Claude-Mem: Carry your coding agent’s project memory into the next session
Added · Updated
Made by Alex Newman
View profile on GitHub (opens in a new tab)A plugin that records what an AI coding agent did, turns those observations into searchable summaries, and brings the relevant decisions and fixes into a later session.
It is built around Claude Code but the project also documents installers or integrations for OpenCode, Codex, Gemini, Hermes, Copilot, OpenClaw and other agents.
What it does
An agent session ends with a familiar problem: the next session has no idea what was already tried, which decisions were deliberate, or where the last fix landed. Claude-Mem installs hooks that observe tool use, stores the observations, and uses an AI provider to compress them into durable summaries. When a new session starts, semantic search selects the memories that match the current work and injects them into the agent’s context. There are also search tools for asking about older work and a Smart Trash path for restoring deleted context. The repository keeps the data in SQLite locally, while its optional hosted observer and cloud-sync paths are separate choices during setup.
You could use it to…
- Pick up a project next week with the earlier decisions still in view
- Search what your agent did in older sessions
- Start a new session with the relevant fixes from the last one loaded
Can I use this?
You'll need Node.js 20+, a supported coding agent and the project installer · Setup needed
Worth knowing The installer can use the project’s hosted memory service, your own provider key or your agent’s plan, so the cost and privacy boundary depends on that choice. Local context is stored in a SQLite database on your computer. Apache 2.0. Licensed under Apache License 2.0.
Why it's here
- 7,482 people starred it on GitHub this week — 114,706 in total.
Numbers as of 8 November 2026, from BuildTube’s latest daily refresh.
Behind it
View sourceBuildTube has not run or verified this project. Everything above is written from what the creator published.
Made this?
Quick questions
- What is Claude-Mem?
- An agent session ends with a familiar problem: the next session has no idea what was already tried, which decisions were deliberate, or where the last fix landed.
- What can I use Claude-Mem for?
- Pick up a project next week with the earlier decisions still in view; Search what your agent did in older sessions; Start a new session with the relevant fixes from the last one loaded
- Can I run Claude-Mem on my computer?
- It runs as an add-on inside an AI assistant you already use.
- What do I need to use Claude-Mem?
- Node.js 20 or newer and a supported coding agent. The installer can set up Bun and uv when needed, then registers the hooks and worker; choose an AI provider during installation or use the agent plan where supported.
- What license does Claude-Mem use?
- Licensed under Apache License 2.0.
More from this week
- RAGFlow Ask questions about your documents and get cited answers
Upload documents and ask questions with traceable citations.
- DeepTutor Study your own documents with a personal AI tutor
Read documents with page-by-page citations, practice daily and get video lessons from an AI tutor.
- Garden Skills Give your coding agent video and web design skills
Install ready-made skills so Claude Code, Cursor or Codex can make short videos, web pages and articles.
Explore by topic
Post this
Download share imageFor makers
Made this? Show it’s on BuildTube — add the badge to your README or embed the card on your site.
README badge (Markdown)
[](https://buildtube.app/p/gh-thedotmack-claude-mem/)Embed card (HTML)
<iframe src="https://buildtube.app/embed/gh-thedotmack-claude-mem/" width="560" height="240" style="border:0;max-width:100%" loading="lazy" title="Claude-Mem on BuildTube"></iframe>