Kokoro Serve

A lightweight Text-to-Speech server powered by Kokoro TTS.

Usage

Run with uv

uv run -m uvicorn main:app --host 0.0.0.0 --port 8000

Run with Docker

docker build -t kokoro-serve .
docker run -p 8000:8000 kokoro-serve

API

GET /voices

POST /generate
    body:
        {
            "text": "string"
            "voice": "string"
        }

Examples

# get available voices
curl -XGET http://localhost:8000/voices

# run synthesis
curl -XPOST http://localhost:8000/generate \
    -H "Content-Type: application/json" \
    -d '{"text":"Hello from Kokoro Serve!", "voice": "af_heart"}' \
    -o hello.wav