TTSFM provides a reverse-engineered API server that mirrors OpenAI's TTS service, offering a compatible interface for text-to-speech conversion with multiple voice options. It is intended for learning and testing purposes, not production use, and targets developers needing a local or self-hosted TTS solution.

How It Works

TTSFM utilizes a Flask web framework for its API server and Celery for asynchronous task processing, managing text-to-speech requests via a Redis broker. This architecture allows for efficient handling of potentially long-running TTS generation tasks, offloading them to worker processes and returning results asynchronously. The project aims for full compatibility with OpenAI's TTS API format, enabling seamless integration for users familiar with that ecosystem.

Quick Start & Requirements

• Install/Run: docker pull dbcccc/ttsfm:latest and docker run -d --name ttsfm -p 7000:7000 -p 6379:6379 -v $(pwd)/voices:/app/voices dbcccc/ttsfm:latest
• Prerequisites: Python 3.13+, Redis server, Docker (optional).
• Demo: Official Demo

Highlighted Details

• Fully compatible with OpenAI's TTS API format.
• Supports multiple voices (alloy, ash, ballad, coral, echo, fable, onyx, nova, sage, shimmer, verse).
• Asynchronous task processing via Celery and Redis.
• Includes a basic frontend interface.

Maintenance & Community

The project is actively under development with a v2 branch noted as not recommended for production. Stable documentation is available for v1.

Licensing & Compatibility

Licensed under the MIT License, permitting commercial use and closed-source linking.

Limitations & Caveats

The project explicitly states it is for learning and testing purposes only and not recommended for production use. The v2 branch is under active development.