TSchonleber

TSchonleber / brainctl

Public

A cognitive memory system for AI agents. Single SQLite file. MCP server included.

11
3
100% credibility
Found Apr 06, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

brainctl provides AI agents with persistent, structured memory in a single SQLite file, including automatic decay, consolidation, knowledge graphs, and integration with tools like MCP servers.

How It Works

1
🔍 Discover brainctl

You hear about a simple way to give your AI helper a real memory that sticks around forever, like a notebook it never forgets.

2
📦 Get it set up

With one easy command, you install it and create a single file where all your AI's memories will live.

3
💭 Start adding memories

Tell it facts like 'User loves dark mode' or 'Fixed the login bug today' – it saves them smartly with categories and links.

4
🧠 Memories start connecting

Watch as it automatically builds a web of knowledge, linking people, projects, and lessons so nothing gets lost.

5
🔎 Ask and get answers

Search for 'login issues' and get exactly the right memories, events, and people – fast and relevant every time.

6
🛌 Let it tidy overnight

Set it to clean up quietly: old stuff fades, duplicates merge, important bits get stronger – all hands-free.

🚀 Your AI remembers forever

Now your helper recalls everything across chats, learns from mistakes, and grows smarter without starting over.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 11 to 11 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 brainctl?

brainctl is a cognitive memory system for AI agents, storing everything—memories, events, entities, knowledge graphs, decisions, and affect states—in a single SQLite file called brain.db. Developers get a Python library and CLI to add facts like "User prefers dark mode," search across tables, build relations between entities, and log events, all without LLM calls or vendor lock-in. It mimics cognitive memory behaviors with automatic decay, duplicate suppression, and consolidation, solving the problem of agents forgetting context between sessions.

Why is it gaining traction?

Unlike mem0, Zep, or MemGPT, brainctl runs server-free with pure SQLite FTS5 search (vectors optional), includes an MCP server for Claude Desktop/VS Code/Cursor, and handles cognitive memory issues like overload or loss via half-life decay and write gates. Multi-agent support lets teams share one brain.db, with token-optimized outputs slashing LLM costs by up to 97%. Neuroscience-inspired features like prospective triggers and affect tracking make agents feel smarter without complexity.

Who should use this?

AI agent builders tired of re-injecting context every run, especially in multi-agent cognitive AI GitHub projects or cognitive science experiments. Devs using Cursor/Claude with MCP for codebases needing persistent project knowledge, or teams tracking cognitive behaviors like decisions and user prefs across agents. Ideal for cognitive memory training in prototypes before scaling to full cognitive services.

Verdict

Try it for lightweight agent memory in cognitive AI setups—installs via pip, works out-of-box with brainctl init—but 11 stars and 1.0% credibility scream alpha stage with thin docs/tests. Promising for agents, but productionize at your own risk until maturity catches up.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.