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

# Quickstart: Get answers in 1 min

> Set up MCP and ask your first marketing question.

Start here if you want answers from OpenRush inside an AI agent.

You can use the API directly later. You do not need to start there.

<div className="or-hero-line or-hero-line-small">
  Connect OpenRush to <span className="or-pill or-pill-violet">Claude</span>, <span className="or-pill or-pill-cyan">ChatGPT</span>, <span className="or-pill or-pill-green">Codex</span>, or any tool that supports MCP.
</div>

## Prerequisites

You need three things:

* An OpenRush account
* Access to [openrush.io/login](https://openrush.io/login)
* An AI client that can connect to MCP, such as Claude, Codex, or another MCP-capable tool

## Set up MCP

<Steps>
  <Step title="Open the Dashboard">
    Go to [openrush.io/login](https://openrush.io/login) and sign in.
  </Step>

  <Step title="Open Setup">
    In the Dashboard, open [Setup](https://openrush.io/dashboard/connect).

    This page shows the MCP URL and setup instructions for supported clients.
  </Step>

  <Step title="Connect your AI client">
    Follow the instructions for your client.

    For Claude, add OpenRush as a custom connector. For Codex, add the MCP URL to your Codex config and run `codex mcp login openrush`.
  </Step>

  <Step title="Approve access">
    Complete the browser sign-in and approve OpenRush access when your AI client asks.
  </Step>

  <Step title="Ask your first question">
    Try a plain-English question:

    > Analyze `stripe.com`. Who are its organic competitors, what keywords matter, and what should I look at next?
  </Step>
</Steps>

## What happens after setup

After setup, your agent can call OpenRush when it needs marketing data.

<div className="or-feature-grid or-feature-grid-compact">
  <div className="or-feature-card">
    <span className="or-feature-icon or-feature-blue">1</span>
    <span>You ask a normal question</span>
  </div>

  <div className="or-feature-card">
    <span className="or-feature-icon or-feature-violet">2</span>
    <span>Your agent calls OpenRush</span>
  </div>

  <div className="or-feature-card">
    <span className="or-feature-icon or-feature-green">3</span>
    <span>You get a useful answer</span>
  </div>
</div>

## Good first questions

Use normal language. Your agent can decide which OpenRush tool to call.

* "Find my top search competitors for `example.com`."
* "Find keyword gaps between my site and these competitors."
* "Inspect the SERP for `best crm for startups`."
* "Audit my website and tell me what to fix first."
* "Which page on this competitor site is winning the most traffic?"

## Check usage and credits

Use the Dashboard for account details:

* [**Activity**](https://openrush.io/dashboard/usage): recent API and MCP requests.
* [**Billing**](https://openrush.io/dashboard/billing): credit balance and purchases.
* [**API access**](https://openrush.io/dashboard/api): API keys for server-side use.

## Developer option

If you are building a server-side integration, create an API key in [API access](https://openrush.io/dashboard/api), then call the REST API directly.

```bash theme={null}
curl -X POST "https://api.openrush.io/v1/tools/inspect_domain" \
  -H "Authorization: Bearer or_your_key" \
  -H "Content-Type: application/json" \
  -d '{"domain":"stripe.com"}'
```

Developers should also call `/v1/capabilities` to see which tools are enabled:

```bash theme={null}
curl "https://api.openrush.io/v1/capabilities"
```

<Tip>
  Most users should start with MCP in the Dashboard. Use direct API calls only when you are building your own software around OpenRush.
</Tip>