Vosk-api is an offline, open-source speech recognition toolkit designed for deployment across a wide range of environments, from small devices like Android smartphones, iOS hardware, and Raspberry Pi to large server clusters and data centers. It supports continuous, large-vocabulary transcription with zero-latency response via a streaming API and offers features such as reconfigurable vocabulary and speaker identification. The system currently supports over 20 languages and dialects, including English, Spanish, French, German, Chinese, Russian, Japanese, and Hindi, with models optimized at approximately 50 MB to ensure efficient performance. Vosk provides native bindings for numerous programming languages, including Python, Java, C, Node.js, C++, Rust, and Go, making it highly accessible for developers. Typical use cases include powering chatbots, smart home appliances, and virtual assistants, as well as generating subtitles for media and transcribing lectures or interviews. Its ability to operate without an