Our take
Quietno commits in the last 90 days- Activity
- 45
- Community
- 25
- Issues
- 21
- Pull requests
- 21
-1 stars in the last 30 days
TTSFM is a practical compatibility layer for experiments, but its reverse-engineered backend, quiet recent activity, and highly concentrated contributor base make it a poor choice for production dependency.
Good fit if
- You need an OpenAI-shaped TTS interface for a prototype or educational project.
- You use Python or Docker and can work within the backend and format capabilities documented by the project.
- Your use is non-commercial and you are comfortable with a project whose recent commit and issue-handling activity is quiet.
Look elsewhere if
- You need a supported production or commercial speech service, which the README explicitly advises against.
- Your deployment needs predictable upstream behavior or active maintenance; no commits, merged pull requests, or closed issues were recorded in the last 90 days, and most contributions come from one person.
- You need advanced formats or speed adjustment in the slim image, which does not include ffmpeg.
All health signals
| Last commit | 2026-06-30 (3 months ago) |
|---|---|
| Commits, last 90 days | 0 |
| Releases, last 12 months | 16 (latest v3.5.0, 2026-05-05) |
| Contributors | 7 (top contributor: 93% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 19 months |
Checked on 2026-10-11 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
738Forks
113Change since Nov 8, 2026: 0 stars·0 forks
Overview
TTSFM offers a Python SDK, CLI, and REST API for generating speech from text using an interface compatible with OpenAI’s text-to-speech API. It aims to make the service accessible to applications that expect that API shape, with a web playground for manual testing.
The project is a reverse-engineered client and service built on the openai.fm backend, not a local speech model. Its README limits the intended use to education and research and warns against commercial or production use.
Key Features
- Python client with synchronous and asynchronous interfaces.
- OpenAI-compatible
POST /v1/audio/speechendpoint and command-line client. - Eleven listed voice choices and six audio formats in the full Docker image.
- Text splitting and audio combination for long inputs.
- WebSocket support for streaming audio generation.
- Optional speed adjustment and additional format conversion with ffmpeg.
- Web playground and capabilities endpoint for trying the service and checking available features.
Use Cases
- Developers can test text-to-speech flows in a prototype that already expects OpenAI-style request fields.
- Python developers can generate speech files from scripts without building an API client themselves.
- Researchers and learners can explore a self-hosted interface to a reverse-engineered speech service, within the project’s stated non-commercial scope.
- Teams evaluating voices, formats, or playback speed can use the web playground before integrating a client.
What you need
Detected in the repository
- Python >=3.8 (from pyproject.toml)
- A Dockerfile, so it can run in a container
- A test suite and automated checks on GitHub Actions
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
Install the client with pip install ttsfm, or start the full Docker image with docker run -p 8000:8000 dbcccc/ttsfm:latest. See the README for client examples, API details, and image variants.
Alternatives
- GPT-SoVITS: GPT-SoVITS offers voice cloning and speech conversion with a WebUI, while ttsfm provides an OpenAI-style client, REST API, and playground.
- ChatTTS: ChatTTS is a dialogue-focused speech model with speaker and prosody controls, while ttsfm wraps a reverse-engineered backend behind an OpenAI-style API.
- tortoise.cpp: tortoise.cpp runs Tortoise-TTS locally in C++ with CPU or CUDA, rather than exposing ttsfm’s Python client and REST API.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| ttsfm | Python | MIT | 739 | Quiet |
| GPT-SoVITS | Python | MIT | 62.7k | Active |
| ChatTTS | Python | AGPL-3.0 | 39.9k | Quiet |
| tortoise.cpp | C++ | MIT | 192 | Inactive |
Considerations
The README describes TTSFM as reverse-engineered and based on openai.fm, so compatibility depends on a backend the project does not present as a stable production service. The repository is marked quiet, with no commits in the last 90 days despite recent releases, and contributor activity is concentrated in one person. The slim Docker variant omits ffmpeg, limiting it to MP3 and WAV and removing speed adjustment and some combining and conversion capabilities. The project requires Python 3.8 or later for package use; Docker is also supported. It is MIT-licensed, but that license does not override the README’s warning to avoid commercial and production use or the need to comply with applicable service terms.
Found this useful?
Share it with someone who would like ttsfm.