Percy vs Applitools vs Chromatic — which visual regression tool wins in 2026

Percy vs Applitools vs Chromatic is the three-way decision most engineering teams hit the moment visual bugs start slipping past their unit and end-to-end suites. The short answer: pick Chromatic if your team lives in Storybook and ships a component library, pick Applitools if you are an enterprise that needs the most mature visual AI and design-to-code validation, and pick Percy if you want broad CI coverage with a genuinely usable free tier — especially if BrowserStack is already in your stack. The rest of this comparison shows the work behind that call: how each tool diffs, what they actually cost in 2026, and where each one breaks down.

TL;DR — the 2026 visual regression tool comparison

• Chromatic — Storybook-native, pixel diffs with Git-aware baselines, free 5,000 snapshots per month, paid from $149/mo. Best for component-driven frontend teams.
• Applitools Eyes — the most mature visual AI, plus a Figma plugin and a Storybook addon shipped in Eyes 10.22 (January 2026). Enterprise pricing, no public tiers. Best for design-system organisations and regulated industries.
• Percy (by BrowserStack) — broadest framework coverage, AI-driven Visual Review Agent shipped in late 2025, 5,000 screenshots free, paid through BrowserStack. Best for mid-market teams already on BrowserStack Automate.
• All three handle the diff. The differences are in noise reduction, ecosystem fit, and price-per-snapshot.
• None of them catch the visual bugs a human spots while clicking around — that is a separate layer.

---

What visual regression testing actually does

Visual regression testing automatically compares screenshots of a UI before and after a code change and flags differences for human review. It catches the class of bug that functional tests miss by definition: a checkout button that turned invisible on mobile, a font-weight cascade that broke a hero section, a CSS variable rename that shifted spacing across forty components.

The hard problem has never been capturing the screenshots. It is filtering the noise. Browser rendering engines paint the same CSS slightly differently across versions, anti-aliasing nudges individual pixels, and fonts hint at sub-pixel variations between operating systems. A naive pixel-by-pixel diff on a real production UI surfaces hundreds of false positives every run — not because something broke, but because Chrome 130 drew a glyph one pixel different from Chrome 129. How each tool handles that noise, and how it stitches into the rest of your workflow, is the comparison that matters.

---

Percy: CI-first visual testing with a new AI review layer

Percy, acquired by BrowserStack in 2020, was built on a simple premise — visual testing should slot into your existing CI pipeline, take minutes to set up, and give your team a clean PR-level review workflow. That premise still holds in 2026, but the product underneath has changed meaningfully.

How Percy works

Percy injects itself into your test runs via the @percy/cli package and a framework-specific SDK. As your functional tests execute, Percy captures DOM snapshots — not raw screenshots — and uploads them to its cloud, where they are rendered consistently across browsers and viewports. That render-on-upload approach removes a class of CI-environment flakiness: your local Chrome version no longer determines whether a build passes. Snapshots get compared against a baseline tied to your Git branch, and when a pull request opens, Percy posts a check with the diff.

Framework coverage

Percy's framework support is the broadest in the comparison. Officially supported SDKs include Selenium, Playwright, Cypress, Puppeteer, WebdriverIO, TestCafe, Nightwatch, Storybook, Ember, Gatsby, Jekyll, Appium, Tricentis Tosca, and Maestro, plus a build-your-own option. Teams typically choose between the standalone Percy SDK and the BrowserStack SDK, which fuses functional and visual testing into a single pipeline on BrowserStack Automate.

The Visual Review Agent

The big 2026 story for Percy is the Visual Review Agent, shipped in late 2025. The old Percy workflow leaned on pixel-level highlighting — every changed pixel got a red overlay, which was technically accurate and humanly unusable. A font rendering shift could light up an entire page in red.

The Visual Review Agent replaces that with AI-driven smart highlights: bounding boxes drawn around meaningful changes, with everything else suppressed. The reviewer reads "the header navigation shifted 4px left and the hero image was replaced" rather than "47,000 pixels changed." BrowserStack reports the agent cuts review time by roughly 3x and filters out about 40% of non-meaningful changes from typical builds. It is opt-in at both build and snapshot level, supports snapshots up to 13,500 pixels tall, and is paid-plan only — plans up to one million snapshots get unlimited AI review, with throttling above that.

