How to Deploy WhisperLiveKit with Docker: Step-by-Step Tutorial

In this tutorial, you will deploy WhisperLiveKit with Docker on your own server and transcribe your first live audio stream. The whole process takes about ten minutes and requires no GPU for small models — a modern CPU works fine for the tiny and base Whisper variants.

Step 1 — Hardware Requirements

Step 2 — Run the Container

The official setup uses the community image vcure/whisperlivekit:latest. Start it with a single command:

docker run -d --name whisperlivekit \
  -p 8080:8080 \
  -v ./data/whisperlivekit:/data \
  vcure/whisperlivekit:latest

The server exposes its transcription API on port 8080. For a more structured setup, use the generated docker-compose.yml from the tool page, which wires up the volume and restart policy for you.

Step 3 — Verify It Is Running

Check that the health endpoint responds:

curl http://localhost:8080/health

If you see a JSON response, the engine is live. Now open the web UI in your browser at http://localhost:8080, grant microphone access, and start speaking — captions appear in near real time.

Step 4 — Connect Your App

WhisperLiveKit speaks OpenAI- and Deepgram-compatible protocols, so you can point an existing client at http://localhost:8080 and swap in your API key. No SDK changes required — your app now runs on your own infrastructure.

Review Verdict

WhisperLiveKit is an excellent drop-in replacement for cloud transcription services. Deploy it once and you have unlimited, private, real-time speech-to-text.