janbjorge

janbjorge / rekal

Public

Long-term memory for LLMs. MCP server backed by hybrid search in a single SQLite file.

pypi.orgprojectrekal llm mcp mcp-server memory python
40
3
100% credibility
Found Apr 13, 2026 at 14 stars 3x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Rekal provides persistent local memory for AI coding assistants, storing and retrieving user preferences and facts across sessions in a single file on the user's machine.

How It Works

1
🔍 Discover Rekal

You hear about Rekal, a simple way to give your AI coding helper a lasting memory that stays right on your computer, remembering your preferences and facts across chats.

2
📦 Set Up Rekal

You quickly prepare Rekal on your computer, and it creates a private memory file just for you.

3
đź”— Connect to Your AI Helper

You link Rekal to your AI assistant like Claude, so it can now save and pull up memories during your work.

4
🧠 Teach Smart Habits

You add helpful guides that teach your AI when to save key details, like your favorite coding styles, without you reminding it every time.

5
đź’¬ Start Chatting with Memory

In your next session, type a command to fill your AI with project knowledge, and it begins remembering everything important automatically.

6
⚡ AI Recalls Perfectly

Weeks later, when setting up tools, your AI instantly remembers 'You prefer Ruff over Black' and sets it up without asking—feels like magic!

🎉 Smarter Sessions Forever

Your AI now grows wiser with every chat, recalling facts, preferences, and lessons effortlessly, making your coding life smoother and faster.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 40 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 rekal?

rekal is a Python MCP server that adds long-term memory to LLMs like Claude Code, storing everything in a single local SQLite file with no cloud or API keys needed. It solves the problem of AI agents forgetting preferences, facts, and procedures across sessions—think recalling "use Ruff over Black" 47 sessions later via hybrid search. Install with pip, add via `claude mcp add rekal`, and pair with its skills plugin for auto-capture and hygiene.

Why is it gaining traction?

Unlike cloud vector DBs, rekal runs fully offline in one portable file, blending keyword, semantic, and recency search for reliable retrieval without setup headaches. Developers dig the 16 MCP tools for storing, linking, superseding memories, plus introspection like conflict detection and timelines—perfect for github long term memory in agents. The skills plugin automates deduping and bootstrapping, making long-term memory agents feel native.

Who should use this?

Claude Code users building persistent AI assistants, or devs crafting long term memory langgraph setups who hate re-explaining project context. Ideal for solo coders tracking preferences and procedures across sessions, or teams needing github long term storage for shared LLM workflows without vendor lock-in.

Verdict

Try rekal if you're experimenting with long-term memory llm github projects—solid docs, CLI (`rekal serve`, `health`, `export`), and 100% test coverage make it reliable alpha software despite 12 stars and 1.0% credibility score. Skip for production until more battle-testing.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.