Ermine.ai is a browser-based application for recording and transcribing audio entirely locally, on the user’s own device. A transcription model loads once before first use and then processes all recording and transcription without sending any data to external servers – genuine on-device AI, not just a privacy claim layered on top of a cloud service.
The interface combines recording and transcription in one tool, with model caching so subsequent uses are faster after the initial download, and both the raw audio and finished transcript are available to download. The project is open source, with its code publicly available on GitHub.
Ermine.ai targets anyone wanting privacy-focused transcription without cloud dependency – journalists, researchers, or anyone handling sensitive recordings. Currently supports English transcription only; the initial model download is about 50MB and may take a few minutes on first use. The tool is free and open source.
