Who it is forA Python developer who wants an offline assistant feature without API keys or per-token bills.
The problemCloud LLM APIs add latency, cost and send user data off the device.
The outcomeA Python script that downloads a model once and generates replies locally.
What you need first
- Rust, Python (uv), Swift or TypeScript (pnpm)
- An Apple device to benefit from unified memory
- Disk space for the downloaded model
Steps
- Run
uv add uzu==0.6.13 - Create the engine with EngineConfig.create() and Engine.create as in the README Python example
- Load the model
alibaba:qwen3.5:0.8b:mirai:mirai-m:4and iterate the download to show progress - Open a chat session and send a system and a user ChatMessage
- Print the reply's reasoning and text
Prompt for your agent
Add trymirai/uzu to this Python project with `uv add uzu==0.6.13` and write a script that downloads alibaba:qwen3.5:0.8b:mirai:mirai-m:4 and answers <question> locally, following the README Python quick start.
Caveat. This card is written from the trymirai/uzu README, not from a test run by Deduped. Before you start, make sure you have Rust, Python, Swift or TypeScript. Check the repo for newer instructions before you rely on these steps.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for trymirai/uzu and "Run a local LLM inside a Python script with Uzu"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer