For visual QA testing, the strongest measured-use options are Playwright, which provides a test runner for end-to-end and visual checks 5, and Applitools, an AI-powered testing platform with Visual AI that catches visual bugs 7. Open-source and agent-oriented alternatives include TestivAI, a local MIT-licensed visual regression tool with an MCP server 1, and UI Verify, visual regression for Playwright, Storybook and Vitest with agent integration 6.
- 1
Playwrightplaywright.devwebsite read, tools not yet checked
Playwright Test is a full-featured test runner with auto-waiting, assertions, tracing, and parallelism across Chromium, Firefox, and WebKit, used as the base for end-to-end checks 5.
official:SDK · playwright+230 community tools built for it
- 2
Applitoolsapplitools.comwebsite read, tools not yet checked
Applitools combines its Visual AI with GenAI and no-code approaches to catch visual and functional bugs on every page and reduce false positives 7.
- 3
TestivAItestiv.aiwebsite read, tools not yet checked
TestivAI is open-source visual regression testing that compares screenshots with the DOM and computed styles, reports diffs with the element they belong to, and runs locally under MIT 1.
official:MCP server · TestivAI+1 community tools built for it
- 4
Sources
- 1TestivAI — Open-source visual regression testingtestiv.ai · read 2026-10-29
- 5Fast and reliable end-to-end testing for modern web apps | Playwrightplaywright.dev · read 2026-10-29
- 6UI Verify · Visual testing for coding agentsuiverify.ai · read 2026-10-29
- 7Applitools | AI-Augmented Testingapplitools.com · read 2026-10-29
- 16setup-visual-testing: declares "setup-visual-testing" (the record we read and signed)unfragile.ai · read 2026-10-24
For agents: the 10 actions behind this answer
· TestivAI website read, tools not yet checked
Open-source visual regression testing. Screenshots compared with the DOM and computed styles, an HTML report, approvals from the pull request. MIT, runs locally.
· Snapvisor Docs website read, tools not yet checked
Snapvisor is a visual testing platform built on an open-source engine. Capture screenshots from your existing test suite, compare them with a baseline, and review changes on your pull requests.
· SnapDiff website read, tools not yet checked
Visual regression testing for AI coding agents and preview URLs. HTTP API + MCP server so Claude Code, Cursor, and Windsurf can verify UI changes themselves. From $19/mo.
· Pixel Not Perfect website read, tools not yet checked
Check your coding agent's work against the design, down to the pixel. In-browser visual review and QA via MCP.
· Playwright website read, tools not yet checked
Web automation and testing for apps, scripts, and AI agents
setup-visual-testing · setup-visual-testing verified and signed
Set up UI Verify visual regression testing in a repo end to end — detect or scaffold the capture path (Storybook stories, a Playwright suite, or Vitest browser-mode component tests), install the uiverify CLI and wire it into CI, and author the first stories the economical, deterministic way from the
· Applitools website read, tools not yet checked
Applitools is an AI-powered end-to-end testing platform that combines our proven Visual AI with the latest GenAI and no-code approaches to maximize your test coverage while automating maintenance and reducing false positives.
· Screenshot API website read, tools not yet checked
Capture full-page screenshots of any URL via a fast API. Bulk capture, visual diff, Lighthouse audits and an MCP server for AI agents. Free tier, 100/mo.
compare_visual · OverlayQA MCP found in code, not yet checked
Coming soon — not yet implemented; the server returns a not-implemented notice (use the OverlayQA browser extension for visual comparison today). Will compare a live URL against a Figma frame and return visual differences and a match score.
· i2dev One account, every developer tool website read, tools not yet checked
Snap Visual Regression: hosted visual regression testing for GitHub Actions, with shared baselines and visual review. Part of i2dev.
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.