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

# Quickstart

> Sign in, connect your laptop, launch an agent and get a pull request in about 10 minutes.

## Before you start

You need:

* macOS. The one-command installer supports macOS only.
* **Node 20 or newer.**
* **Kiro CLI** installed and signed in (`kiro-cli whoami` works). Credits are billed to your own Kiro plan.
* **GitHub CLI** signed in to the org you work in (`gh auth status`). The runner clones with your login.
* Access to a Code-it project that has an Asana board.
* Optional, only if the agent deploys to AWS: a named AWS profile on your laptop.

**1. Sign in**

Open [app.aiml-code-it.automat-it.io](https://app.aiml-code-it.automat-it.io) and sign in with Google.

**2. Install the runner**

Open **Connect your laptop** in the web app (`/connect-laptop`). It shows the install command for this deployment. It looks like this:

```bash theme={null}
curl -fsSL https://app.aiml-code-it.automat-it.io/install.sh | sh
```

The script checks Node and `kiro-cli`, downloads the runner, verifies its sha256, installs it to `~/.code-it/runner` and registers a LaunchAgent so it starts at login.

**3. Approve this laptop**

The installer opens a browser page. Approve the short code shown in your terminal. Nothing is pasted. The runner stores its token in the macOS Keychain, or in `~/.config/code-it/runner.json` when the Keychain is not used.

When it connects, the web app shows your laptop as **online**. Check from a terminal with `code-it-runner status`.

**4. Create a project**

Open **Projects**, then **Create**. Set the name, the repository and base branch, and the Asana workspace and board. A project is the thing an agent launches against. See [Projects](https://docs.aiml-code-it.automat-it.io/concepts/projects.md).

> [!NOTE]
>
> Creating a project needs an admin role. If you cannot see Projects, ask an admin to create one or add you to an existing project.

**5. Launch your first agent**

Open **Agents**, pick a project and an Asana task, and press **Start**. You can also pick a git branch instead of a task. The API creates the agent as `waiting_for_runner`, then sends a launch message to your runner.

The runner clones the repo and the ticket branch into `~/code-it/<project-slug>/<ticket>`, writes the `.kiro/` workspace and starts `kiro-cli` on the task brief. The grid moves to **running**.

**6. Chat with the agent**

Open the agent. You see the conversation stream, tool calls and any question the agent is blocked on. Type a prompt to steer it. See [Chat features](https://docs.aiml-code-it.automat-it.io/guides/chat.md).

**7. Approve**

When the agent wants to run `terraform apply`, `cdk deploy`, `cloudformation deploy` or `sam deploy`, the `code-it-apply` gate turns the plan into a pending question. Review it in the web app and approve. The agent applies only after you approve.

If the grid shows `aws_session_expired`, run `aws sso login --profile <aws-profile>` and the next turn picks it up.

**8. Review the pull request**

When the work is ready, the PR hook opens a pull request on the ticket branch. The agents grid links to it.

## If something does not work

Run `code-it-runner doctor` and read [Install and troubleshoot the runner](https://docs.aiml-code-it.automat-it.io/guides/runner.md) and the [FAQ](https://docs.aiml-code-it.automat-it.io/faq.md).


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