vosk-server

Why we included this project

Vosk-server puts the Vosk offline speech engine behind a network interface, so you can run it as a service rather than embedding it in a single app. It ships ready-made servers for WebSocket, gRPC, MQTT, and WebRTC, letting you pick the transport that fits your setup and stream audio with low latency while keeping recognition on your own hardware. Smart-home voice control, PBX integrations like FreeSWITCH or Asterisk, chatbots, and browser-based apps are all covered, with Docker images and client bindings for several languages. Since it wraps the same Vosk-API as the rest of the ecosystem, the models and accuracy you already know carry over, which saves a lot of work when moving from a prototype to a deployed service.