Percy vs Applitools vs Chromatic

These three end up on almost every shortlist, and they are not really competing for the same job. One tests components, one tests whatever your suite drives, and one is an enterprise comparison engine you attach to any of it. Getting that straight settles most of the decision before pricing enters the conversation.

The one-paragraph answer

Chromatic if your design system lives in Storybook — best baseline engineering in the category, published pricing from $179/month. Applitools if you are an enterprise with an existing test suite, need cross-browser rendering at scale, and false positives are what is blocking adoption — $667/month billed annually. Percy if you are already inside BrowserStack and want route-level coverage from your existing suite, accepting that paid pricing is now sales-gated.

At a glance

1. Capture model

The most consequential technical axis, and the one least often explained.

• Percy serialises the DOM, CSS and assets and re-renders them in its own cloud. One capture yields several viewport widths without re-running your suite — genuinely cheap responsive coverage. The trade: Percy's renderer decides the pixels, so what you approve is not literally what your test saw.
• Applitools captures through its SDK inside your test and compares in the cloud. Its Ultrafast Grid then renders that single capture across Chrome, Safari, Firefox and Edge without you running a browser matrix — the strongest answer to cross-browser cost in the category.
• Chromatic builds your Storybook and renders it in a standardised cloud browser that watches for loading, painting and reflow to choose the capture moment. It also archives the DOM and styling at capture time, so you can inspect a change in your own browser instead of re-running to reproduce.

None of the three captures in the browser your tests actually ran in. If that matters, local-capture tools are a different category — see Chromatic vs Argos.

2. Comparison engine

This is where Applitools has a real, non-marketing advantage. Pixel comparison with a tolerance is a blunt instrument: set it tight and antialiasing fails your build, set it loose and a genuinely broken layout slips through. Perceptual comparison sidesteps that trade-off by modelling what a human would notice.

Applitools' own comparison material frames it exactly this way — match levels configured in the SDK rather than thresholds hand-tuned per project — and on the underlying technique the claim holds up. They also group similar differences across browsers and pages into a single review, so one approval updates every affected baseline, which matters once a matrix is wide.

3. Baselines and branching

Chromatic wins this on engineering depth, with caveats worth knowing.

Baselines live alongside git history and persist through branch and merge. Each branch keeps its own; a new branch inherits from its branch point; squash and rebase merges are detected through git provider APIs because those operations destroy the commit the last build was tied to.

The caveat, from Chromatic's own documentation: approving a change on main does not update other branches' baselines, so you can see "changes that look new but were already approved and merged elsewhere". The preferMergedBaselines flag fixes it — but it is account-level with no per-project override.

Percy compares a feature branch against main, which is adequate for linear workflows and strains on telescoping pull requests. Applitools sits between the two, with baselines scoped per branch and batch.

4. What it actually costs

Two of the three publish enough to model. The third does not, which is itself a planning fact.

Chromatic's TurboSnap multiplier is the other number to model honestly. It bypasses builds only when no story's dependency graph changed, and their docs note that editing any file in that graph — a comment counts — prevents the bypass. Global Storybook file changes force a full rebuild, and rebase or squash breaks the computation. Detailed in Percy vs Chromatic.

5. The axis none of them compete on

All three require you to author the coverage. Percy needs a snapshot call in a test you wrote. Applitools needs an SDK call per state. Chromatic needs a story someone maintained. In all three, a page nobody wrote a test or story for is invisible.

This is the single best predictor of whether a visual testing programme still exists in a year, and no comparison page in this category — including the vendors' own — treats it as a dimension. Coverage decays at exactly the rate team discipline does: routes get added and never captured, a noisy test gets skipped, and eventually the suite is green and meaningless.

It is also, transparently, the problem we built VisualMax to remove, so weigh the framing accordingly. The point stands regardless of vendor: when you evaluate these three, add a brand-new route to your application and run the tool without touching its configuration. All three will miss it. If that is your real constraint, a whole category of tools now generates its own coverage — Meticulous, TestDriver and VisualMax get there by different routes.

Choosing

And Percy if you are already committed to BrowserStack, your coverage is route-level from an existing suite, and you are comfortable negotiating a price that is no longer published. It remains a capable product — the commercial signals are the thing to weigh, not the technology. More in Percy alternatives.