Connect an AI agent

Let an AI coding agent like Claude Code or Cursor read your PageProofer notes, fix the issues in your code, and report back on the note.

PageProofer has an MCP server, the standard way for AI agents to connect to other tools. Once it’s connected, you can ask your agent things like “fix the open notes on the Northwind site” or “what’s wrong with note 12?”. The agent can:

• list your sites and their open notes, filtered by status, #tags, page, assignee or overdue
• read a note in full: the feedback, the page it’s on, the element it was pinned to, the reporter’s browser and window size, console errors and failed network requests, and every comment
• look at screenshots and read attached text files
• comment on the note, for example to say what it changed and in which commit
• change the status, priority, assignee or due date

AI agent access is available on all paid plans.

Create a token

1. Go to My Settings and open the AI agents tab.
2. Give the token a name you’ll recognise, such as “Claude Code on my laptop”, and click Create token.
3. Copy the token straight away. For your security it’s only shown once.

You can create as many tokens as you like, for example one per computer or tool, and revoke any of them at any time. The tab shows when each token was last used.

Connect Claude Code

The AI agents tab gives you the full command with your token filled in. It looks like this:

claude mcp add --transport http pageproofer https://app.pageproofer.com/mcp --header "Authorization: Bearer YOUR_TOKEN"

Run it in your project folder, then start Claude Code and ask it about your notes.

Connect other agents

Any agent that supports MCP servers over HTTP can connect. Add a server with:

• URL: https://app.pageproofer.com/mcp
• Header: Authorization: Bearer YOUR_TOKEN

Using the agent

Once your agent is connected, talk to it in plain language. For example prompts and tips on finding notes, tracking down the code behind them, fixing issues and reporting back, see How to use the AI agent.

What the agent can and can’t do

The agent acts as you. It sees the same sites and notes you do, and its comments and changes show your name. In the note’s Activity tab, changes made by an agent are marked “via AI agent”.

Agents can’t mark a note Complete or Closed. When an agent has fixed something, it moves the note to Review so a person can check the fix and close it.

Notifications work as usual: a comment from an agent emails the people on the note just like any other comment.

Keeping your token safe

Treat a token like a password. Anyone who has it can read and comment on your notes as you.

• Don’t commit it to a code repository or share it in chat.
• If a token might have leaked, revoke it in My Settings > AI agents and create a new one.
• Tokens stop working if your account’s subscription ends.