ctxr-dev

ctxr-dev / memory

Public

Inspectable local project memory for AI coding agents.

agent ai dify hooks llm
11
0
100% credibility
Found May 12, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

A local memory system using structured knowledge storage to help AI coding agents retain project details, lessons learned, and avoid repeating mistakes.

How It Works

1
🔍 Discover AI memory boost

You hear about a helpful tool that lets your AI coding assistant remember project details and stop repeating mistakes.

2
📥 Add to your project

You copy the memory folder into your project's main directory to get started.

3
🛠️ Run easy setup

You follow simple steps or ask your AI to handle the initial preparation for you.

4
🚀 Launch the service

With one command, you start the background helper that powers the memory.

5
🔗 Link your AI tools

You connect it to your AI apps like Claude or Cursor so they can use the memory.

6
📚 Share project knowledge

You let it absorb your documents and session notes to build a smart knowledge base.

🧠 AI learns and remembers

Your AI now recalls past lessons, avoids old errors, and gets better with every use.

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 memory?

Local Dify-powered RAG for AI coding agents, storing project knowledge as typed atoms like decisions, bug causes, and self-improvement lessons. Hooks distill sessions via flush-compile pipeline, exposing MCP tools (search_memory, recall_lessons, save_lesson) over stdio for Claude, Cursor, Copilot. JavaScript/Docker setup runs inspectable memory github claude or copilot style, cutting repeat errors in agent workflows.

Why is it gaining traction?

Ditches noisy transcripts for deduped, metadata-filtered atoms—precise retrieval without embedding bloat. Self-improvement dataset auto-checks corrections pre-task; on-demand absorb_files indexes docs. Git-pull updates and client snippets make it a drop-in github memory manager for llm agents.

Who should use this?

Backend devs using Cursor/Claude Code on evolving codebases, frontend teams with Copilot handling form quirks, or anyone building coding agents needing github mcp memory beyond session limits. Suits mid-size projects where agents forget fixes across restarts.

Verdict

Worth cloning for memory github llm experiments—early at 11 stars and 1.0% credibility, but thorough docs, tiered verification, and MCP smoke tests signal polish. Restart agents post-setup; scale cautiously until stars climb.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.