Coqui TTS

For several years the coqui-ai/TTS repository was the default open toolkit for speech synthesis: one Python package with a command line tool, a model zoo, training recipes, and implementations of Tacotron 2, Glow-TTS, VITS, YourTTS, Bark, Tortoise and the XTTS cloning models. Its maker, a startup founded by former Mozilla speech engineers, closed in January 2024. The original repo took its last commit in February 2024, and its final PyPI release, TTS 0.22.0 from December 2023, installs only on Python 3.9 to 3.11, which rules out current distributions unless you keep an older interpreter around. The code itself is sound, and the community carried it forward: Idiap Research Institute maintains a fork published on PyPI as coqui-tts, with releases through 2026, support for Python 3.10 to 3.14, prebuilt wheels for Windows and macOS, and the same TTS import path, so most old scripts run after swapping the package name. Code stays MPL-2.0, while each pretrained voice has its own license. For new work, install the Idiap fork, or try lighter current models such as Kokoro or Piper for plain narration.