The process

Two URLs in. One verdict out.

MigrateKit takes your old site URL and your new site URL. It runs seven verification suites and returns a pass/fail scorecard. You fix what failed, re-run, and ship when everything is green.

Three steps

  1. Point it at both sites.

    The old site live, the new one on staging. Any platform, any stack: Wix, Squarespace, Duda, WordPress, Webflow, Next.js - anything with a URL.

    No CI integration required on the old site. That's the point. Most rebuilds start from a site that never had a CI pipeline. MigrateKit reads rendered pages, not source code. If it loads in a browser, it can be scored.

  2. The suite runs.

    Seven checks, every breakpoint, before/after metrics. A full pass takes minutes.

    What runs: responsive visual comparison at four breakpoints (1440, 1024, 768, 375px), SEO tag preservation (titles, canonicals, OG tags, structured data), content completeness (every text block, link, image, heading), Core Web Vitals before/after, WCAG 2.1 A/AA regressions, link and form functionality, image quality and format.

  3. Read the scorecard, fix, re-run.

    Pass/fail per check, with the exact failing pages, tags, and elements listed. Fix what failed, re-run, and repeat.

    Re-run as you fix things. Fix what failed, re-run, and repeat until the rebuild is clean.

Seven verification suites

Each suite has a target score and a pass/fail gate. All seven must pass for the overall scorecard to pass.

  • responsive

    target: 85%

    Pixel-level visual comparison at 1440, 1024, 768, and 375px. Flags layout shifts, missing elements, and size regressions at each breakpoint.

  • seo

    target: 80/100

    Compares title tags, meta descriptions, canonical URLs, Open Graph tags, heading structure, structured data (JSON-LD), robots directives, and image alt text between old and new.

  • content

    target: 85/100

    Checks that every text block, internal link, image, and heading from the old site is present on the new site. Word count matched. Missing content listed exactly.

  • perf

    target: 80/100

    Measures LCP, FCP, CLS, TBT, and TTFB before and after. Checks page weight and request count. Flags any Core Web Vital that regresses. No regression = pass.

  • a11y

    target: 90/100

    Runs WCAG 2.1 A/AA checks on both sites. Reports only violations the rebuild introduced - if an issue existed on the old site, it's noted but not penalized.

  • functional

    target: 90/100

    Checks all internal links resolve, forms submit correctly, mobile menu opens, and anchor links scroll to the right position. Broken links listed individually.

  • image

    target: 85/100

    Verifies all images load, none are broken, resolution is preserved, no images are upscaled or distorted, and modern formats (WebP/AVIF) are maintained where used.

What you get back

Pass/fail scorecard

A clear overall verdict (pass or fail) with scores per suite, target thresholds, and the exact failing pages, tags, and elements. Readable by a developer, presentable to a client.

Machine-readable JSON

Every scorecard exports to JSON - the same structure as the visual report. Drop it into a CI pipeline, a Slack notification, or an agent workflow. The JSON is the same data as the scorecard, nothing extra, nothing missing.

Store catalog and link check

For stores moving to Shopify, the comparison also covers every product, price, and picture, and verifies every old link still lands somewhere sensible on the new store. That is where most store-move problems actually hide. The result comes as JSON you can hand to whoever did the move.

How this differs from Lighthouse and Percy

Lighthouse scores one site in isolation. Percy and Chromatic compare a site to its own last commit on the same codebase. Both are excellent tools for what they do.

MigrateKit does something different: it compares two different sites on two different stacks - which is what a rebuild actually is. There is no commit history to diff against. The old site is the reference. No CI integration on the old site required, because you usually don't have one.

The baseline is what the client already had: the live site that built the rankings you're trying to keep. Not a style guide, not a spec.

Ready to run it on a real project?

The free scan reads one page in about 20 seconds, no account. The full Verification Report runs both sites across all seven checks, and it is free.

Get free scan See pricing