README.md

metadata

title: Text-to-Speech API
emoji: 🗣️
colorFrom: blue
colorTo: purple
sdk: docker
app_file: app.py
pinned: false

Text-to-Speech API with Coqui TTS

A minimal Text-to-Speech API built with FastAPI and Coqui TTS VITS model, designed for Hugging Face Spaces.

Features

• High-Quality TTS: Uses Coqui's VITS model (tts_models/en/ljspeech/vits)
• Simple API: Clean GET and POST endpoints
• Automatic Model Loading: Downloads model automatically on first startup
• CPU Optimized: Runs efficiently on CPU-only environments
• Browser Compatible: Returns WAV files playable in browsers

Model Information

This app uses the LJSpeech VITS model from Coqui TTS:

• Model: tts_models/en/ljspeech/vits
• Language: English
• Voice: Female (LJSpeech dataset)
• Quality: High-quality neural synthesis

API Usage

Simple GET Request

curl "https://your-space-url/tts?text=Hello%20world"

POST with JSON

curl -X POST "https://your-space-url/tts" \
  -H "Content-Type: application/json" \
  -d '{"text": "Hello world, this is a test."}'

POST with Form Data

curl -X POST "https://your-space-url/tts" \
  -F "text=Hello world"

Endpoints

• GET / - Health check and model information
• GET /tts - Text-to-speech conversion (GET)
• POST /tts - Text-to-speech conversion (POST)
• GET /health - Detailed health status

Response

All TTS endpoints return a WAV audio file that can be:

• Played directly in web browsers
• Downloaded as speech.wav
• Used in audio players and applications

Local Development

# Install dependencies
pip install -r requirements.txt

# Run the application
python app.py

The API will be available at http://localhost:7860

Example Usage in Browser

You can test the API directly in your browser:

https://your-space-url/tts?text=Welcome%20to%20the%20text%20to%20speech%20API

Deployment Notes

• Hugging Face Spaces: Optimized for CPU-only inference
• Model Loading: VITS model downloads automatically (~50MB)
• Memory Usage: Approximately 500MB RAM
• Response Time: 2-5 seconds for typical sentences

Error Handling

The API provides clear error messages for:

• Missing or empty text input
• Model loading failures
• Audio generation errors