LakyFx

LakyFx / CogniLayer

Public

Shared brain for Claude Code & Codex CLI — one memory, two agents, all your projects. 13 MCP tools, crash recovery, cross-project search, deployment safety. Save ~100K tokens/session.

ai-memory anthropic claude claude-code codex-cli
10
1
100% credibility
Found Mar 02, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

CogniLayer creates a shared, persistent memory system for AI coding assistants, enabling knowledge retention across sessions, tools, and projects with visualization tools.

How It Works

1
🔍 Discover CogniLayer

You find this tool while searching for ways to make your AI coding helpers remember what they've learned across different sessions.

2
📥 Download and set up

Grab the files and run a quick setup script that prepares everything in your home folder.

3
🧠 Start your first AI session

Open your favorite AI coding assistant and watch it instantly recall project details, past fixes, and decisions without you repeating anything.

4
🔄 Switch AI tools seamlessly

Jump between different AI helpers during the day – they all share the same smart memory, so no lost context.

5
📊 Check your memory dashboard

Launch the colorful screen view to browse facts, see hot tips, spot gaps, and review session history at a glance.

🚀 Work smarter every day

Your AI partners now build on real experience across projects and crashes, saving hours of re-explaining and boosting your productivity.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 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 CogniLayer?

CogniLayer builds a shared brain for Claude Code and Codex CLI agents, letting them share one persistent memory database across sessions, projects, and even different tools. Switch from Claude in the morning to Codex in the afternoon, and it recalls debugging steps, decisions, or gotchas without re-explaining—saving 80-100K tokens per session via compact MCP tool injections. Written in Python, it hooks into both CLIs with crash recovery, cross-project search, and a TUI dashboard launched via `cognilayer`.

Why is it gaining traction?

No other tool offers true agent interoperability like this shared brain setup, turning solo AI coding into a shared github workflows experience where knowledge compounds like twins with one brain cell (minus the meme). Token savings pay off instantly on big codebases, plus deployment safety gates prevent "wrong server" disasters. Slash commands like `/recall query` or `/status` make it zero-learning-curve, with hybrid search pulling facts faster than re-reading repos.

Who should use this?

AI-assisted coders juggling Claude Code and Codex CLI on multiple projects, especially full-stack devs debugging auth flows or deploying to shared servers. Perfect for solo devs or small teams treating agents as shared brainstorming boards, avoiding context loss in long-lived github shared repositories.

Verdict

Early alpha with 10 stars and 1.0% credibility—docs are solid, install is 30 seconds via `python install.py`, but expect bugs in concurrent use. Try the demo TUI (`cognilayer --demo`) if you're deep in agent workflows; skip otherwise until more polish.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.