a2328275243

a2328275243 / mempalace-evolve

Public

Self-evolving memory palace for AI agents — persistent memory with automatic learning, knowledge graph, and multi-agent support

33
1
80% credibility
Found May 29, 2026 at 49 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

MemPalace Evolve is a self-evolving memory system that gives AI assistants persistent, searchable memory across conversations. It stores information in organized categories, builds knowledge graphs of relationships, and automatically learns from past interactions to improve over time.

How It Works

1
💡 You realize your AI keeps forgetting

You notice your AI assistant doesn't remember important decisions from last week, so you look for a way to give it persistent memory.

2
📦 You install the memory palace

You add the memory system to your project with one simple command, and everything sets itself up automatically.

3
🏠 Your AI gets its own memory space

Each of your projects gets its own private memory area where it can store decisions, errors, and important facts separately.

4
🧠 Your AI starts remembering things

You tell your AI to remember important things, and it stores them in organized categories like decisions, errors, and architecture.

5
You can search memories in two ways
🔎
Ask questions naturally

Search memories by asking 'how did we handle authentication?' and get relevant past decisions instantly

🗺️
Explore connections

See how your project pieces fit together — like 'API connects to PostgreSQL and Redis'

6
✨ The system learns on its own

After conversations, your AI automatically reviews what happened and stores important new learnings without being asked.

🎉 Your AI now has perfect memory

Everything you discussed weeks ago is still there. Your AI understands your project history and never makes the same mistake twice.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 49 to 33 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-evolve?

MemPalace Evolve is a persistent memory system for AI agents that automatically learns and evolves over time. Built in Python, it stores memories using vector search (ChromaDB) and maintains a knowledge graph (SQLite) for structured relationships. The system organizes memories into rooms like "decisions," "errors," and "architecture," with support for multiple isolated project spaces called wings. An evolution pipeline automatically reviews memories, promotes valuable ones, and discards stale content based on usage patterns and importance scores.

Why is it gaining traction?

The killer feature is the self-evolving loop: every recall updates access timestamps, cross-wing hits get tracked, and the evolve command runs candidate extraction, review, and consolidation automatically. Unlike static memory stores, this learns what matters. The multi-adapter approach is also smart -- it works as a CLI tool, REST API, MCP server for Claude Code/Cursor, or LangChain tools. The contradiction detection is particularly clever: when you store a new decision, it marks the old one as superseded rather than creating conflicting entries.

Who should use this?

AI agent developers building long-running or multi-session applications will get the most value. If you're tired of re-explaining context to every new agent session, this solves that. It's also useful for teams running AI coding assistants (Claude Code, Cursor) who want persistent project memory without manual curation. The MCP server setup is particularly painless for Claude Code users.

Verdict

At 33 stars with a 0.800000011920929% credibility score, this is early-stage but actively developed with a solid architecture. The Python-only dependency (ChromaDB) and multiple integration paths make it easy to adopt incrementally. The 4-layer memory stack, knowledge graph, and evolution pipeline show real thought about memory management. Worth trying if you need persistent agent memory -- start with the demo and CLI to feel it out before committing to an adapter.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.