Vendor page
OpenWhisper README on GitHub
OpenWhisper, GitHub, richardwu/openwhisper repository, n.d.
Read the originalRead on October 22, 2026
- What it is
- The README of OpenWhisper, a free local voice-to-text menu bar app for Mac, in the maker's GitHub repository. It describes transcription with on-device models, custom vocabulary, pasting the result into the active app, and the internet connection needed only for initial model downloads and updates.
- How much weight it carries
- The maker's own repository README, read in October 2026. It is the only official page read for this app. It gives no user price, plan or data-training statement, so no row is written for those. The offline row reads its download statement together with its on-device design, and the caveat says so.
Facts taken from this source
- OpenWhisper's README says that audio and transcription stay on the user's Mac.
“Local transcription ... audio and transcription stay on your Mac.” (README, Features list)
- OpenWhisper's README lists custom vocabulary, where you pin names and jargon and add comma-separated terms.
“Custom vocabulary ... pin names and jargon, add comma-separated terms, search, and remove entries.” (README, Features list)
- OpenWhisper's README says its Nemotron 3.5 ASR 0.6B streaming model supports 28 languages, plus automatic detection.
“Model Mode Languages in OpenWhisper ... Nemotron 3.5 ASR 0.6B Streaming 28 languages, plus automatic detection” (Models table, Languages in OpenWhisper column)
- OpenWhisper's README says the app pastes the transcription result into the active app.
“OpenWhisper lives in your menu bar, transcribes with your choice of on-device models, and pastes the result into your active app.” (README, opening description and Requirements)
- OpenWhisper's README says the app transcribes with a choice of on-device models.
“OpenWhisper lives in your menu bar, transcribes with your choice of on-device models, and pastes the result into your active app.” (README, opening description)
- OpenWhisper's README says its Voxtral Mini 4B Realtime streaming model supports 13 languages, plus automatic detection.
“Model Mode Languages in OpenWhisper ... Voxtral Mini 4B Realtime Streaming 13 languages, plus automatic detection” (Models table, Languages in OpenWhisper column)
- OpenWhisper's README says its Whisper Base, Small, Medium and Large models support 99 languages, plus automatic detection.
“Model Mode Languages in OpenWhisper ... Whisper Base / Small / Medium / Large Batch 99 languages, plus automatic detection” (Models table, Languages in OpenWhisper column)
- OpenWhisper's README says model files remain on your Mac and an internet connection is needed for initial downloads and updates.
“Model files download when needed and remain on your Mac. An internet connection is needed for initial downloads and updates.” (README, Models section)