Most AI visibility tools give you a dashboard. You log in, check the numbers, log out. Fine for reporting. Less useful when you want to do something about what you find.
Connecting AEO Copilot to Claude changes that. Instead of opening a dashboard, you ask a question: which prompts mention my competitors but not me? What's my visibility trend this month? What should I fix first? Claude pulls the data and answers directly.
This is how to connect them.
The API
Five endpoints, all GET requests, all returning JSON.
Authentication is a Bearer token. Every request needs this header:
Getting your API key
Go to Settings > API Keys and generate a key. It starts with aeo_. Keys don't expire, but you can revoke them at any time.
A basic request
Grab the brand id from the response and use it with the other endpoints.
Filtering results
The /results endpoint takes optional query parameters:
from and to take ISO dates. limit defaults to 100, max 500. topicId filters to a single topic.
The MCP server
MCP is the protocol that lets AI assistants connect to external tools. The AEO Copilot MCP server wraps the API so you can ask Claude about your brand's visibility data without leaving the conversation.
The server is open source: github.com/sofianbettayeb/aeo-copilot-mcp
Install in Claude Code
Install in Claude Desktop
Edit claude_desktop_config.json:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
- Windows:
%APPDATA%\Claude\claude_desktop_config.json
Restart Claude Desktop after saving.
What it looks like in practice
Once connected, you talk to Claude like you would any conversation:
- "What's my brand's AI visibility score?"
- "Which prompts mention my competitors but not me?"
- "Show me my visibility trend for the last 30 days"
- "What should I fix to improve my AI visibility?"
- "How does my brand compare to [competitor] in AI-generated answers?"
Claude calls the API and translates the response into plain language. You can follow up, filter by date, drill into a specific topic. It works the way you'd expect a conversation to work.
What's coming
The MCP server now covers writes too: create a brand, add topics and prompts, run them across all four LLMs, and run the technical scan. That means you can build a full client report from a single Claude Code conversation, which is the workflow in Build an AI visibility report for a client in minutes. Next on the roadmap: webhooks and pushing results into external tools.
If you build something on top of it, or hit something that doesn't work as expected, open an issue on GitHub: github.com/sofianbettayeb/aeo-copilot-mcp.
Frequently asked questions
Is there an AI visibility tool that works with Claude Code over MCP?
Yes. AEO Copilot ships an open-source MCP server. Add it to Claude Code with the one claude mcp add command above, and Claude can read your brand's visibility, results, and recommendations directly.
How do I check my brand's ChatGPT visibility from inside Claude Code?
Once the server is connected, ask in plain language: "What's my brand's AI visibility score?" or "Which prompts mention my competitors but not me?" Claude calls the API and answers in the conversation. No dashboard, no export.
Can I plug it into my own AI agent workflow?
Yes. Everything the MCP exposes is also a REST API, so you can wire AEO Copilot into any agent stack or automation, not just Claude. API and MCP access is on the Freelancer plan and up.
Get your API key →