Your voice is not
a subscription.
In the AI era, talking to your computer is a basic need — like copy and paste, like a keyboard. Basic needs should never be gated by payment. That is the whole mission, and everything else follows from it.
Talking is how humans work. Typing is the workaround.
People speak at roughly 150 words a minute and type at roughly 40. A Stanford study measured speech input at about three times the speed of typing on mobile keyboards. Voice has always been the faster interface — what was missing was software that respects you enough to run on your own machine, for free.
Pay for tokens, not permission.
Companies charge $144 a year to stand between your voice and your screen — for compute your Mac already has. We think that's backwards. Dictation in OpenVoiceFlow is free for everyone, with no account and no tier: Whisper runs on your machine, your audio never leaves it, and there is nothing to upgrade to.
The one optional extra — AI cleanup that polishes grammar and filler words — is bring-your-own-key. Point it at OpenRouter, OpenAI, Anthropic, or Groq and you pay the model provider directly, in cents, for tokens you actually use. Or run Ollama and pay nobody. Your money should buy compute, not permission — go buy tokens, go build something, and keep the $144.
The best version of this app is the one that becomes unnecessary.
Our goal is not to own a market. It is for voice input this good to be a default feature of every operating system — built in, free, private — so that third-party developers no longer need to build apps like this one. Until that day, we build it in the open: MIT-licensed, source public, every privacy claim checkable in code.
If you're a developer who believes typing shouldn't be a toll booth, this is an invitation. File an honest bug report. Add a test. Improve accuracy for your language. Join us on GitHub.
Who builds this
OpenVoiceFlow is built by Shimoverse Studios and open-source contributors. The code is MIT-licensed — free for any use, with credit required. The name and logo are protected by a brand policy, so anything wearing the OpenVoiceFlow name keeps the promises on this website: audio on-device, no telemetry, no account, no paid tier. Read the privacy policy — the short version is that we couldn't sell your data even if we wanted to, because we never have it.