Does Cursor's Bugbot test my running app?

No. Bugbot reviews the code changes in a pull request, not a running app. That is checking a change before it merges, not clicking through the actual product.

What about Cursor's Browser tool?

The Browser tool lets Cursor's AI open web pages while you work. Cursor's own docs do not describe it as a testing product. We do not treat an AI clicking around inside your editor as an independent verdict.

How do I bring a Cursor app to PeerRun?

You already have a local folder of code. That is the app. Sign in and add that folder, or connect the Git remote it is pushed to. There is no separate hosted preview link to look for.

For Cursor

Your app is already on your computer. That's your way in.

Cursor doesn't give you a hosted preview link the way some other tools do. There's no website to paste. Cursor's AI wrote files in a folder on your machine, one you already own. Someone still has to actually run those files to see if they work. Having the folder is all the permission you need.

Last verified 18 August 2026. This page sticks to what Cursor's own docs say. We are not inventing a first run we have not actually watched.

the-app-cursor-just-wrote

Not like Lovable, Bolt, or v0

There's no preview link to argue about, because there isn't one.

Every other builder we cover makes you deal with some kind of confusing link: a preview that looks done but isn't, an error page, a link that asks you to sign in, a link that expires. Cursor skips all of that. What comes out is an ordinary folder of code on your computer. That's the whole product. If the AI said your app works, it said so in the exact same folder you're sitting in right now.

That makes this the simplest way to bring an app to PeerRun. You already have everything. You don't need to export your way out of a closed system. You also don't get a reassuring public link that looks like proof but isn't.

Cited as trivially true in TARGET_ACCESS §5.7 (Cursor row). Not separately footnoted; there is no hosted preview to fetch.

What Cursor already offers

Bugbot reviews your code. It doesn't run your app.

Code review

Bugbot

Reviews the code changes in a pull request, never a running app. It's genuinely useful for catching a bad change before it merges. But it has no way of knowing whether checkout still works once the app is running. We are not saying Bugbot is useless. We are saying it does not run your app.

Agent tool

The Browser tool

Lets Cursor's AI open and click around web pages. Cursor's own documentation doesn't describe this as a testing product. An AI clicking around inside your own editor is still the builder checking its own work, not a result you can rely on.

Cited: Bugbot, Browser tool. TARGET_ACCESS §3B.2.

How to bring a Cursor app

Add the folder on your computer, or connect the Git remote.

Sign in to PeerRun. Point it at your project folder, or at GitHub (or wherever else you already pushed it). That's the whole hand-off. We don't have a Cursor plugin or install anything into your editor. We also don't quote Cursor's plan limits or usage caps here, because we don't have reliable numbers for those and won't make them up.

If your app needs a local settings file you don't also give us, or expects a database that isn't running, PeerRun will say BLOCKED or UNKNOWN, and explain why. It won't call your app broken just because a key is missing.

the-app/
  package.json
  src/
  .env.example
  .env          yours, only if you offer it
  .git/

Honesty

What PeerRun will not be able to tell you

Signed-in product: Unknown until you grant access. A login screen running on your own computer is still a login screen.
A bot wall on wherever you deployed: If you later put the app behind something like Cloudflare and send us that link, a challenge screen means Blocked, not broken.
Secrets that never left your machine: We do not go digging through your local settings file. If you do not give us what the app needs to start, we name the gap, we don't guess around it.
What we do not claim: No Cursor plugin. No usage or plan-limit figures. No claim that Bugbot is worthless. No first-hand account of Cursor's signed-out first run, because we do not have one.

Questions people actually ask

Cursor, specifically

The AI said the app runs fine on my machine. Why doesn't it work for other people?

Your machine working is one situation, not proof it works everywhere. PeerRun's job is to run what's actually in your folder and tell you what it could see. Often that means only the parts anyone can reach without logging in, until you give us more access.

Can I just paste my localhost address?

An online checker cannot reach an address that only exists on your own computer. Bring the folder instead. Do not expect a pasted local address to count as the real thing.

Does PeerRun replace Bugbot?

No. Bugbot reviews code changes. We run the actual app. Different job. Keep both if both help you.

Sources

What this page is allowed to say

Cursor Bugbot
Cursor Browser tool
In-repo: TARGET_ACCESS §3B.2, §5.7. Quotas: COMPETITOR_FIRSTRUN_FLOWS §8 marked UNKNOWN, omitted here.

Cursor wrote the code. Someone still has to run the app.

We will not invent a pass. Anything behind your sign-in stays Unknown until you grant access. A robot-check is Blocked, not broken. The first look is on us.