Percy pricing in 2026

Percy's free tier is 5,000 screenshots per month with unlimited team members — one of the most usable free tiers in this category. Paid plans scale with snapshot volume and run through BrowserStack's sales motion. BrowserStack does not publish a clean pricing grid for Percy beyond the free tier, so teams expecting to exceed 5,000 snapshots should expect a quote conversation. Parallelisation — running multiple snapshot jobs simultaneously to keep CI fast — is metered on most plans, which matters once your suite passes a few hundred snapshots.

Where Percy fits

Percy works well for mid-market teams already on BrowserStack that want visual coverage stitched into the same dashboard, or any team that values framework breadth across Cypress, Playwright, Selenium, and Storybook. The Visual Review Agent is the reason to reconsider Percy in 2026 if pixel-diff noise was why you abandoned it in 2023.

---

Applitools: enterprise visual AI with design-to-code coverage

Applitools Eyes is the most technically sophisticated tool in this comparison and the one with the longest head start in visual AI. Where Percy and Chromatic started as screenshot-comparison products and added AI later (or, in Chromatic's case, not at all), Applitools built AI-first comparison as its core differentiator from launch.

How Applitools works

Applitools' engine — branded Visual AI — is a perceptual comparison model trained on hundreds of millions of UI screenshots. It does not compare pixels; it compares what those pixels mean. The model understands that a button is a button, that a navigation bar holds a structural relationship to the page around it, and that a 2-pixel shift may or may not be a regression depending on context. Teams tune sensitivity through four match levels — Strict, Layout, Content, and Exact. Framework support is wide and includes mobile (Selenium, Playwright, Cypress, WebdriverIO, Appium, plus SDKs for the major language stacks), and snapshots run through the Ultrafast Test Cloud — Applitools' parallelised render grid that cuts wall-clock test time significantly on suites that would otherwise serialise across browsers and viewports.

Eyes 10.22 — Figma plugin and Storybook addon

In January 2026, Applitools shipped Eyes 10.22, which closed two long-standing gaps.

The Figma plugin exports frames — components, full pages, or entire prototypes — straight into Eyes as visual baselines. Developers then run their normal visual tests against those design baselines, and Visual AI flags every place the built implementation drifts from the approved design. The plugin also supports design-to-design comparison, so designers can diff two Figma versions to track iteration history.

The Storybook addon brings component-level testing into the Storybook UI itself — no separate test file, no CI round-trip. Tests run inside Storybook, results group by component, and a reporter widget surfaces what needs review first. The addon supports React, Vue, and Angular Storybooks, and gives Applitools a credible answer to teams who would otherwise have defaulted to Chromatic for component workflows.

Applitools pricing in 2026

Applitools is the most expensive option here and the one that requires the most sales conversation. There is no public pricing grid. The free tier is a trial — useful for evaluation, not for production work. Paid plans are quote-based and scale on usage measured in pages, components, or test executions, with dedicated and on-prem deployment options for regulated industries. For small teams running a few hundred snapshots a week, Applitools is more tool than you need. For large organisations running tens of thousands of snapshots across complex multi-platform applications, the visual AI typically cuts review time enough to justify the spend.

Where Applitools fits

Applitools is the right answer for enterprise teams with significant design systems, multi-platform coverage requirements, and budget for a tool that pays back through reviewer time saved. The Figma plugin is the genuine 2026 differentiator — no equivalent exists in Percy or Chromatic. For regulated industries needing on-prem or dedicated deployments, Applitools is effectively the only credible option of the three.

---

Chromatic: visual testing built around Storybook

Chromatic is built by the team that maintains Storybook, and that lineage shapes every product decision. If Percy is CI-first and Applitools is AI-first, Chromatic is Storybook-first — designed for teams that build UI as isolated components, develop inside Storybook, and want visual regression testing native to that workflow.

How Chromatic works

Chromatic captures snapshots of your Storybook stories — one per story, per browser, per viewport — and compares them against a baseline. Changes surface in a review UI; approvals are wired into Git so the next merge inherits a clean baseline. Beyond Storybook, Chromatic also supports Playwright and Cypress for full-page and interaction testing, narrowing the "Storybook only" framing the tool used to live with.

