milla-jovovich

The highest-scoring AI memory system ever benchmarked. And it's free.

981
89
100% credibility
Found Apr 07, 2026 at 841 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

MemPalace is a local tool that structures and searches AI conversations and project files into an organized memory system accessible via simple commands.

How It Works

1
🔍 Discover MemPalace

You hear about a simple way to save and find everything from your AI chats and project notes so nothing gets forgotten.

2
📦 Get it set up

With one easy command, you add it to your computer and it's ready to use right away.

3
🏠 Build your memory home

Tell it about your projects and people by pointing to your folders, and it organizes everything into neat sections.

4
📂 Add your stories

Feed in your chat logs and files, and it stores every word safely in your personal memory palace.

5
💡 Find anything instantly

Ask about a past decision or idea, and it pulls up the exact words from months ago like magic.

6
🤖 Let your AI remember

Share the memories with your AI helper so it knows your history without starting over each time.

🎉 Never forget again

Now your AI chats and projects stay alive forever, helping you build on the past with confidence.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 841 to 981 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 mempalace?

MemPalace is a Python-based AI memory system that ingests your project files, code, docs, and chat exports from Claude, ChatGPT, or Slack, organizing everything into a searchable "palace" of wings for projects/people, halls for memory types like decisions or events, and rooms for specific topics. Instead of AI summaries that lose details, it stores full verbatim content locally with ChromaDB, then makes it instantly findable via semantic search—CLI commands like `mempalace mine ~/chats --mode convos` or `mempalace search "GraphQL switch"`. A lossless AAAK dialect compresses context 30x for any LLM, keeping your stack offline.

Why is it gaining traction?

It claims the highest-scoring benchmarked memory system ever on LongMemEval (96.6% raw recall@5, 100% hybrid—all zero API calls and free), beating paid tools like Mem0 or Zep by 10-15 points while staying local. The palace structure alone boosts retrieval 34% via wing/room filtering, and MCP integration gives Claude 19 tools like `mempalace_search` without manual prompting. Developers dig the `wake-up` command loading ~170 tokens of critical facts on demand.

Who should use this?

Solo devs juggling multiple repos who lose architecture debates in chat histories; team leads mining Slack/Claude for "who owns auth?" without Jira; AI power users building agent workflows needing persistent, offline recall across sessions. Perfect for Python shops on macOS/Linux tracking decisions in convos.

Verdict

Grab it if persistent AI memory sounds useful—pip install and mine your chats in minutes for top benchmarked results at zero cost. With 209 stars and 1.0% credibility, it's early beta (solid docs, benchmarks folder), so test on non-critical data first.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.