harrylettering

harrylettering / CodeMemory

Public

A memory layer for Claude Code

11
0
100% credibility
Found May 03, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A local-first plugin for Claude Code CLI that stores engineering decisions, constraints, failures, and conversation summaries in SQLite for retrieval during long coding sessions, refactors, and debugging.

How It Works

1
🔍 Discover CodeMemory

While coding with your AI helper Claude Code, you learn about CodeMemory, a tool that remembers your past decisions and mistakes to make coding smoother.

2
📦 Install easily

Add the memory tool from the built-in marketplace with a few simple steps, and it sets up your personal memory storage automatically.

3
🚀 Start your coding session

Open Claude Code in your project, and CodeMemory quietly starts watching and saving key moments like decisions, errors, and fixes.

4
💾 It remembers everything important

As you code, it saves summaries of tough problems solved, choices made, and constraints, keeping your history safe and searchable.

5
⚠️ Get helpful reminders

Before repeating a past mistake, like editing a tricky file or running a failing command, it warns you with details from before.

✅ Code faster and smarter

You build projects quicker, avoiding old pitfalls and staying focused on new ideas, with your AI always remembering what worked and what didn't.

Sign up to see the full architecture

4 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 CodeMemory?

CodeMemory is a TypeScript plugin for Claude Code, acting as a memory layer for AI agents that persists decisions, constraints, failures, and summaries in local SQLite. It solves context loss in long sessions by injecting relevant history into user prompts and warning of prior failures before edits or bash tools. Developers get durable recall across refactors and debug loops without rebuilding context each time.

Why is it gaining traction?

It hooks directly into Claude Code's lifecycle—prompt submission, tool use, session start/end—for seamless retrieval via a fast per-session daemon over Unix sockets. Standouts include prior-failure alerts, DAG-based compaction to avoid github memory limit bloat, and tools like codememory_mark_decision or /codememory-status for quick inspection. No external services needed, just marketplace install and env vars.

Who should use this?

Claude Code users in extended refactors where design rationale fades, or debugging loops repeating failed fixes on specific files/commands. Ideal for backend engineers managing complex state, or AI agent builders needing a memory layer for llm without copilot's session resets. Skip if you're on short tasks or broad RAG needs.

Verdict

Worth installing for Claude Code heavyweights despite 11 stars and 1.0% credibility—docs are thorough, benchmarks hit sub-200ms lookups, but watch for edge cases in untested scale. Strong start on memory layer for claude code; fork and contribute if it clicks.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.