MarcoPorcellato

MarcoPorcellato / matryca-logseq-llm-wiki

Public

Bringing Andrej Karpathy's LLM Wiki to the Outliner Paradigm. Turn any AI Agent into a spatial Knowledge Architect using Logseq's atomic nodes.

agentic-memory ai-agents andrej-karpathy autonomous-agents knowledge-graph
14
2
85% credibility
Found May 22, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Matryca Logseq LLM Wiki is a tool that connects AI assistants directly to your Logseq note-taking folder. Instead of copying and pasting notes into chat, you can ask your AI to read, search, edit, and organize your notes automatically. The AI understands the hierarchical structure of your notes and can create new content, find information across your entire knowledge base, discover topics you mentioned but didn't link, and help reorganize messy notes. All changes are written safely back to your notes folder with automatic backups. This is useful for anyone who uses Logseq and wants AI assistance in maintaining and growing their personal knowledge base.

How It Works

1
📓 You take notes in Logseq

You've been using Logseq to capture ideas, tasks, and knowledge in a folder of Markdown files on your computer.

2
🤖 You connect your AI assistant to your notes

You point your AI assistant to your notes folder so it can read and understand everything you've written.

3
🔍 Your AI reads and understands your notes

The AI sees the full structure of your notes—the hierarchy, the connections between ideas, and the tags you've used.

4
You ask the AI to help
🔎
Find information

Search across all your notes by keyword to find exactly what you're looking for

✏️
Create or edit notes

Add new ideas, update existing content, or fix typos in your notes

🔗
Discover hidden connections

Find places where you mentioned a topic but forgot to link it

🧹
Organize and clean up

Reorganize messy notes, split long blocks, or unify inconsistent tags

5
💾 Changes are saved safely

The AI writes changes directly to your notes folder, with automatic backups before making edits.

🎉 Your knowledge grows smarter

Your notes stay organized and connected, with AI helping you maintain and expand your personal knowledge base.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 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 matryca-logseq-llm-wiki?

Matryca is a Python-based MCP server that transforms your local Logseq knowledge base into an AI-accessible workspace. It lets you connect any MCP-compatible AI assistant (Claude Desktop, Cursor, and similar) directly to your Logseq vault without requiring the desktop app to run. The tool operates entirely through atomic file operations on Markdown, reading and writing blocks, properties, and journal entries while respecting Logseq's native outliner structure. A CLI tool called `matryca` handles everything the MCP server does, plus background service management for persistent setups.

Why is it gaining traction?

The project solves a real pain point: turning passive note-taking into active knowledge management. Its X-Ray mode compresses page content into numbered aliases that reduce context noise by up to 35x, making long pages usable in LLM conversations. The headless architecture means no Electron app hogging RAM, and the path sandboxing ensures AI writes never escape your graph directory. Git snapshots before mutations provide rollback safety, and the zero-dependency search (BM25 + structural traversal) avoids vector database complexity.

Who should use this?

Knowledge workers running Logseq who want AI agents to read, search, and write their notes. Researchers building personal knowledge bases with AI-assisted synthesis. Developers using Logseq for project documentation who need automated graph refactoring. Anyone frustrated by Logseq's API limitations and seeking a local-first alternative to cloud-connected tools.

Verdict

At 14 stars, this is early-stage but production-quality code with 162 passing tests and strict type checking. The 0.8500000238418579% credibility score reflects solid engineering practices despite low visibility. Worth trying if you want AI agents to interact meaningfully with your Logseq knowledge graph, but expect to encounter rough edges typical of young projects.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.