v1.2.0-rc

Visual Regression Testing
that actually understands your UI.

Stop fighting 1-pixel false positives. BugHunters Vision uses Multimodal AI to evaluate Playwright tests like a human QA. Zero cloud storage. 100% InfoSec friendly.

Request Access → Try the demo →

Not ready to run it? See a real report from the live demo ↗

Trusted by QA teams who hate flaky tests. Built by BugHunters.

75%

of visual test failures in our E-Commerce suite
were false positives

0

false positives
with BugHunters Vision AI

€0

AI cost on identical screenshots
Fast Pixel Match runs locally

See the full side-by-side comparison →

The problem

Same change. Completely different verdict.

A banner shifted 2 pixels. Pixel diffing panics. Our AI understands context.

Traditional Pixel Diffing

myapp.com/checkout

✗  BUILD FAILED

ScreenshotDiff: threshold exceeded (2.3%)

banner shifted 2px. Again.

BugHunters Vision AI

myapp.com/checkout

✓  PASS

"The banner moved slightly, but all buttons and text remain fully visible and functional."

— BugHunters Vision AI Evaluator

Why teams choose us

Built for QA teams who care about reliability.

No more babysitting visual tests. No more security trade-offs.

🛡️

InfoSec Approved. Zero Cloud Storage.

We don't want your data. Baseline images stay in your repository. Reports are generated locally. Our API is a stateless evaluator — images are destroyed immediately after evaluation.

No cloud upload Stateless API Zero retention

🧠

Human-Like Evaluation via Prompts.

Dynamic dates? A/B tested colors? Tell our AI what to ignore — in plain English. Stop maintaining fragile ignore-regions that break with every redesign.

          prompt: 'Ignore the dynamic date.
        Ensure Pay button is visible.'
        

💰

Smart Cost Control. Pay only when it matters.

Don't pay AI for identical screenshots. Our built-in local pixelmatch intercepts 100% matches for free. You only pay for an API call when the UI actually changes.

FREE identical

API on change

⚡

Native Playwright DX. Zero learning curve.

Plugs into your playwright.config.ts. VS Code Playwright extension & UI mode work out of the box. Results appear as annotations & attachments directly in the native Playwright HTML report too.

VS Code compatible playwright.config.ts TypeScript native

🔗

Multi-step & Soft Assertions. Built for long E2E flows.

Call vision.check() multiple times per test — each step gets its own baseline and report entry, auto-grouped under the parent test. Enable bvFailMode: 'soft' to let a test run to the end even when a visual step fails.

Multi-step grouping bvFailMode: soft Multi-project

🚀

Works in any CI/CD pipeline. Rich Markdown summaries.

One command publishes a full test summary — passed checks, AI failure reasons, method badges — directly into GitHub Step Summary, GitLab, Bitbucket, CircleCI, or Jenkins. Pipeline turns red exactly when BHV detects a real regression.

GitHub Actions GitLab CI Correct exit codes See CI/CD integration →

How we compare

The only tool that gets all three right.

Open-source misses dynamic data. Enterprise tools demand your UI screenshots. We solve both.

Feature	Playwright open-source	Applitools / Percy visual AI platforms	BugHunters Vision you are here
Dynamic data tolerance Live metrics, timestamps, chatbot widgets, maintenance banners	✗ Fails on data change	✓ Visual AI	✓ Claude Vision AI
Data Security & InfoSec Where do your UI screenshots go?	✓ Everything stays local	✗ Your UI goes to their cloud	✓ Zero cloud storage, stateless API
Pricing model What does a visual check actually cost?	✓ Free — but no AI	~ Enterprise pricing, every check billed	✓ Credits deducted only when pixels differ

Banking, health, fintech teams: your UI screenshots never leave your infrastructure. See the full demo →

Integrations & Ecosystem

Works where your tests live.

Native Playwright support today. Cypress and Robot Framework arriving next.

Playwright

Live v1.2.0

Drop-in integration — vision.check(), locator support, soft/hard fail mode, CI/CD Markdown summaries and an interactive HTML report out of the box.

          npm install @bughunters/vision
        

Try it in 5 min →

Cypress

Coming Q3 2026

Native cy.task() plugin — pixelmatch locally, AI call only on real diff. Zero config changes to your existing Cypress setup.

Get early access + launch discount:

Robot Framework

Coming Q4 2026

Python library compatible with SeleniumLibrary and Browser. Pillow/OpenCV Fast Pixel Match. Built for enterprise & Paranoia Tier teams.

Get early access + launch discount:

Developer experience

One reporter. One call. That's it.

Register the reporter in playwright.config.ts, replace screenshot assertions with vision.check(). Your existing test structure stays exactly the same.

playwright.config.ts

export default defineConfig({
  reporter: [[
    '@bughunters/vision/reporter',
    { snapshotsDir: './bhv-snapshots' }
  ]],
  use: {
    screenshot: 'off',   // BHV handles screenshots
    bvMode: 'ai',       // 'ai' | 'strict' | 'off'
    bvFailMode: 'hard', // 'hard' | 'soft'
  } as Record<string, unknown>,
});

checkout.spec.ts

import { test } from '@playwright/test';
import { vision } from '@bughunters/vision';

test('Checkout page', async ({ page }) => {
  await page.goto('/checkout');

  // Replaces the native toHaveScreenshot()
  await vision.check(page, 'checkout',
    'Ignore the dynamic delivery date.
      Ensure the Pay button is enabled.'
  );
});

Request Access →

CI/CD Integration

Works in any pipeline. Reports everywhere.

BugHunters Vision runs natively in GitHub Actions, GitLab CI, Bitbucket Pipelines, CircleCI, and Jenkins — anywhere Playwright runs. One command publishes a rich Markdown summary directly into the workflow summary.

BugHunters Vision GitHub Step Summary showing passed and failed visual checks

GitHub Actions Step Summary — live run against bughunters.dev

✓

One-line GitHub Step Summary
cat bhv-report/summary.md >> $GITHUB_STEP_SUMMARY
✓

Pipeline turns red on real regressions only

Correct exit codes — false positives never block your deploy.
✓

Baselines in your repo — no external state

bhv-snapshots/ committed with your code. Every runner gets the same reference via git checkout.

GitHub Actions GitLab CI Bitbucket CircleCI Jenkins

Pricing

Simple, transparent pricing.

Scale from Pay-As-You-Go to Enterprise. Zero vendor lock-in, no hidden fees.

Pay-As-You-Go

€29

one-time purchase

1,000 AI Visual Checks
API token delivered instantly
Zero cloud storage
Credits never expire

Contact us

Self-serve purchases are temporarily paused.

Stop chasing pixels.
Ship with confidence.

Self-serve trials and purchases are temporarily paused. Existing API tokens keep working. Interested in access, an on-premise setup, or a pilot? Drop us a line — we read everything.

Contact us →

Try it in 5 minutes →

Open Source

Try the demo before you commit.

Our demo kit is public on GitHub. Clone it, add your token, run npm run setup and see all 9 scenarios — AI passes vs. native Playwright false positives — on your own machine.

bughunters-dev/promo-demo

9-scenario demo · clone → install → run

bughunters-dev/bughunters-playwright-example

@bughunters/vision package source