# Ask Claude or Cursor about your AI visibility

> Connect AEOmeter's remote MCP server to Claude, Cursor or any MCP client with OAuth, or use the REST API and OpenAPI spec to work with your AI visibility data.

AEOmeter runs a remote MCP server at `https://aeometer.com/mcp`. Connect it to Claude, Cursor or another MCP client, sign in with OAuth, and ask about your prompts, answers and competitors in plain language. The same operations are available as a REST API with an OpenAPI spec.

## Questions you can ask your assistant

The assistant calls AEOmeter tools for you and answers from your data. A few examples:

- "Which prompts mention PostHog but not us this week?"
- "Show me what Perplexity said for 'best product analytics tool for a seed-stage startup' and which sources it cited."
- "Which domains get cited most for my pricing prompts this month?"
- "Import these 20 prompts and tag them alternatives."
- "Is my browser ready to run checks?"
- "Set my daily schedule to 9:00 in Europe/Berlin."

## Connect it in a minute

Add a remote MCP server with the URL `https://aeometer.com/mcp` and sign in when your client asks. The OAuth scope is `mcp:tools`. Most clients that support remote servers accept a config like this:

```json
{
  "mcpServers": {
    "aeometer": {
      "url": "https://aeometer.com/mcp"
    }
  }
}
```

In Claude Code:

```bash
claude mcp add --transport http aeometer https://aeometer.com/mcp
```

In Claude on the web or desktop, add it as a custom connector with the same URL. Config formats vary by client, so check your client's docs if a field name differs. If your client caches tool lists, reconnect after AEOmeter releases.

## What the tools cover

The MCP tools use the same handlers and permissions as the web app. Every project tool checks that you own the project.

| Workflow | Tools |
| --- | --- |
| Projects and brands | `list_projects`, `get_project`, `create_project`, `update_project` |
| Prompts and tags | `list_prompts`, `create_prompt`, `import_prompts`, `add_starter_prompts`, `update_prompt`, `bulk_update_prompts` |
| Browser monitoring | `get_browser_status`, `get_schedule`, `update_schedule`, `list_runs`, `create_run`, `cancel_job`, `cancel_queued_jobs` |
| Results and evidence | `list_observations`, `get_analytics`, `get_answer`, `reanalyze_project` |

## What to know before you connect

Starting a run from your assistant queues checks. They are collected when your Chrome is connected, and at that point real prompts go to the AI engines. Browser status reports what your extension advertises. It does not verify that you are logged in to each engine.

Editing prompts or brands through MCP starts a new measurement baseline, exactly as in the web app. Credential management and the extension's collector endpoints are not exposed over MCP. Whatever the tools return goes to the AI client you connected, under that client's terms.

## REST API and OpenAPI spec

Everything in the dashboard is also available over HTTP under `/api/v1`. The OpenAPI spec lives at [/openapi.json](https://aeometer.com/openapi.json) and the browsable reference at [/docs](https://aeometer.com/docs). Use it for scripts, internal dashboards or your own exports.

## Questions

### Which MCP clients work?

Any client that supports remote MCP servers over HTTP with OAuth sign-in. That includes Claude, Claude Code and Cursor.

### Can the assistant start checks?

It can queue a run, optionally for selected prompts. The checks run in your Chrome when the extension is connected and the engines are signed in. Nothing runs on a server on your behalf.

### Do I need an API key for MCP?

No. Your client opens an AEOmeter sign-in page and receives an OAuth token with the `mcp:tools` scope.

### Is the MCP data the same as the dashboard?

Yes. MCP and the web app share the same schemas and handlers, so the numbers match what you see in the app.

---

- Web page: https://aeometer.com/features/mcp
- Start tracking your brand: https://aeometer.com/sign-up
- All pages for AI agents: https://aeometer.com/llms.txt