The cloud platform that tests every pixel of your interface - built by the people who gave the web Storybook.
It is a Tuesday afternoon and somewhere a developer pushes a one-line CSS change. Nothing dramatic. A little padding. Before the branch even finishes building, Chromatic has rendered the affected components in Chrome, Safari, Firefox and Edge, compared each one against yesterday's version pixel by pixel, and flagged the four screens where that "harmless" padding quietly shoved a button off the page. No human took a screenshot. No QA engineer clicked through anything. The bug was caught before lunch.
That is Chromatic in 2026: a cloud service that does the part of frontend work everyone agrees is important and nobody wants to do by hand. It runs visual, interaction, and accessibility tests across real browsers, then hands the results to a review workflow where designers, product managers, and engineers sign off on what the interface actually looks like. The company is small - around 68 people - but its reach is not. The same team maintains Storybook, the open-source tool used by Airbnb, Microsoft, and LEGO to build UI components in isolation.
Which sounds simple, in the way that "just make the internet look right" sounds simple. It is not.
Backend code has a comforting property: it is either right or wrong, and a test can tell you which. Frontend code does not work that way. A component can pass every unit test and still render as a smear of overlapping text on Safari. The truth of a user interface lives in how it looks and behaves - and for years the only tool for verifying that was a person, a browser, and a lot of patience.
So teams did what humans do with tedious work: they skipped it. They eyeballed the big screens, shipped, and discovered the broken dropdown when a customer emailed about it. Manual QA caught some of it. Nobody enjoyed manual QA. The bugs that slipped through were almost always visual - the category of defect that is most obvious to a user and most invisible to a test suite.
The founders had a front-row seat to this particular flavor of misery. They had, after all, built a tool that thousands of teams used to make components in the first place.
Zoltan Olah, Tom Coleman, and Dominic Nguyen did not start with Chromatic. They started with Percolate Studio, a web app consultancy founded in 2012. In 2015 the open-source framework Meteor acquired the studio, and the three got a close look at what makes developer tools spread - and what makes them stall.
Out of that came Storybook, an open-source workshop for building UI components on their own, away from the rest of the app. It caught on. The obvious question for any team that maintains a beloved free tool is the uncomfortable one: how do you eat?
Their answer, in 2017, was Chromatic. Keep Storybook free and excellent. Build a paid cloud service on top of it that solves the next problem in the same workflow - testing and reviewing all those components at scale. Open source for trust and distribution; a SaaS product for revenue. The bet was that the people building components would happily pay to stop manually checking them.
Give away the tool everyone builds with. Sell the service that tests what they built. Storybook drives adoption; Chromatic pays the maintainers.
EST. 2017 · CHROMA SOFTWARE, INC.
Three former consultants who decided the most valuable thing they could sell was looking at pictures of websites very carefully.
To a developer, Chromatic is almost boringly simple: it runs in the CI pipeline and comments on the pull request. Underneath, it is rendering interfaces in real cloud browsers, capturing snapshots, diffing them against a baseline, and routing anything that changed to the right reviewer. Here is what it actually does.
Catches bugs in layout, fonts, color, and spacing across Chrome, Safari, Firefox, and Edge - down to the pixel.
Verifies clicks, typing, and drags behave the way they should, using your existing Playwright and Cypress suites.
Flags WCAG violations in the same run, so accessibility stops being a separate quarterly fire drill.
Tests only the components a change actually touched instead of the whole app. Less time, less cost.
Visual diffs plus stakeholder approvals - designers, PMs, and engineers sign off in one place.
A hosted, searchable component library with docs and a Figma plugin for design handoff.
// A 13-year arc through frontend tooling
You do not have to take the pitch on faith. The teams that test their interfaces with Chromatic include GitHub, Adobe, the BBC, Square, Monday.com, Vercel, Retool, DocuSign, The Guardian, Indeed, Nespresso, Toyota, Perplexity, Netlify, and Mozilla. These are organizations with the budget to build visual testing in-house. They chose not to.
A partial guest list. Several of these companies could build their own visual testing tool. Time, it turns out, is also a budget line.
// Illustrative comparison of frontend QA approaches - directional, not audited benchmarks
Relative testing speed gains reflect Chromatic's stated ~4x year-over-year improvement from Capture, TurboSnap, and Storybook 8. Figures are directional.
Steers the company that turned open-source goodwill into a paying business.
A core Storybook maintainer who helps build the infrastructure rendering millions of snapshots.
Long-time voice of Storybook and Chromatic, championing component-driven development.
Four of Storybook's core maintainers work at Chromatic - which is either a conflict of interest or the entire point, depending on how you feel about open source.
Most mission statements are wallpaper. Chromatic's is closer to a job description. Every snapshot the service captures is one more interface that renders correctly for one more person. The company makes money by removing a specific, daily category of frustration - the broken layout, the clipped text, the button that vanishes on mobile - from the web at large.
The business model keeps it honest. Storybook stays free, which keeps the team accountable to a community of developers who would notice immediately if the open-source work were neglected. Chromatic charges for the cloud service, priced on snapshots, so revenue grows in step with how much testing teams actually do. The incentives point the same direction: more tested UI, fewer bugs reaching users.
The interesting twist is what happens when the developer pushing that one-line change is not a person at all. AI coding agents now generate interface code at a pace no human review process was designed for. They are fast, confident, and occasionally produce a login form that looks perfect in the diff and renders as chaos in Firefox. The faster the code gets written, the more valuable it becomes to automatically verify what it looks like.
That is the ground Chromatic is moving onto: a validated UI testing layer for humans and agents alike. The same machinery that catches a careless padding change catches a confidently wrong AI one. The category the founders bet on in 2017 - automatically checking how interfaces look - turns out to matter more, not less, as more of the web gets generated rather than written.
So return to that Tuesday afternoon. A change goes out - maybe from a developer, maybe from an agent. Somewhere, four screens quietly break. And before anyone ships, before any user notices, Chromatic has already looked at every pixel, found the problem, and flagged it. The internet stays slightly less broken than it would have been. That is the whole job. They do it a few million times a day.