ComfyUI Node
Speaker Diarizer (Up to 5 Speakers)
A ComfyUI node in Audio/Isolation with 10 inputs and 7 outputs.
Speaker Diarizer (Up to 5 Speakers)
- audio
- speaker_1_audio
- speaker_2_audio
- speaker_3_audio
- speaker_4_audio
- speaker_5_audio
- summary
- segments_json
◄hf_token►
◄deviceauto►
◄min_speakers1►
◄max_speakers5►
◄merge_gap_ms200►
◄keep_only_detected_speakerstrue►
◄pyannote_python►
◄worker_script►
◄timeout_seconds600►
CategoryAudio/Isolation
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| audio | AUDIO | — | |
| hf_token | STRING | Hugging Face token for pyannote community-1 | |
| device | COMBO | auto | Device used by the external pyannote worker |
| min_speakers | INT | 11–5 | Minimum expected speakers |
| max_speakers | INT | 51–5 | Maximum expected speakers |
| merge_gap_ms | INT | 2000–5000 | Merge adjacent same-speaker segments up to this gap |
| keep_only_detected_speakers | BOOLEAN | true | Kept for compatibility. Output is timeline-preserving. |
| pyannote_python | STRING | Absolute path to Python executable of the external pyannote environment | |
| worker_script | STRING | Absolute path to pyannote_worker.py | |
| timeout_seconds | INT | 60030–7200 | Timeout for the external worker |
Outputs (7)
| Name | Type | Description |
|---|---|---|
| speaker_1_audio | AUDIO | — |
| speaker_2_audio | AUDIO | — |
| speaker_3_audio | AUDIO | — |
| speaker_4_audio | AUDIO | — |
| speaker_5_audio | AUDIO | — |
| summary | STRING | — |
| segments_json | STRING | — |