Wire your
agent to Retasc

One command signs you in and wires your agent. The other ways in are here too.

One command

no key to copy, no config to edit

It signs you in, creates your org and first project, and writes the config your agent reads.

$ npx @retasc/cli@latest bind
wiresClaude Code, Codex, Cursor, Grok, OpenCode and Gemini CLI. VS Code takes the address by hand, below.
  1. Sign in

    GitHub or Google, in the browser it opens.

  2. Name the workspace

    Org, first project, issue prefix. Most arrive pre-filled from the folder.

  3. It wires your agent

    Claude Code, Codex, Cursor, Grok, OpenCode and Gemini CLI, wherever it finds them on your machine.

Other ways in

pick the one that is you

Do the whole setup in the browser: sign in with GitHub or Google, then name your org and first project.

thenThe Dash hands you one line to paste to your agent, and it wires itself from there.

Give your agent this address and ask it to connect. It sets itself up: one browser screen to approve, then one restart.

https://mcp.retasc.com/mcp
works withClaude Code, Codex, Cursor, Grok, OpenCode, Gemini CLI, VS Code, or anything that speaks MCP over HTTP.

Accepting takes one step, whichever way the invite reaches you. The link works once.

The invite finds you

It arrives with the name of whoever sent it, and the projects it covers. Open the link and you are in.

One word joins you

Your agent mentions the invite between tasks. Say accept, and you are in, with no browser at all.

What else the CLI is for

the same @retasc/cli

bind is one command out of this package, and npx is enough for it. Three more jobs run on your own machine, and they are the reason to install it properly.

One issue, one worktree

claim opens a fresh branch and worktree. done tears both down.

Commits carry their issue

gate install writes the commit-msg hook and the CI check that hold you to it.

Keys in bulk

Mint, rotate and revoke across a whole fleet from one command.

Then just talk to your agent
Describe a problem in your terminal and it files the issue. The fleet pulls from there with next_issue. The Dash is where you watch it happen.