How to Get Comments on a Cursor Canvas (2026)

By Marcin, founder of Markloop · Last verified September 2026

What shared canvases can and can't do

Cursor shipped shared canvases on 20 May 2026, and inside a team they work well: hit Publish, copy the link, and a teammate opens a live snapshot in the browser. Cursor's docs say only members of your team can open a share, and publishing needs a paid plan and a team. So there's no link you can send a client or a beta user. Cursor staff said on the forum in June they're "looking at ways to expand this", but it's still team-only.

In practice a teammate opens the canvas, spots a problem in the pricing table, and tells you in Slack. A client or a beta user can't open it at all, so they get a screenshot and reply by email. Either way you re-type their words into Cursor, guess which element they meant, and do it again next round.

Turn the canvas into an HTML file with one prompt

There's no export button on the canvas toolbar, and the canvas source won't help. According to threads on the Cursor forum, Cursor keeps it outside your repo, as a .canvas.tsx file under ~/.cursor/projects/<workspace>/canvases/. It imports from cursor/canvas, which exists only inside the IDE, so the file doesn't render anywhere else.

The agent that built the canvas can write the same content out as plain, self-contained HTML:

Save this canvas as a standalone HTML file in the project root.

Open the file in a browser before you send it, to check nothing got lost on the way out of the canvas.

Get comments on it from inside Cursor

Markloop publishes the HTML file at one link, people comment on it in the browser, and your agent pulls the comments back into Cursor chat. The next version goes up at the same link.

1 · Connect Markloop to Cursor

Add one entry to ~/.cursor/mcp.json. Cursor opens a browser sign-in, and Settings → MCP then shows Markloop as Connected.

{
  "mcpServers": {
    "markloop": {
      "url": "https://app.markloop.io/api/mcp"
    }
  }
}

2 · Publish the spec and send the link

Say your canvas is an onboarding spec and your designer isn't on your Cursor team. In the chat:

Save this canvas as onboarding-spec.html and publish it to my Markloop project.

Turn on commenting for the link and send it. Your designer types a name and comments in the browser, no account.

3 · Pull the comments and publish v2

When comments come in:

Pull the comments on the onboarding spec, make the changes, and publish the next version.

Each comment arrives with the text it was left on and the element it points at. The agent marks which ones it addressed, and the rest stay open for the next round.

No setup needed to try it: download the HTML file your AI made and upload it to Markloop. Or connect Claude, ChatGPT, Cursor or Grok once (setup is a one-time step, shown in the app after signup) and ask your AI to publish. Watch the 2-minute walkthrough.

Sending the doc to a client instead? See how to write a client proposal with AI and get comments on one link. For who can see your link and where your data lives, see the FAQ.

FAQ

Related: What Claude artifacts are · Claude artifact sharing options compared · How to share Claude artifacts outside your organization · HTML is the new Markdown - reviewing it is the missing half