HKUDS

HKUDS / CatchMe

Public

"CatchMe: Capture Your Entire Digital Footprint"

hkuds.github.ioCatchMe ai-agent clawdbot-plugin llm recall-ai retrieval-systems
54
4
100% credibility
Found Mar 31, 2026 at 54 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

CatchMe records your computer activity like mouse movements, keystrokes, and app usage, organizing it into a privacy-focused, AI-searchable memory tree for easy recall and agent integration.

How It Works

1
🔍 Discover CatchMe

You hear about a simple tool that remembers everything you do on your computer, helping your AI assistants understand your daily work.

2
📦 Set it up quickly

Download and launch it with a few easy steps, no complicated setup needed.

3
🧠 Connect your AI helper

Pick a smart AI service to make sense of your activities, keeping everything private on your machine.

4
▶️ Start capturing your day

Turn on recording and let it quietly track your mouse clicks, typing, apps, and files as you work.

5
🌳 Explore your memory tree

Open the dashboard to see your day organized into a beautiful, searchable tree of summaries.

6
💬 Ask and recall anything

Chat naturally to relive moments, find files, or trace your steps from yesterday.

✨ Your AI knows you perfectly

Now your personal assistants have full context of your work, making them truly helpful companions.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 54 to 54 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is CatchMe?

CatchMe is a Python tool that silently captures your entire digital footprint—mouse moves, keystrokes, window switches, clipboard changes, notifications, and file access—across macOS and Windows. It organizes raw events into a searchable hierarchy (day, session, app, location, action) with LLM summaries, letting you query via CLI like `catchme ask "What was I coding yesterday?"` or a web dashboard for timelines and chat. Privacy-focused, it stores everything locally in SQLite, no cloud required.

Why is it gaining traction?

Unlike vector-heavy screen recorders or ActivityWatch clones, CatchMe skips embeddings for tree-based LLM retrieval, keeping RAM under 0.2GB and costs low (~$0.40 for 2 hours intensive use). Agent devs love the zero-config CLI skills for tools like Cursor or Claude, turning your PC history into instant context. Multilingual docs and offline LLM support (Ollama) make it dead simple to deploy.

Who should use this?

AI agent builders feeding personal context to LLMs, like HKUDS NanoBot users. Researchers replaying reading sessions or code edits. Power users in coding marathons or deep dives needing "catchmeifyoucan" recall of catchment areas in their digital day—without the watchmen overhead.

Verdict

Promising for agent workflows, but 54 stars and 1.0% credibility signal early days—docs shine, but expect rough edges in edge cases. Try if you're prototyping personal AI; skip for production until more battle-tested.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.