> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getcliq.io/llms.txt
> Use this file to discover all available pages before exploring further.

# What is Cliq?

> Ready-made AI teams. Your requirements. No code. From first install to production pipelines in three commands.

**Ready-made AI teams. Your requirements. No code.**

cliq is a multi-agent orchestration tool. Assemble a team of specialized AI agents, point them at a requirement, and run. Three commands. Production-quality output.

```bash theme={"theme":{"light":"github-light","dark":"one-dark-pro"}}
cliq init
cliq assemble @cliq/hello-world
cliq run
```

(With the daemon, `cliqd`, running — see [Get started](/get-started).)

***

## Why cliq

**Ready-made teams, not a toolkit.** Other multi-agent frameworks require you to write Python to wire agents together. Install teams from CliqHub — TDD pipelines, feature development with git automation, human-in-the-loop review, and more — then assemble and run.

**Build your own.** Need a custom team? Build it in the CliqHub team builder, or write `team.yml` and role files by hand. Check it with `cliq team validate`, then install or publish it with `cliq team install` / `cliq team publish`.

**Configure, don't code.** Under the hood, roles are markdown files, workflows are YAML DAGs, and gates are checks the process must pass. Everything is inspectable and editable. No SDK required to start. No classes. No framework abstractions in your way.

**Built to share.** Teams are portable packages — roles, workflow, and gates bundled together. Build a team once. Use it on your laptop, share it with your org, publish it for others. CliqHub is the registry where teams are discovered and installed.

**Any domain.** Official engineering teams ship on CliqHub, but the model is domain-agnostic: phases, dependencies, and gates. The same primitives power content pipelines, research workflows, compliance reviews, or any process where specialist agents collaborate.

***

## What a run feels like

1. You attach a team to your project
2. You start a run with your requirement or inputs
3. Specialists do their part, in dependency order
4. Quality checks decide whether the work continues, goes back for revision, or needs a person
5. You inspect what happened — status, logs, and the result of each phase

When something fails, you can see *which phase* failed and *what it produced* — not scroll a buried chat.

***

## Start here

<Columns cols={2}>
  <Card title="Get started" icon="rocket" href="/get-started">
    Install Cliq and finish your first run in a few minutes.
  </Card>

  <Card title="Concepts" icon="book-open" href="/concepts">
    Phases, agents, roles, gates — how a team is built.
  </Card>

  <Card title="Running teams" icon="play" href="/running-teams">
    Assemble, pass inputs, run, inspect, and resume.
  </Card>

  <Card title="Agents" icon="bot" href="/agents">
    Built-in agents, settings, and phase inputs.
  </Card>

  <Card title="Auth" icon="key-round" href="/auth">
    Sign-in, user tokens, and machine enroll tokens.
  </Card>

  <Card title="Local development" icon="laptop" href="/local-development">
    Build and test teams locally with cliqd --no-login.
  </Card>

  <Card title="CLI" icon="terminal" href="/cli">
    Commands and flags when you already know what you want.
  </Card>
</Columns>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.