Repository navigation
BetterWispr
Typing is slower than speaking, and many dictation services send recordings to a remote server. BetterWispr is a free, open-source dictation app for macOS that turns speech into text in whatever application has focus, with speech processed on the Mac itself.
The user holds Option + Space, speaks, and releases, and the cleaned-up text is inserted at the cursor (product site). The app removes filler words, adapts writing style per application, and supports custom vocabulary, learned corrections, and voice commands such as "new line" (README).
It also transcribes meetings with summaries and action items, and can transcribe audio, video, and video links. MCP-compatible apps can read meeting notes in a read-only mode that is off by default.
On-device options are Apple's built-in speech recognition, NVIDIA Parakeet v3 (25 European languages), Parakeet v2 (English), and Whisper large-v3 turbo. Optional connections exist for Sarvam AI, Smallest AI, and any OpenAI-compatible endpoint. Installation is a DMG from the Releases page, with Microphone and Accessibility permissions granted at first launch.
The repository was created on 2026-10-07 under the Apache-2.0 license. At the time of review it had 374 stars and 13 open issues. It requires macOS 14 or later, and Apple Silicon is recommended.
BetterWispr is placed in Assess within Tools. Local processing, no account, and no cost make it cheap to try, but the project is only days old and no first-person production use has been confirmed. Cheap trial and early star count describe a technology worth exploring, which is Assess.
