Auris — Speech-to-Text

Auris is Synth’s Speech-to-Text subsystem. It transcribes audio files (voice messages, uploads) into text so they can enter the normal message chain and be understood like any other input.

The active engine is selected with ACTIVE_AURIS_ENGINE, chosen from the Auris engine registry (core/auris_registry.py). Engines are interchangeable.

Actions

Action	Purpose
`stt_transcribe`	Transcribe an audio file to text via the active Auris engine.

Configuration

Key	Purpose
`ACTIVE_AURIS_ENGINE`	Active STT engine.
`AURIS_ENGINE_SETTINGS`	Per-engine settings blob.
`VOSK_MODEL_PATH` / `VOSK_LANGUAGE`	Local VOSK model path and language.
`MODEL_AUTO_DOWNLOAD`	Auto-download the STT model when missing.