms":750.73,"first_audio_ms":1122.78},"at":28993.39999771118},{"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":314.38,"finish_reason":"stop","usage":{"prompt_tokens":2227,"completion_tokens":14,"total_tokens":2241,"cost":0.0022033341,"is_byok":false,"prompt_tokens_details":{"cached_tokens":1024,"cache_write_tokens":0,"audio_tokens":0,"video_tokens":0},"cost_details":{"upstream_inference_cost":0.00222559,"upstream_inference_prompt_cost":0.00220473,"upstream_inference_completions_cost":2.086e-05},"completion_tokens_details":{"reasoning_tokens":0,"image_tokens":0,"audio_tokens":0}},"llm_first_sentence_ms":370.92,"generation_calls":1,"llm_total_ms":370.94,"response_complete":true,"answer_first_audio_ms":1122.77,"tts_first_audio_ms":750.73,"first_audio_ms":1122.78,"tts_server_stages":[{"generation_ms":1086,"alignment_ms":40,"scoring_ms":0,"speaker_preparation_ms":0,"total_ms":1249,"performance":{"device_seconds":0.8771105589866639,"raw_audio_seconds":3.04,"aggregate_rtf":0.28852321019298155,"audio_seconds_per_second":3.4659256679250876,"batches":[{"batch_size":1,"device_seconds":0.8771105589866639,"audio_seconds":[3.04],"model_wall_seconds":0.9726785719394684,"first_audio_seconds":0.45694873046875,"graph_capture_seconds":0,"live_callback_seconds":0.09530941396951675}],"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":1256.67,"audio_s":2.91}],"total_ms":1628.73,"audio_s":2.91,"sentences":1,"generation_rtf":0.5597},"at":29500.09999847412}],"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":0.994},"at":82423.79999923706},"named_speakers_present":true,"source_unchanged":true}