Diffing is pixel-based — no AI engine equivalent to Percy's Visual Review Agent or Applitools' Visual AI. The trade Chromatic makes instead is Git-aware baseline management: baselines persist through branches and merges without the conflicts that show up in tools with simpler branch comparison logic. For teams running multiple feature branches in parallel, that baseline discipline is a real practical advantage.

Speed, parallelisation, and TurboSnap

Chromatic runs snapshots with unlimited parallelisation on all plans — every story captures simultaneously, and the tool publishes benchmarks showing it can snapshot and diff thousands of tests in under two minutes. That contrasts with Percy, which meters parallelism on most plans, and Applitools, whose parallelism is bundled into Ultrafast Grid pricing. TurboSnap is the related cost lever: it uses Webpack/Vite dependency analysis to skip snapshots of components unaffected by a given commit, dropping snapshot usage on a typical three-component PR in a 200-component library by an order of magnitude.

Chromatic pricing in 2026

• Free — 5,000 snapshots per month, unlimited contributors. Suitable for small teams and open-source projects.
• Pro — from $149/month, with higher snapshot allowances, approval workflows, Figma integration, and unlimited team members. Overage is metered per snapshot.
• Enterprise — custom pricing with SSO/SAML, audit logs, and on-prem options.

Meaningfully cheaper than Applitools at the same team size, and competitive with Percy once you account for parallelisation costs.

Where Chromatic fits

Chromatic is the right pick when your team is Storybook-first and maintains a component library as a deliverable in its own right. It is the wrong pick if you do not use Storybook, if your visual testing needs are dominated by full-page user-flow regressions, or if you specifically need AI-powered noise reduction on high-volume snapshot runs. The newer Playwright and Cypress integrations help, but the tool's real strength is still upstream of those, inside Storybook.

---

Head-to-head — the dimensions that actually decide the call

Diffing technology

Applitools has the most mature visual AI of the three — the longest training runway and the most semantic understanding. Percy's Visual Review Agent narrows the gap meaningfully in 2026, though the model is newer. Chromatic ships no AI layer at all. Its pixel-based diff is accurate but generates more false positives on rendering-heavy UIs, which is part of why noise-pressured Chromatic teams often pair it with Percy rather than replacing it.

Framework and platform coverage

If your stack is plural — a mix of Cypress for end-to-end, Storybook for components, and Appium for mobile — Percy and Applitools both span it. Chromatic plus a second tool is the more common pattern in that scenario.

CI/CD integration

All three integrate with GitHub Actions, GitLab CI, Jenkins, CircleCI, and Bitbucket Pipelines, and all three post review status as a PR/MR check. The differences sit in baseline management — Chromatic's Git-aware baselines handle branch-heavy workflows cleanest, while Percy and Applitools both work fine on trunk-based teams but get noisier in long-lived feature branch setups.

Pricing summary

Chromatic is the only one with fully public pricing. Percy publishes the free-tier number and routes everything else through BrowserStack sales. Applitools publishes nothing and expects a quote conversation.

---

How to choose — three decision paths

The right answer is not the best tool. It is the best tool for the shape of your team.

Storybook-heavy frontend team → Chromatic

If your team builds and maintains a component library, develops inside Storybook, and treats components as a deliverable in their own right, Chromatic is almost always the right pick. Git-aware baselines, TurboSnap, unlimited parallelisation, and a Storybook-native review UI all compound. See the best visual regression testing tools roundup for how Chromatic sits against open-source options like BackstopJS.

Enterprise with a design system → Applitools

If you are a large organisation with a real design system, multi-platform requirements (web plus iOS plus Android), and procurement that expects on-prem and SSO/SAML, Applitools is the credible answer. The Figma plugin shipped in Eyes 10.22 is the differentiator that justifies the price — design-to-code validation against an approved Figma frame is something the other two tools cannot do. Plan for a quote conversation and a real onboarding rollout rather than self-serve signup.

Mid-market team already on BrowserStack → Percy

If BrowserStack Automate already runs your cross-browser functional tests, the path of least resistance is Percy. The BrowserStack SDK fuses functional and visual testing into one pipeline, billing rolls into your existing contract, and the Visual Review Agent makes the reviewer experience meaningfully better than the 2023 version some teams remember.

