#116 · Primary category: Speech & Audio

pykaldi

asr clif feature-extraction kaldi language-model numpy openfst python speech speech-recognition wrapper

A Python wrapper for Kaldi

Project last updated:11/30/25

GitHub Stars

1.0K

Forks

247

Contributors

32

License

Apache-2.0

Why we included this project

Kaldi is powerful but C++-heavy, and PyKaldi exists to close that gap: it wraps both Kaldi and OpenFst so you can call low-level functions, inspect and modify recognizer internals, and write new tools without dropping back into C++. The people who get the most out of it are researchers and engineers already committed to Kaldi who want programmatic control over its guts. If that sounds like more than you need, the asr, alignment, and segmentation modules provide a gentler path for developers who just want to run a pre-trained Kaldi pipeline inside a larger Python application. PyKaldi deliberately does not try to replicate Kaldi's command-line recipes; it is meant to sit beside an existing Kaldi setup and extend it. That makes it a practical fit for teams that want to escape the C++ layer without rebuilding their stack.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category