Coqui TTS
coqui-ai/TTSReviewed by AI
Checked by RepoPicks Radar on
What is Coqui TTS?
Coqui TTS is a Python toolkit that reads text aloud with ready-made models in many languages, and can imitate a voice from a sample recording (the XTTS and YourTTS models) or convert one voice into another. It also has the training code for making your own voice models. You use it from the terminal with the `tts` command, from Python code, or through a small web page served by its Docker image. It is a tool for people comfortable with Python; there is no installer, and the project has not been updated since February 2024.
Coqui TTS is free and open source (MPL-2.0) and can replace ElevenLabs and Murf.
How easy is it to install Coqui TTS?
D · terminal, several tools or a specific graphics card · macOS
- Windows E
- macOS D
- Linux D
None of the 98 releases has a file to download. On Linux, which the README says it is tested on (Ubuntu 18.04), it is `pip install TTS` on a Python between 3.9 and 3.11, then a command such as `tts --text "..." --out_path speech.wav` that downloads the model on first use; a Docker image with a small web page is the other route, and a graphics card is what makes it usable in practice. macOS follows the same pip route without a CUDA card. Windows is worse than a D because the README has no instructions of its own: it links to an answer someone wrote on StackOverflow.
Coming soon: install in 1 click with the RepoPicks app.
Coqui TTS on GitHub
- Stars
- 46,208on Nov 10, 2026
- Downloads
- 7,019,881all releases
- Growth
- +29stars in 7 days
Stars over time
Facts
- Runs on
- Windows, macOS, Linux
- You need to install
- Python, Docker, Graphics card
- To build from source
- Python, Docker, Graphics card
- Disk space
- 950 MB–4.3 GB · estimate
- Accounts or keys
- None
- Cost
- Free · The code is free software under MPL-2.0 and runs on your own machine, so there is nothing to pay. The pretrained models are downloaded separately and each carries its own terms.
- License
- MPL-2.0 (Copyleft)
- Can replace
- ElevenLabs, Murf
- Last change on GitHub
- Feb 10, 2024
Coqui TTS risk signals (8)
- Attentionstale releases
Last release on 2023-12-12 (1057 days ago)
automatic check
- AttentionDependencies with known vulnerabilities
Pinned dependencies with known vulnerabilities (OSV): pylint: GHSA-22qq-2fwx-7p67; bokeh: GHSA-793v-589g-574v, PYSEC-2026-1223
automatic check
- AttentionLow OpenSSF Scorecard
OpenSSF Scorecard 2.9/10
automatic check
- Attentioncopying someone s voice
The README's own examples clone a voice from an audio file — `tts.tts_to_file("This is voice cloning.", speaker_wav="my/cloning/audio.wav", ...)` — and the news section advertises "🐶Bark ... with unconstrained voice cloning", plus a voice conversion example that turns one person's recording into another person's voice. As a rule, making a recording that sounds like a real person without that person's agreement is restricted by law in several countries.
AI review
- Attentionno commits for almost three years
The last commit is from 2024-02-10, 997 days before this check, the newest release v0.22.0 is from 2023-12-12, and only 2 issues are open against 46,178 stars, which is what a closed issue tracker looks like. The project is not marked archived, but nobody is fixing anything here.
AI review
- Infoneeds an older python
setup.py raises an error on anything outside `python >= 3.9 and < 3.12` and requirements.txt pins old libraries (`numpy==1.22.0` for Python up to 3.10, `bokeh==1.4.0`, `pylint==2.10.2`), so you need an older Python than current systems ship.
AI review
- Infomodels are downloaded at run time
The code names models by id (for example `tts_models/multilingual/multi-dataset/xtts_v2`) and fetches them when first used, so what actually runs arrives from the internet after you install, and each model's own licence applies rather than the repository's MPL-2.0.
AI review
- Infotracking pixel in the readme
The README embeds `https://static.scarf.sh/a.png?x-pxid=cf317fe7-2188-4721-bc01-124bb5d5dbb2`, an image from the Scarf analytics service, so simply opening the project page reports a view to a third party.
AI review
Evidence is shown in its original language.
What Radar could not check about Coqui TTS
- Whether `pip install TTS` still resolves today, with the old pinned libraries and a Python older than 3.12
- The licence of each pretrained model, which is separate from the repository's MPL-2.0 and can forbid commercial use
- How good the result sounds and how closely a cloned voice matches, which only listening can tell
- Real disk and memory use after the models download, since the figure here is an estimate
- Whether it runs usably without an NVIDIA graphics card