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

# Quickstart

> Give an AI agent access to pay-per-use tools with Cralo.

This guide takes your AI agent from a clean terminal to its first successful Cralo call.

## Prerequisites

* Node.js 22 or newer
* npm
* A [Cralo account](https://app.opentooler.ai) for interactive agent setup

Cralo supports macOS, Linux, WSL, and native Windows.

<Steps>
  <Step title="Install the CLI">
    ```bash theme={null}
    npm install --global @cralo/cli@stable
    ```
  </Step>

  <Step title="Check your setup">
    ```bash theme={null}
    cralo doctor --json
    ```

    Confirm that `connectivity.reachable` is `true`.
  </Step>

  <Step title="Sign in">
    ```bash theme={null}
    cralo auth login
    cralo auth status --json
    ```

    Top-level aliases exist for both commands: `cralo login` for `cralo auth login`, and `cralo status` for `cralo auth status`.
  </Step>

  <Step title="Give the agent a first tool">
    ```bash theme={null}
    cralo tools describe opentooler.ping --json
    cralo tools call opentooler.ping --input '{}' --json
    ```

    `opentooler.ping` is a deterministic, zero-price tool that verifies the full execution path.
  </Step>
</Steps>

## Next steps

* [Teach your agent to discover and call tools](/docs/guides/discover-and-call)
* [Set spending and safety controls](/docs/guides/billing-and-safety)
* [Give your agent the CLI reference](/docs/cli/overview)

<Warning>
  Do not put an API key in an agent prompt or command-line argument. For
  unattended agents, provide `CRALO_API_KEY` through a secret manager or
  controlled environment.
</Warning>