glommer

glommer / memelord

Public

Memelord is an in-process agentic memory system

110
14
100% credibility
Found Feb 18, 2026 at 35 stars 3x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Memelord adds persistent memory to AI coding agents so they remember project-specific corrections, insights, and patterns across sessions to improve performance over time.

How It Works

1
💡 Discover Memelord

You notice your AI coding helper keeps forgetting project details between sessions, then learn about Memelord to fix that.

2
📁 Set up in your project

Go to your project's folder and run a quick setup to create a private memory space just for this project.

3
🚀 Restart your AI helper

Restart your AI coding assistant, and it instantly connects to the memory, feeling like it remembers everything.

4
🧠 See past lessons load

At the start of each session, your AI gets key reminders from before, like file locations or common fixes.

5
🔍 Get task-specific tips

When starting a new task, your AI pulls up the most relevant past experiences to guide its work.

6
📝 Save new learnings

As you work, correct mistakes or share tips, and your AI stores them safely for future use.

🎉 AI improves over time

Session after session, your AI makes fewer errors, works faster, and feels like a true project expert.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 35 to 110 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 memelord?

Memelord delivers persistent, per-project memory to coding agents, so they stop forgetting fixes, project quirks, and user instructions across sessions. Built in TypeScript with Turso for local storage and local embeddings, it captures insights via MCP tools like `memory_start_task` and `memory_report`, then refines them through agent ratings and decay. Run `memelord init` in your repo for instant setup with Claude Code hooks that auto-inject top memories.

Why is it gaining traction?

Unlike basic agent state or cloud RAG, memelord runs fully in-process with agentic feedback loops—agents self-rate memories, contradict bad ones, and build a weighted knowledge base that evolves. CLI commands like `memelord status` and `memelord search` let you inspect and purge, while benchmarks show it tackling SWE-bench tasks smarter over time. No APIs, no keys: pure local memelord tech that makes agents feel less amnesia-prone.

Who should use this?

AI coding agent users on mid-sized repos, like Rust DB devs fixing Turso bugs or Django maintainers debugging subqueries, where agents repeat file path hunts or tool fails. Ideal for Claude Code power users tired of re-explaining "use pnpm, not npm" every session.

Verdict

Early alpha with 18 stars and 1.0% credibility—solid README and benchmarks, but unproven at scale. Worth `npm install -g memelord` if you're deep in agentic workflows; skip for production until more battle tests.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.