← VoiceStrip

A more accurate model, free

VoiceStrip comes with a fast speech model. If it mishears names or longer sentences, this larger one gets more right — it just takes a moment longer to write.

Accurate (English)

181 MB · works offline once imported

Download the accurate model
SHA-256 bfdff4894dcb76bbf647d56263ea2a96645423f1669176f4844a1bf8e478ad30

Put it in VoiceStrip

  1. Download the file on your phone. It lands in Downloads.
  2. Open the VoiceStrip app and tap Models.
  3. Tap Import a speech model… and choose the file you downloaded.
  4. VoiceStrip checks the file, then lists it. Select it, and you're done.

Your browser does the downloading. VoiceStrip itself still has no internet access — it only reads the file you hand it. You can delete the download afterwards; VoiceStrip keeps its own copy.

Other languages and models

VoiceStrip runs any whisper.cpp model — import it the same way. Models without “.en” in the name understand other languages; once one is selected, VoiceStrip shows a language setting.