uzu
by trymirai
uzu is an engine for running AI language models directly inside your own app, on the user's device, rather than through a cloud service. It has libraries for Rust, Python, Swift and TypeScript, and handles downloading models and generating replies.
- Stars
- 2.9k
- Licence
- MIT
- Runs on
- Your computer
- Setup
- Technical
- Language
- Rust
- Age
- 491 days
What you can do with it
- Add a chat assistant to an iPhone, Mac, web or server app that runs a model locally
- Download supported open models automatically and generate replies with a few lines of code
- Use structured output and tool calls so a local model can feed into other parts of your app
Make money with it
A freelance developer could build and sell privacy-focused, offline AI features for mobile or desktop apps to clients who cannot send data to the cloud.
An idea, not financial advice. Check the licence before charging for anything built on it.
Ways a business could use it
- Build a private AI feature into a product without paying per-request fees to a cloud AI provider
- Offer customers an offline-capable assistant, so sensitive data never leaves their device
Can you use it commercially?
Free to use, change and sell, including in paid products. Keep the copyright notice.
Not legal advice. Read the project's licence file before building a product on it.
Getting it running
You add it as a dependency using Cargo, uv or pip for Python, Swift Package Manager, or pnpm for TypeScript, then write code that calls it.
Follow the official instructions in the project README.
Things to check
- It is a developer library, not a ready-made app, so coding is required
- Optimised for Apple devices (Metal), so performance on other hardware may vary
- Version is still below 1.0, so the interface may change
- Some examples use a cloud chat option that may involve the vendor's services
Who's talking about it
- José da Cruz: GitHub Weekly: This Week in Rust (October 12 – October 18, 2026) Featured in a roundup
Checked 28 Oct 2026. Summary drafted by AI from the project's README.