Open-source voice dictation
Open-source voice dictation you can actually trust
OpenBroca is free, MIT-licensed voice dictation built in the open. Speak into any app and have your words turned into polished text — using cloud providers for speed or fully local models for privacy, with nothing hidden behind a closed binary or a subscription.
What is open-source dictation?
Open-source dictation is voice-to-text software whose source code is published under a license that lets anyone read, modify, and redistribute it. Instead of trusting a vendor's marketing claims about how your speech is processed, you (or anyone in the community) can inspect the code and verify it. Most popular dictation apps — Wispr Flow, Typeless, Monologue — are polished but closed-source commercial products. OpenBroca takes the opposite approach: the entire app is open and free.
Why open source matters for dictation
Dictation is uniquely sensitive: it captures your raw voice and everything you say — drafts, messages, private notes. That makes transparency and control more than nice-to-haves.
Free and permissively licensed
MIT-licensed means you can use, fork, and ship OpenBroca with no subscription and no usage caps.
Auditable, not a black box
Every line is on GitHub. Read exactly how recognition and rewriting work — and how your audio is handled.
Private by design
Run models fully on-device so audio never leaves your machine, with credentials in OS-backed secure storage.
No vendor lock-in
Swap between 16+ language models and two speech engines through open interfaces. Your workflow is yours.
How OpenBroca delivers open-source dictation
OpenBroca turns speech into text and intent into action as a system-wide input layer — dictate into your editor, browser, chat, or terminal. Speech recognition runs through open interfaces, so you can pick Deepgram in the cloud for speed or Sherpa-ONNX entirely on-device for privacy. From there, your choice of 16+ language-model providers can polish, translate, or summarize what you said. Because the code is MIT-licensed and on GitHub, you can audit those data paths, file issues, send pull requests, or fork the project entirely. It ships as native desktop builds for macOS, Windows, and Linux.
Open-source vs closed dictation apps
Commercial apps often win on first-run polish and managed convenience, and that is a legitimate reason to choose them. But they ask you to trust a closed stack, accept a recurring subscription, and stay on whichever models the vendor selects. Open-source dictation trades a little hand-holding for ownership: no lock-in, no per-seat fees, and the ability to verify and extend everything. See how OpenBroca stacks up against Wispr Flow, Typeless, and Monologue.