Free speech recognition and text-to-speech
Speech models have long run locally and for free — even on an ordinary laptop.
8 tools8 open source
| # | Tool ↕ | What is free | License | Platforms | GitHub ↕ | |
|---|---|---|---|---|---|---|
| 1 |  Whisper OpenAI |
Code and model weights are MIT-licensed and free to run locally; only OpenAI's separate cloud transcription API is paid. | Open source MIT |
WindowsmacOSLinuxCLI | ★ 109,500 | |
| 2 | whisper.cpp ggml-org |
Free and open source; runs fully offline on CPU, GPU or NPU. | Open source MIT |
WindowsmacOSLinuxCLIDocker | ★ 53,900 | |
| 3 | Handy CJ Pais |
Completely free and open source; transcription runs offline with Whisper or Parakeet models. | Open source MIT |
WindowsmacOSLinux | ★ 32,100 | |
| 4 | Chatterbox Resemble AI |
MIT-licensed models are free to self-host and use commercially; every output carries a PerTh watermark. | Open source MIT |
WindowsmacOSLinux | ★ 26,500 | |
| 5 | faster-whisper SYSTRAN |
Free MIT-licensed library that runs Whisper models locally on CPU or an NVIDIA GPU. | Open source MIT |
WindowsmacOSLinux | ★ 25,500 | |
| 6 | Buzz Chidi Williams |
Buzz Classic is free and open source for offline Whisper transcription and translation. | Open source MIT |
WindowsmacOSLinux | ★ 21,600 | |
| 7 | Kokoro TTS hexgrad |
Code and weights are Apache-2.0: free to run locally and to use commercially. | Open source Apache-2.0 |
WindowsmacOSLinux | ★ 9,000 | |
| 8 | Piper Open Home Foundation |
Free and runs fully offline (GPL-3.0); the project is currently looking for maintainers. | Open source GPL-3.0 |
LinuxWindowsmacOSCLIDocker | ★ 5,700 |
Nothing found.
Need a free model for these tools? Pick one in the catalog — every model card has a base URL and a code sample. Coding models · Models for agents