SEPIA server to support open-source speech recognition via WebSocket connection.
coqui-model-en-1.0.0
(English Coqui model v1.0.0 w/o scorer for testing) - Apache 2.0 licenseThis is a complete rework of the 2018 STT server with a lot of new features, better performance and even smaller size ^^. This new server will be fully supported by the upcoming v0.24.0 of the SEPIA Client. Please see the README for more info and a script to convert your old STT models.
vosk-model-small-en-us-0.15
- Lightweight wideband model for Android and RPi - Apache 2.0 licensevosk-model-small-de-0.15
- Lightweight wideband model for Android and RPi - Apache 2.0 licensevosk-model-spk-0.4
- Model for speaker identification (all models) - Apache 2.0 license