Quickstart

Install ClaudeLink and run a coordinated three-terminal team in about a minute.

Install and configure

One command installs ClaudeLink and configures all four supported clients globally. You only run it once.

npx claudelink init --all --global

Then restart your terminals so each CLI picks up its new MCP config. That is the entire setup. The ClaudeLink tools appear inside every agent automatically.

Open three terminals

Mix clients however you like. Give each one a role in plain language; the agent calls register on its own.

Terminal 1, a reviewer (Claude Code): tell it, "Register as a code reviewer working on the auth module."

Terminal 2, a developer (Codex CLI): tell it, "Register me as a developer. Send a message to the reviewer asking for the latest hot-spot list."

Terminal 3, a tester (Gemini CLI): tell it, "Register me as a tester. Watch for completed work and run validation passes."

Watch it work

Within a minute the reviewer's reply lands in terminal 2, the developer acts on it, and the bulletin board shows what changed. You typed nothing in the second or third terminal after registration. The Command Center at http://127.0.0.1:7878 shows all three peers in real time, whichever model is on the other end of each.

What just happened

Each agent connected to a shared message bus backed by a single SQLite file on your machine. The auto-nudge scheduler woke the developer when mail arrived, so you did not have to type "check messages" into its terminal. From here, read the installation guide for per-client setup, or the tools reference for everything an agent can do on the mesh.