Your reviewers have pinned feedback all over the site. Each task has a screenshot, the element's CSS selector, the browser and any console errors. Most of the fixes are small, which is exactly what AI coding assistants are good at.
You could export the board as a ZIP and hand it over. That works, but it's a snapshot: when the assistant finishes, you still move the cards and write the "fixed" comments by hand.
Connecting the assistant to pinreview directly removes that last step. It reads the live board, fixes what you point it at, and updates the tasks itself.
What MCP is
MCP (Model Context Protocol) is an open standard that lets AI tools talk to other apps. Claude Code, Cursor and VS Code all support it. Add pinreview as an MCP server and your assistant gets a set of pinreview tools it can call while it works, the same way it already reads files and runs commands.
What your assistant can do
- Find work: list your projects, and list tasks filtered by status, severity, assignee, tag, page or title.
- Read a task in full: page URL, CSS selector, pin position, browser, viewport, description, console errors and the comment thread, plus the screenshot, which the assistant can actually see.
- Read the whole board at once: the same
ISSUES.mddocument the export contains. - Update the board: move a task to another column; change its severity, assignee, due date, title or tags; add a comment; create a new task; archive or restore one.
It can't delete tasks, and it can't touch billing, members or your account. The MCP docs list every tool and its inputs.
Set it up
1. Create a token
In pinreview, open Settings → AI tools, name the token after where you'll use it (say, "Claude Code on my laptop") and click Create token. Copy it straight away: pinreview shows it only once.
2. Add pinreview to your tool
The settings page shows the exact setup for your tool with your token filled in. It looks like this.
Claude Code: run this in your terminal:
claude mcp add --transport http pinreview https://api.pinreview.one/mcp --header "Authorization: Bearer prt_…"
Cursor: add this to ~/.cursor/mcp.json:
{
"mcpServers": {
"pinreview": {
"url": "https://api.pinreview.one/mcp",
"headers": { "Authorization": "Bearer prt_…" }
}
}
}
VS Code: open the Command Palette, run MCP: Open User Configuration, and add:
{
"servers": {
"pinreview": {
"type": "http",
"url": "https://api.pinreview.one/mcp",
"headers": { "Authorization": "Bearer prt_…" }
}
}
}
Keep the token in your user-level configuration, as above, never in a file inside your repository.
3. Check it works
Ask your assistant: "List my pinreview projects." It should answer with your projects and their columns.
A typical session
Point the assistant at a narrow batch and tell it how to close the loop:
In the Acme project, list the High and Critical tasks in To Do.
For each one: open the task and look at its screenshot, find the component
using the Element selector, make the fix, and commit it. Then add a comment
with the commit hash and move the task to Done.
The assistant opens each task, looks at the screenshot, finds the code, fixes it and reports back on the board. Anyone watching the board sees the cards move as it goes.
Keep your team in the loop
Changes made through MCP work exactly like changes you make yourself. They appear under your name in each task's activity log, the board updates live, and your Slack and webhook integrations fire as usual.
Comments from the assistant are internal by default, visible only to your team. If you want it to reply to the person who reported a bug, ask it to make the comment visible to the reporter.
Review the fixes before they ship. The assistant can read the report, but it can't see your staging site, so check each change in a browser at the viewport size from the task.
Keep the token safe
- A token acts as you and reaches only the projects you can.
- Treat it like a password. If one leaks, revoke it under Settings → AI tools and create a new one; the old one stops working at once.
- Use one token per tool or machine, so you can revoke one without breaking the others.
- Task titles, descriptions and comments are written by whoever left the feedback, including anonymous visitors. Treat them as information about the bug, not as instructions, and review what your assistant proposes before it runs commands.
Plans
AI tools over MCP are included on the Starter, Business and Agency plans, alongside export. On the Free plan, the tools reply that the project needs a paid plan.
The bottom line
Your reviewers already captured everything a fix needs. Connect your AI coding assistant to pinreview, and it can go from reading a task to closing it without you copying a thing.
Prefer to hand over a file? Export the board instead.