Kaldi provides an open toolkit for speech recognition and speech processing research.
Vosk
Vosk provides an offline speech recognition API for mobile devices and servers, with several programming language interfaces.
About Vosk
Vosk is an offline, open source speech recognition toolkit distributed through a GitHub repository. It provides speech-to-text on Android, iOS, Raspberry Pi and servers, with bindings for Python, Java, Node.js, C#, C++, Rust, Go and other languages. The project states support for 20+ languages and dialects, and its models are around 50 MB while still offering continuous large-vocabulary transcription. Additional capabilities described include a streaming API with zero-latency response, a reconfigurable vocabulary and speaker identification. Because recognition runs offline, no network connection is required for the core task, and the toolkit scales from small devices such as a Raspberry Pi or Android smartphone up to large clusters. Pricing is not stated in the source material.
Key features
- Offline speech recognition across Android, iOS, Raspberry Pi and servers
- Bindings for Python, Java, Node.js, C#, C++, Rust, Go and others
- Recognition for 20+ languages and dialects
- Models around 50 MB with continuous large-vocabulary transcription
- Streaming API with zero-latency response and reconfigurable vocabulary
- Speaker identification
Best for: Developers building offline speech features for chatbots, smart home appliances, virtual assistants, subtitles, or lecture and interview transcription across small devices and servers.
Good for
- Open source project released under the Apache-2.0 license
- Recognition runs offline, so the core task does not depend on a network connection
- Scales from small devices like Raspberry Pi or Android smartphones to big clusters
Vosk at a glance
- Website
- github.com/alphacep/vosk-api
- Category
- AI & Machine Learning › AI Voice & Audio
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Platforms
- Android, iOS
- Open source
- Yes · Source code
- Language
- English
- Last checked
- Oct 6, 2026 · Online when checked
- Listed
- Oct 5, 2026
Vosk alternatives
Tools and sites you could use instead, ranked by how closely they match.
The whisper.cpp project ports OpenAI’s Whisper speech recognition model to C and C++.
Open source
24 alternatives
ESPnet includes tools for end to end speech processing and related speech tasks.
Open source
24 alternatives
The whisper-ctranslate2 command-line client works with the OpenAI Whisper client and uses CTranslate2.
Open source
24 alternatives
The wav2letter project provides an automatic speech recognition toolkit from Facebook AI Research.
Offer Bull provides AI interview practice, real-time speech recognition, and live answer suggestions for job seekers preparing for online interviews.
Popular in AI Voice & Audio
Well-known listings from the same topic.
Google Illuminateilluminate.google.comAlteredaltered.aiELSA Speakelsaspeak.comKreadoAIkreadoai.comPodwisepodwise.aiAddsubtitleaddsubtitle.ai
Vosk FAQ
Does Vosk require an internet connection to recognize speech?
No. It is described as an offline speech recognition toolkit, so recognition runs locally rather than through a network service.
Which programming languages can I use with Vosk?
Bindings are implemented for Python, Java, Node.js, C#, C++, Rust, Go and others.
What platforms does Vosk run on?
Android, iOS, Raspberry Pi and servers, scaling from small devices up to big clusters.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.