Visual regression, on autopilot

Catch every visual
regression before
your users do.

Side By Side screenshots your sites across every environment and viewport, then pixel-compares each deploy against a trusted baseline. Layout breaks, broken images, console errors and style regressions surface in seconds, not support tickets — and when a run finds them, Side By Side emails you instead of waiting for you to open a dashboard.

Open the Dashboard See how it works
3environments
prod · staging · dev
∞viewports
desktop · tablet · mobile
AIdiff triage
false-positive aware
What it does

One platform for everything that can visually break.

Pixel-perfect diffs

Capture a baseline, then compare every later run against it. The dashboard highlights exactly which regions changed and reports a diff percentage per URL and viewport.

Shift-tolerant comparison

Content that simply moved down the page isn't a bug. The comparator detects vertical displacement and reports a corrected diff so you chase real regressions, not noise.

Console error checks

Run console-only passes to catch JavaScript errors and failed requests across your tracked URLs — no screenshots required.

AI diff triage

Claude reviews each diff image and classifies it — content change vs. real regression — so reviewers skip the obvious false positives.

Multi-environment

Production, staging and dev, each with its own URL and auth. Compare like-for-like or promote a baseline between environments.

VPN / proxy aware

Test sites behind a firewall or geo-gate through per-environment SOCKS5 proxies — gated staging is a first-class citizen.

How it works

From register to regression report in five steps.

  1. 01

    Register a site

    Add a site with its environment URLs and the viewports you care about. Organize sites under your organization.

  2. 02

    Discover & pick URLs

    Crawl the sitemap or add paths manually. Record pre-screenshot actions (click, hover, scroll) with the Chrome extension.

  3. 03

    Capture a baseline

    The runner screenshots every URL × viewport and stores it as the reference of record.

  4. 04

    Run a comparison

    After a deploy, run again. Each screenshot is diffed against the baseline; changed regions are highlighted.

  5. 05

    Review & classify

    Triage diffs side by side — mark false positives or intentional content changes. AI pre-sorts the queue.

Built for real workflows

The details that make it trustworthy.

Job queue

Queue baselines, comparisons, console checks, retakes and thumbnails — processed by a background worker.

Single-screenshot retake

Re-capture one URL/viewport inside an existing run without rerunning the whole site.

Viewport presets

Define global presets once; sites reference them by name. Per-URL viewport filters when you need them.

Pre-screenshot actions

Click, hover, scroll, type, wait — captured visually so dynamic states render consistently.

Chrome extension

An in-page drawer + action recorder, with a live queue badge, right where you browse.

Bookmarks & compare modes

Side-by-side, overlay and onion-skin comparison, with bookmarks for the results that matter.

When something changes

You don't have to be watching.

A monitoring tool that only reports into a dashboard is a dashboard. When a run finishes, Side By Side emails the people in your organisation who asked to hear about it — once to say the run is done, and again if it actually found differences worth looking at.

  • Run finished — one email when a baseline, comparison or console run completes, fails or comes back partial
  • Differences found — a second email listing the pages and viewports that changed most, only when the run reported at least one diff
  • Two switches on your account page, both on by default, both honoured — including the one-click unsubscribe link in the footer of every message
  • Included on every plan, Free included. Notification is the product, so we do not sell it back to you

What this is not: there are no outgoing webhooks, no Slack integration and no usage-threshold emails at 80% or 100% — none of those are built, on any plan. Email on run completion and on differences found is the whole of what we send about your runs.

Run finished“[clientsite.com] Run Completed — 3 differences” · 90 screenshots compared, 3 to review, with a link straight into the run.
Differences found“[clientsite.com] 3 visual differences to review” · the largest ones listed by page and viewport: /pricing/ desktop 28.20%, /pricing/ mobile 31.70%.
Every planFree, trial and paid alike — there is no plan check anywhere in this path.
Organizations & access

Multi-tenant by design.

Every site belongs to exactly one organization. Invite your team — they get access to all of that organization's sites, and nothing else.

  • Organizations own sites and users
  • Org admins manage their own members
  • Members operate every site in their org
  • Strict isolation — no cross-tenant access
Super adminManages all organizations & global config
Org adminCRUD sites + users within their org
MemberRuns & reviews their org's sites
ServiceAutomation account for the test pipeline

Stop shipping visual regressions.

Open the dashboard, register your first site, capture a baseline in minutes — then let it email you when something changes.

Open the Dashboard