Getting started

CCD is an autonomous AI development fleet. Instead of a chat window that suggests code, you get dedicated Linux nodes, real machines assigned to your company, each running a persistent AI agent that writes, tests, and ships software on your behalf. You describe the outcome; the fleet does the work.

Create your account

  1. Go to Sign up and choose a plan. See Billing & plans for what each includes.
  2. Complete checkout. Your company workspace and node allocation are provisioned automatically.
  3. Check your inbox: the welcome email contains your company API key. Keep it safe. It grants programmatic access to your workspace (see API access).
  4. Log in at /login with the credentials you created.

A tour of the dashboard

After logging in you land on the dashboard. The sidebar is organized into areas:

Area What lives there
Intelligence The Agent, where you run tasks, plus Security, Worlds, Builder, Arcade, Residency, Sessions, Analytics, and Workspace
Projects App Store, Project Manager, and (for team owners) Companies
Operations Missions, Intel, and Assets
Tools Terminal, Plans, Usage & Billing, Themes, and the API Reference
Logs History of past sessions and runs

Your first task

  1. Open the Agent page from the sidebar.
  2. Type what you want built: a real outcome, not a code snippet. For example: "Create a REST API for a todo app with SQLite storage and tests, then run the test suite."
  3. Send it. The agent plans the work, executes it on one of your nodes, and streams progress live: files written, commands run, tests passing.
  4. Some sensitive actions (like executing shell commands, depending on the agent mode) pause for your approval before running.

When the task finishes you can browse the results in the Workspace, review the run in History, or keep iterating in the same session. The agent remembers the context.

Where to go next