kitio.aiAdded Nov 5, 2026
Kaldi
FreeOpen-source toolkit for building speech recognition systems
What is Kaldi?
Kaldi is a free, open-source C++ toolkit for speech recognition research and production systems, started by Daniel Povey. Researchers and engineers use it to train acoustic models and build custom recognizers rather than call a hosted API.
The project ships training recipes for common public speech datasets, along with tools for feature extraction, acoustic modeling and decoding with weighted finite-state transducers. Many academic and commercial speech systems were built on it.
The source code lives on GitHub under the Apache 2.0 license and can be cloned or downloaded as a zip. The website hosts documentation, a help forum and a small set of pretrained model builds.
Features
- Acoustic model training
- WFST-based decoding
- Recipes for public datasets
- Feature extraction tools
- Pretrained model downloads
Tags
Releases
+ Submit new release- Listed
Kaldi was added to the directory.
Organization
Daniel Povey and Kaldi contributors
Listed by kitio.ai · unverified owner
Pricing
- Pricing model
- Free
- Paid options from
- 100% free
Sponsored
Your AI tool here
A Featured launch puts your tool in this spot and on the home page for a week.
Top Kaldi alternatives
View allAudio & voice
Meta's open-source library for MusicGen and AudioGen models
Audio & voice
OpenAI's open-source speech recognition and translation model
Audio & voice
Open-source text-to-speech with zero-shot voice cloning
Audio & voice
Free desktop app for AI vocal removal and stem separation
If you liked Kaldi
Minimal open-source coding agent that works through bash alone
FreeCodingMade in United StatesOpen-source framework and research community for multi-agent systems
FreeAgents & automation1 clicks to the websitePrivate, open-source research notebook that runs on your own machine
FreemiumResearch & search1 viewsIBM speech recognition API for transcription, captions and call analytics
FreemiumAudio & voiceMade in United States
Pros and cons
Pros
- Free under a permissive license
- Full control over the speech pipeline
- Long track record in research
Cons
- Steep learning curve for newcomers
- Few pretrained models on the site
- Needs your own servers and training data
Prompts & results
No prompts yet. Add your own prompt and output to help others use this tool.
Share a prompt
Reviews
Q&A
What is Kaldi?
Kaldi is a free, open-source C++ toolkit for speech recognition research and production systems, started by Daniel Povey. Researchers and engineers use it to train acoustic models and build custom recognizers rather than call a hosted API.
Where can I get the Kaldi source code?
The code is on GitHub at kaldi-asr/kaldi, which you can clone with git or download as a zip.
How much does Kaldi cost?
Kaldi is 100% free.
Is there a free version of Kaldi?
Yes. Kaldi is completely free to use.
What are the main features of Kaldi?
Acoustic model training, WFST-based decoding, Recipes for public datasets, Feature extraction tools, Pretrained model downloads.
Who makes Kaldi?
Kaldi is made by Daniel Povey and Kaldi contributors.
What are the best alternatives to Kaldi?
Popular alternatives include AudioCraft, Whisper, Chatterbox.