🐾 Now in public beta

Find the bugs before your users do.

Bugpup's AI agent explores your Next.js app in a real browser β€” clicking buttons, filling forms, and hunting for what's broken. Every find comes with a screenshot, repro steps, and a severity. No test scripts. No config.

No credit card GitHub connect in one click First report in minutes
median run
24s
median run
pages explored per run
14
pages explored per run
avg cost per run
$0.02
avg cost per run
https://fetch-demo.vercel.app/adopt/applyexploring
πŸ• FetchNext.js 15
Next.js demo app
Adopt a best friend

Every pup deserves a home. Browse adoptable dogs and start an application in seconds.

Find your pup
🐢
Biscuit
2 yr Β· lab mix
πŸ•
Mochi
2 yr Β· lab mix
🐩
Waffles
2 yr Β· lab mix
πŸ• Bugpup agentlive run

πŸ• Bugpup is starting…

What it catches

The pup sniffs out six kinds of trouble

Not scripted assertions β€” real bugs, found the way a user would hit them.

JavaScript errors

Uncaught exceptions, console errors, and render failures on every page the pup visits.

Cannot read properties of undefined

Broken user flows

Checkout, signup, search β€” journeys that die mid-way get flagged with the exact step.

Payment page 500s after login

Network failures

Failed requests, 4xx/5xx responses, and API calls that hang or time out.

GET /api/checkout β†’ 503

Accessibility violations

WCAG 2.1 AA checks that catch users being locked out of your app.

Button missing accessible name

Visual regressions

Layout shifts and styling breaks caught between runs, before they ship.

Navbar overlaps hero on mobile

Performance regressions

Routes that get slower between releases, with LCP and CLS tracked per run.

LCP 4.2s on /products

Every bug is verified by the agent in a real browser β€” no flaky, scripted noise.

A catch up close

Every bug comes with proof

No vague "something broke". Each report tells you exactly what broke, where, how serious it is, and how to reproduce it.

  • Screenshot & trace

    The exact page state when the bug happened β€” not a log line.

  • Step-by-step repro

    The agent's own path, replayable in one click.

  • Severity & type

    Critical, high, medium, or low β€” so your team knows what to fix first.

  • PR & Slack integration

    Findings land on the pull request before anything merges.

HIGHbroken_flow
open26s ago

Application accepts invalid email

/adopt/apply Β· run 8f3a21c2 Β· project fetch-demo

fetch-demo.vercel.app/adopt/apply
notanemail
bug #1
  1. 1Open /adopt/apply
  2. 2Enter notanemail in the email field
  3. 3Click Send application β€” no validation error, submission proceeds
PR #482 Β· acme/fetch-demoπŸ”” Slack alert sent$0.004 of compute
How it works

Three steps to a sniffed-out app

No test scripts. No config files. Just connect and go.

Step 1

Connect your repo

GitHub OAuth in one click. Bugpup reads your Next.js source and maps every route, component, and auth flow β€” no config files, no YAML.

Step 2

The pup explores

A real browser, driven by AI. It clicks, types, and navigates like a user would β€” including the paths you'd never think to test.

Step 3

Get the catch report

Bugs land in your dashboard with screenshots, repro steps, and severity β€” plus PR comments and Slack alerts before anything merges.

Why Bugpup

Better than writing tests

Traditional tools make you write scripts. Bugpup just starts hunting.

CapabilityBugpupPlaywrightCypress
Zero config
AI-powered exploration
No test scripts needed
Code-coverage aware
PR comments
Accessibility checks
Self-healing selectors
Pricing

Simple, transparent pricing

Start free. Upgrade when you need more.

Free
For side projects and solo devs.
$0/mo
  • 1 project
  • 25 runs/month
  • Basic bug reports
  • Community Discord
Get started
Most popular
Pro
For teams shipping regularly.
$49/mo
  • 5 projects
  • 500 runs/month
  • Slack alerts
  • PR comments
  • Priority support
  • Code coverage reports
Start free trial
Team
For agencies and large teams.
$199/mo
  • Unlimited projects
  • 2,000 runs/month
  • CI/CD integration
  • Team dashboard
  • API access
  • Custom run schedules
Contact sales
FAQ

Got questions?

Do I need to write any test code?

No. Bugpup's AI agent explores your app autonomously β€” clicking, typing, navigating, and finding bugs. You just connect your repo.

Does it work with server components and App Router?

Yes. Bugpup is built specifically for Next.js 13+ with App Router, server components, layouts, and all the latest features.

How does authentication work?

Configure test credentials for form-based auth, or pass a session cookie/token for OAuth-based apps. The pup logs in just like a real user.

What kinds of bugs can it find?

JavaScript errors, broken network requests (4xx/5xx), accessibility violations (WCAG 2.1 AA), broken user flows, slow page loads, and visual regressions between runs.

How much does a run cost?

A typical run exploring 10-15 pages costs about $0.02-0.05 in AI API fees. We handle model routing to keep costs low while maintaining quality.

Can I run it in CI/CD?

Yes. Bugpup integrates with GitHub Actions and any CI pipeline β€” trigger runs on push, on pull requests, or on a schedule.

Ready to let the pup loose?

Connect your Next.js repo and get your first bug report in minutes. Free while you're in beta.