Three commands to shared memory
Install the CLI, sign in, and connect your agents. Memkith runs as an MCP server your coding agents talk to, so the setup is done in a terminal once per machine and once per repository.
Before you start
- Homebrew, which is how the CLI is distributed today.
- A Git repository. Branch and commit scope are what give a memory its address.
- A Memkith account. Create one if you have not.
Run these in order
The first command is per machine. The second is per machine and account. The third is per repository, and it is the one you run again whenever you add an agent.
01
Install the CLI
Installs the command-line tool and the MCP server that connect Memkith to your coding agents.
02
Sign in
Opens a browser to connect this machine to your account. Terminals without a browser take --no-browser, which prints the sign-in URL instead.
03
Connect your agents
Prints the MCP setup for every supported agent. Name one instead of passing --all to set up a single agent.
One command per agent
integrate prints the native MCP configuration for an agent so you can paste it where that agent keeps its settings. Pass a name to set up one, or --all for every one of them.
- claude
- Claude Code
- cursor
- Cursor
- codex
- Codex
- zed
- Zed
Any other MCP-capable agent can be pointed at the same server by hand. These four are the ones integrate writes configuration for.
The rest of the CLI
The three commands above are the whole install path. These are what you reach for afterwards, and memkith --help stays the authority.
Report whether this machine is signed in, and to which account.
--refresh forces a token refresh before reporting.
Install a post-commit hook so pending memory anchors fill in automatically as code lands.
Stamp the commit anchor on memories whose scoped code has since been committed. Idempotent, so it is safe to run at any time.
--quiet prints nothing, for use from the post-commit hook.
Sign this machine out.
Print the installed version.