Run a tiny 8 to 29 MB AI model on a phone or microcontroller
Made by Cactus Compute, Inc.
View profile on GitHub (opens in a new tab)Needle is a small foundation model built specifically for devices too limited for a normal chatbot model — phones, wearables, smart-home hubs, robots and microcontrollers — packaged as a single binary as small as 8 to 29 megabytes.
It gives up general conversation ability in exchange for being very good at three narrow, useful jobs: picking the right function to call from a list your app provides and filling in its arguments correctly, pulling structured fields out of messy text (like a booking or an invoice), and turning a sentence into a searchable vector. The creator reports it beats models ten times its size on mobile tool-calling tasks.
You could use it to…
- Run reliable tool-calling AI on a phone, no internet needed
- Pull structured fields out of messy text, on-device
- Turn a sentence into a searchable vector locally
Can I use this?
You'll need A device to run the binary on, or the browser demo · Setup needed
Worth knowing Benchmark comparisons against larger models are the creator's own reported numbers rather than an independent test, and the model deliberately does not do open-ended chat, only the three tasks it was built for. Licence not stated in the excerpt available. BuildTube has not run or verified this model.
Why it's here
Not trending this week
Last seen on the GitHub trending list on 25 September 2026. It stays on BuildTube so you can still find it.
- 13,340 people have starred it on GitHub.
Numbers checked on 6 October 2026; not refreshed since.
Behind it
See the code on GitHubBuildTube has not run or verified this project. Everything above is written from what the creator published.
Made this?