Some teams legitimately run two: Percy plus Chromatic is a common pattern — Percy for full-page CI coverage, Chromatic for daily component review. Applitools typically replaces both, at a price that only makes sense at scale.

---

Where automated visual regression testing stops working

All three tools are automation-first. They catch regressions introduced by code changes — the visual diff between commit N and commit N+1. They are not designed for the exploratory phase of QA: actually using the product, finding edge cases nobody wrote a test for, and reporting visual bugs that fall outside snapshot coverage.

Automated visual regression covers the paths you thought to test. Manual exploratory QA covers the paths you did not. A pixel shift in your header navigation gets caught by Percy on the next PR. A visual bug in a modal that only appears after a specific sequence of user actions, on a viewport you did not write a test for, with a feature flag enabled — that one shows up when a human is clicking around, not in any pipeline.

This is where Crosscheck sits. Crosscheck is a free Chrome extension for manual visual bug reporting — annotated screenshots, screen recordings, console logs, network requests, and Instant Replay (a retroactive DOM replay of the last few minutes of the session) filed straight to Jira, Linear, ClickUp, GitHub, or Slack. Pair it with one of the three regression tools above, and you cover both layers of the visual quality story — the regressions caught in the pipeline, and the bugs caught when a human is actually using the product. The perfect bug report template is the standard way to ensure those manual reports include everything a developer needs to reproduce them.

---

FAQ

What is the main difference between Percy, Applitools, and Chromatic?

Percy is a CI-first visual testing platform with broad framework coverage and a new AI review layer. Applitools is an enterprise-grade visual AI engine with design-to-code validation via its Figma plugin. Chromatic is a Storybook-native visual testing tool with Git-aware baselines and unlimited parallelisation. The difference is not whether they catch visual bugs — all three do — but how they handle noise, where they sit in your workflow, and what they cost.

Which visual regression tool has the best free tier in 2026?

Percy and Chromatic both offer 5,000 screenshots or snapshots per month for free, with unlimited team members on Chromatic and unlimited contributors on Percy. Applitools' free tier is a trial-grade allowance and not suitable for ongoing production use. For most small teams or open-source projects, Chromatic and Percy are the two realistic free-tier options.

Is Chromatic worth it if I do not use Storybook?

Chromatic's Playwright and Cypress integrations are real, but the tool's strongest features — Git-aware baselines, TurboSnap, the review UI — are all tuned for component-driven Storybook workflows. If you are not on Storybook, Percy or Applitools will be a better fit. If you are considering adopting Storybook anyway, Chromatic becomes a credible reason to do it.

Does Applitools really justify its price?

For enterprise teams running tens of thousands of snapshots across complex multi-platform applications, yes — the visual AI typically cuts reviewer time enough to recoup the cost. For small or mid-sized teams running a few hundred snapshots per week, no — Percy or Chromatic will deliver most of the value at a fraction of the price. The Figma plugin is the case where Applitools' price becomes uniquely defensible: design-to-code validation is something neither Percy nor Chromatic can replicate today.

How do visual regression tools handle dynamic content like ads or carousels?

All three offer ways to ignore dynamic regions. Percy has Intelli-ignore, Applitools has match levels and explicit ignore regions, and Chromatic supports CSS-selector-based ignores. Percy's Visual Review Agent and Applitools' Visual AI also handle some dynamic content automatically without explicit configuration, which is part of why noise reduction has become the most important capability differentiator in 2026. For framework-level choices around Cypress, Playwright, and Selenium, the Selenium vs Playwright vs Cypress comparison covers the trade-offs in depth.

---

Start catching the visual bugs your test suite cannot

Visual regression testing is one layer of visual quality assurance. It catches the bugs that ship inside a code change. It does not catch the bugs that show up the first time a real person clicks through a flow you did not think to write a test for — and those are the bugs that reach production.

That is the gap Crosscheck fills. Manual bug reporting with full developer context — annotated screenshots, screen recordings, console logs, network requests, Instant Replay — sent straight to Jira, Linear, ClickUp, GitHub, or Slack from the Chrome extension. Free, no usage limits, no paid tier.

Try Crosscheck free