CodeAbra

CodeAbra / iai-mcp

Public

The best-benchmarked open-source memory system for AI coding assistants

ai-agents claude claude-code embeddings llm-tools
13
1
100% credibility
Found May 07, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A local memory system that automatically captures and recalls conversation history for AI coding assistants like Claude using the MCP protocol.

How It Works

1
🔍 Discover a smart memory helper

You hear about a tool that helps your AI coding buddy like Claude remember your past chats automatically, so conversations pick up right where you left off.

2
📥 Get it set up easily

Download and run a simple setup helper that prepares everything on your computer in a few minutes.

3
🔗 Link it to your AI chats

Tell your Claude setup to use this memory helper, and it starts working behind the scenes with one easy connection.

4
💬 Chat as usual

Keep having your coding conversations – every message from you and the AI gets saved automatically without you lifting a finger.

5
🩺 Check it's healthy

Run a quick health check to see everything is running smoothly and your memories are secure.

🧠 Your AI remembers you

Now your AI recalls details from past chats, asks fewer repeat questions, and feels like a true long-term partner.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 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 iai-mcp?

IAI-MCP is a Python-based, open-source memory system for AI coding assistants like Claude Code. It runs as a local daemon, automatically capturing every user and assistant turn verbatim via MCP protocol hooks, then injecting a relevant history slice at new session starts. Developers get persistent, cue-driven recall without manual prompts like "remember this" or cloud dependencies.

Why is it gaining traction?

It stands out as the best-benchmarked option, shipping runnable harnesses proving 99% verbatim recall, under-100ms latency at 10k records, and low token costs (3k warm, 8k cold). Fully local with AES-256 encryption and no telemetry, it preserves exact details over aggressive summarization in alternatives. Automatic consolidation during idle time adapts to your style without intervention.

Who should use this?

AI coding assistants users on macOS/Linux with Claude Code, needing history across sessions for ongoing projects like refactoring or architecture decisions. Ideal for solo devs or teams building custom MCP tools, where precise recall of past decisions beats re-explaining context.

Verdict

Try it if local AI memory fits your workflow—install is one script, doctor checks health, benchmarks verify claims. At 13 stars and 1.0% credibility, it's experimental with no Windows support or GUI, but solid docs and MIT license make it low-risk for tinkerers.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.