Skip to content

STT Whisper

Whisper.cpp based VoiceDock STT implementation. Provides gRPC API for high quality speech-to-text (from raw PCM stream) based on Whisper.cpp project. Provides download of new language packs via API.

Features

  • Browse and download language packs (models in ggml format)
  • Speech to text conversion for 99+ languages
  • Automatic language recognition
  • GPU support

Installation

Create directories for model data and configuration: 

mkdir dataset
mkdir config
Create a config/sttwhisper.json configuration file or download an example configuration: 
curl -o config/sttwhisper.json https://raw.githubusercontent.com/voicedock/sttwhisper/main/config/sttwhisper.json


docker run --rm \
    -v "$(pwd)/config:/data/config" \
    -v "$(pwd)/dataset:/data/dataset" \
    -p 9999:9999 \
    ghcr.io/voicedock/sttwhisper:latest sttwhisper

docker run --rm --gpus all \
    -v "$(pwd)/config:/data/config" \
    -v "$(pwd)/dataset:/data/dataset" \
    -p 9999:9999 \
    ghcr.io/voicedock/sttwhisper:gpu sttwhisper

Configuration

See on Github.

Languages support

en, zh, de, es, ru, ko, fr, ja, pt, tr, pl, ca, nl, ar, sv, it, id, hi, fi, vi, iw, uk, el, ms, cs, ro, da, hu, ta, no, th, ur, hr, bg, lt, la, mi, ml, cy, sk, te, fa, lv, bn, sr, az, sl, kn, et, mk, br, eu, is, hy, ne, mn, bs, kk, sq, sw, gl, mr, pa, si, km, sn, yo, so, af, oc, ka, be, tg, sd, gu, am, yi, lo, uz, fo, ht, ps, tk, nn, mt, sa, lb, my, bo, tl, mg, as, tt, haw, ln, ha, ba, jw.