Ermine

Ermine is a client-side transcription tool that converts spoken audio into text directly in the browser, without sending data to external servers. It runs entirely locally using WebAssembly and on-device models, enabling secure, low-latency transcription from your device’s microphone. Users can start recording in one click, see live transcription appear in real time, and export or copy the resulting text for further use.

Key capabilities include continuous speech recognition, automatic punctuation, and support for long-form audio, such as meetings, lectures, interviews, or personal dictation. Because processing is done on the client, Ermine is suitable for sensitive or confidential conversations where cloud-based transcription is not acceptable. The tool is optimized for modern browsers and can work offline after initial model loading, making it useful in low-connectivity environments.