MADRE CLI
MADRE puts Codex, Claude Code, Gemini CLI, Cursor and OpenCode into one local chat with shared memory and permission modes.
- 127 ★ on the repository
- Last commit
Broad access
Why this level
- In mode #4 agents run commands and can git push and deploy with your sessions.
- Project content goes to the providers of the connected agents.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/madreDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- In mode #4 agents run commands and can git push and deploy with your sessions.
- Project content goes to the providers of the connected agents.
This is third-party code. Review the repository files before installing.
What it does
MADRE starts with `npx @jossuealcala/madre`, opens in the browser on `127.0.0.1` and joins already installed coding agents into one chat inside a project. An agent reads what the others said, and the user decides what each may change. Decisions are kept in project memory. Every message carries a permission mode from `#0` to `#4`, from an incognito mode with no ledger entry to full access with command execution. Before edits MADRE takes a restore point, then shows the change list and an undo button. It also offers multi-agent plans with steps, a module market and a diagnostic via `npx @jossuealcala/madre doctor`.
Who it is for. Developers who keep several coding agents and want to compare their answers in one project without copying context between terminals.
Good fit when
- The same code question goes to Claude Code and Codex and the answers are needed side by side.
- An agent enters a project read-only and edits are enabled as a separate decision.
- An orchestrator hands plan steps to different agents with different rights.
- Project decisions must survive session close and be shared across agents.
Not a fit when
- A larger organization uses MADRE at work: the BSL license requires a commercial agreement.
- You need a cloud service with accounts and team features.
- You use a single agent and need no permission separation.
Example request
@claude explain why the build fails, @codex propose a fix. Mode #1, change nothing, then show how the answers differ.Limitations
Node 22.5 or newer and at least one signed-in agent are required. Windows runs through WSL2. The BSL 1.1 license is free for individuals and teams up to 10, larger organizations need a commercial one; it is not open source, and each version turns into AGPL v3 after four years. Whatever an agent reads goes to its provider, and madre.run receives a daily anonymous ping plus diagnostics that can be disabled. The project is in public beta, last commit on 5 November 2026.
How to disable. Stop the MADRE process and delete project data: `.madre/` and `.pulse/` in projects and `~/.pulse/`. Each agent's sign-in lives with the agent itself.
FAQ
What do modes #0 to #4 mean?
`#0` incognito, `#1` read-only (the default), `#2` create new files, `#3` edit the project, `#4` full access with commands. `#3` and `#4` unlock with a separate key.
Can agent edits be undone?
In `#3` a restore point is made before the turn and the undo button restores files. Anything sent outside, such as `git push` in `#4`, cannot be taken back.
Is a MADRE account required?
No. The app runs locally and agents sign in with their own accounts.
How do I check that agents are connected?
Open the diagnostics section or run `npx @jossuealcala/madre doctor`.
What license is MADRE under?
Business Source License 1.1. Free for individuals and teams up to 10; larger organizations need a commercial license.
Discussion
No questions or reviews yet.
Related
CLI toolsA local control panel for coding agents: it runs Claude Code and Codex, keeps a task queue and merges verified work into main.
Mjolnir runs Claude Code, Codex and other agent sessions in a terminal, in containers and over SSH, and keeps them running.
A CLI that indexes memory, rules and sessions of Claude Code, Codex, Cursor and Gemini and hands durable decisions to every agent
A local dashboard for running many agents across machines: memory, code graph and agent-to-agent messaging
A tmux-based TUI that runs Claude Code, Codex and other agents in one list: status, quick prompts, worktrees and diff review
A memory and coordination layer for coding agents: specs and issues live in git while agents work in parallel
README in short
The app describes itself as a shared chat over agents you already use. Keys are not copied; each agent keeps its own session.