| 1 | # Whisper STT |
| 2 | |
| 3 | Built-in speech-to-text plugin backed by Whisper. |
| 4 | |
| 5 | ## Responsibilities |
| 6 | |
| 7 | - Registers Whisper as the active STT provider when the plugin is enabled. |
| 8 | - Owns the microphone runtime, device selector UI, message delivery mode, and plugin APIs. |
| 9 | - Keeps dependency installation and model bootstrap on the Docker/bootstrap path. |
| 10 | |
| 11 | ## Config |
| 12 | |
| 13 | - `model_size`: Whisper model name |
| 14 | - `language`: language hint or `auto` |
| 15 | - `message_mode`: `send` to send final transcriptions immediately, or `draft` to leave them in the composer |
| 16 | - `silence_threshold`: frontend threshold before recording starts |
| 17 | - `silence_duration`: silence window before waiting state |
| 18 | - `waiting_timeout`: delay before transcription dispatch |
| 19 | |
| 20 | ## API |
| 21 | |
| 22 | - `POST /api/plugins/_whisper_stt/transcribe` |
| 23 | - `POST /api/plugins/_whisper_stt/status` |