agent-memory CLI
Long-term memory for AI agents: Markdown files as the source of truth, a SQLite index as a cache, shared by Claude Code, Codex and others
- 3,404 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Hooks read agent sessions and write memory
- Installed from a clone with no PyPI package
Install
- Works in
Manual install
git clone https://github.com/tigerless-labs/agent-memory.git
cd agent-memory
uv sync --all-packagesThe mem, mem-mcp and mem-hook commands appear in .venv/bin. Python 3.12 or newer and uv are needed.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Hooks read agent sessions and write memory
- Installed from a clone with no PyPI package
This is third-party code. Review the repository files before installing.
What it does
agent-memory keeps agent memory in plain Markdown files that remain the single source of truth. Next to them lies a SQLite index as a cache that can be deleted and rebuilt. The storage is shared by Claude Code, Codex CLI, Muse Code and any agent that can run a command. A cloned repository gives the `mem`, `mem-mcp` and `mem-hook` commands: there is an MCP mode and hooks. Python 3.12 or newer and uv are needed, and there is no PyPI package.
Who it is for. For people who work with several agents and want a shared memory that can be read, edited and kept in Git as ordinary files.
Good fit when
- You need shared memory for Claude Code and Codex in the form of Markdown files
- You need memory you can browse and edit by hand instead of hiding in a database
- You need an MCP mode or hooks so the agent writes and reads memory itself
Not a fit when
- You need a ready PyPI install: there is no package and it installs from a clone
- You are not ready for hooks to read sessions and write memory
Example request
Write into the project memory that we chose PostgreSQL and why, and find past decisions about the databaseLimitations
Hooks read agent sessions and write memory, so check what lands in the files. There is no PyPI release: install from a clone with `git clone`, `cd agent-memory` and `uv sync --all-packages`, and the commands appear in `.venv/bin`. Python 3.12 or newer and uv are required. The SQLite index serves as a cache and the source of truth is the Markdown files.
How to disable. Remove the mem-hook hooks from the agents' settings and delete the repository clone and the memory files folder if it is no longer needed.
FAQ
Where is memory stored?
In Markdown files, which are the source of truth. The SQLite index next to them serves as a cache and is rebuilt.
Which agents does it work with?
The storage is shared by Claude Code, Codex CLI, Muse Code and any agent that can run a command.
How do I install it?
From a clone: git clone, cd agent-memory, uv sync --all-packages. There is no PyPI package.
What is needed to run?
Python 3.12 or newer and uv.
Discussion
No questions or reviews yet.
Related
CLI toolsLocal persistent memory for coding agents: hooks record every session, and Claude Code, Codex or Cursor recall past decisions through 54 MCP tools
Long-term agent memory in local Markdown files with a link graph, semantic search and MCP access
Shared semantic memory for agents: conversations are saved as markdown and recalled by search across Claude Code, Codex and other agents
Long-term memory for agents: a remote MCP server, plugins for Claude Code, Cursor and Codex, and an API for your own apps
Long-term memory for agents: it distills knowledge from session history and brings it back in new tasks, one store shared across Claude Code, Codex, Cursor and
iai-memory
iai-memory (Personal Memory Engine)
A local long-term memory MCP server for agents: keeps conversations verbatim and returns the right context at the start of a session
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| agent-memory (this entry) | no | 3,404 | ||
| OpenWolf | no | 2,387 | shared local project memory with a project map and token tracking | |
| MCP Knowledge Graph | no | not required | 888 | memory in a local knowledge graph, .aim files |
| MemoraX Code | no | 2,192 | a memory plugin for architecture and repository rules, version 0.1.x |
README in short
The README describes a long-term memory runtime for AI agents: Markdown files, a SQLite index, the mem, mem-mcp and mem-hook commands and a shared store for several agents. Installation from a clone with uv. MIT license.