We built a free and open-source alternative to Cotypist because we're sick of paying subscriptions for apps that run on our own hardware
If you haven't saw the recent news around Cotypist, they're charging a subscription for an app that runs entirely on your hardware using free, open-source models.
They're gating which models you can run on your own machine behind pricing tiers, paywalling clipboard awareness and per-app customization, and capping at 100 words a day.
I'm not sure about you, but this sounds insane to me for local inference, on your own Mac.
That's why we've been working on an open-source version called Tabby. It's the same idea but 100% free. A macOS menu bar app that brings inline autocomplete to any text field.
What you get, for free, until we both drop dead:
- Works in any macOS text field (Mail, Notes, Slack, iMessage, Safari, etc.)
- 100% on-device inference, nothing ever leaves your Mac
- Visual context via screenshot OCR + clipboard awareness
- Apple Intelligence or any GGUF model via llama.cpp
- Use ANY model. Drop a .gguf file in and go. No tiers, no paywalls on model size
- Unlimited completions. No daily word caps
- Works on as many Macs as you own
- Every single new update
AGPL licensed, open-source on GitHub. No accounts, no licenses, no kill switches, no hardcoded expiry dates, no forced updates that lock you out. We built this because we want to help people save time, not take advantage of their wallet.
Tabby is in beta, and definitely not as polished yet. Custom writing instructions, per-app rules, tone personalization/learning are on the roadmap. We're just two devs so this is very much a community project. If you want to help out, please feel free to star the repo, report bugs, or open a PR. It all makes a real difference.
GitHub: https://github.com/FuJacob/tabby
Website: https://tabbyapp.dev