Siteimprove governs an estate. We hand a developer the next change.
Siteimprove crawls whole digital properties, including documents, and reports a score upward to whoever owns the programme. We scan the pages you name, rank the fixes by how many issues each one clears, and say plainly which criteria were never checked.
Scan a public URL free · no credit card required
Choose Siteimprove if
- You own the standard across many sites you did not build, and need to know the size of the problem, whether it is shrinking, and who is responsible for which part.
- Documents are part of the problem. PDFs and media are scanned too, which for a public-sector body or a bank is frequently the bigger half.
- Somebody who will not read a scan report is going to ask you for a number — a score, a dashboard, a board-ready report.
- Non-developers need to act in their own tools: design, editor and publishing workflows, plus analytics, search and content in the same platform.
Choose TestKase if
- You have a fortnight, a backlog, and a component missing a label in ninety places, and you need to know what to change first.
- You want to know whether yesterday's change worked — fixed, introduced and unchanged, matched issue by issue.
- You want criteria with no automated rule reported as not checked rather than quietly counted as passing.
- You want it in the same workspace as the device testing, the automated suites and the test cases, with the price on the page.
The head-to-head
A difference in kind, not in quality. Most of these rows go to whichever product matches the job you are doing.
| Item | How the two compare | For us |
|---|---|---|
| Coverage | Siteimprove crawls entire digital properties rather than a list of addresses. A scan here takes up to ten URLs you choose, on every plan including the free one — deliberately small: the pages under change this sprint, not the estate. | Theirs |
| Documents and media | PDFs and media are scanned as part of their estate coverage, mapped against WCAG, ADA, Section 508 and the European Accessibility Act. Nothing here touches a document. | Theirs |
| Reporting upward | An accessibility score, a wider certainty index, executive dashboards and management-ready reporting that ties improvement to outcomes a board recognises. That is exactly the artifact a programme office needs and exactly the thing we chose not to build. | Theirs |
| Reaching non-developers | Checks surface inside design tools, editors and publishing workflows so a content author can act without learning the standard. We have no CMS publishing integration. | Theirs |
| Platform breadth and services | Accessibility sits alongside analytics, search and content strategy in one platform, with professional services beside it. For a marketing organisation that is a consolidation argument we cannot answer. | Theirs |
| What the report opens with | Ours opens with the ranked list of code changes, ordered by how many issues each one clears, with severity as the tie-break, a running total per row and the selector the fault concentrates on — so you can tell a one-file fix from a forty-file one. | Ours |
| Did my change work | Every run is compared with the one before it and reports fixed, introduced and unchanged, matched on the rule, the page and the element selector. A falling total is not the same claim as a fixed problem. | Ours |
| Where it lives | The scanner shares a workspace with manual sessions on real devices, automated suites, visual baselines and test cases, so an accessibility finding becomes a defect in the same tracker as everything else. | Ours |
| Seeing a price | Siteimprove routes to a demo and its pricing page carried no figures when we read it. Every plan here is published, and the scanner has a free tier that scans real addresses rather than a sample report. | Ours |
| Proving conformance | Neither can, and no automated tool can. Siteimprove sells professional services alongside the platform, which is a genuine route towards a defensible position that we do not offer. | Neither |
- Addresses per scan, on every plan including the free one
- 10
- WCAG 2.2 A and AA criteria with an automated rule behind them
- 23 of 55
- Numbers every run reports: fixed, introduced, unchanged
- 3
- Scan history and schedules kept
- 90 days
Our figures, from the Accessibility product pages in this site. There is no comparable Siteimprove figure to put beside them: their pricing page carried none when we read it on 11 September 2026 and routes to a demo instead.
These are not the same purchase
Somewhere in a large organisation there is a person responsible for accessibility across forty websites they did not build, in six languages, owned by teams who do not report to them. They need to know the size of the problem, whether it is shrinking, and which department is responsible for which part of it. They will be asked for a number by somebody who will not read a scan report.
Somewhere else there is a developer who has a fortnight, a backlog, and a component that is missing a label in ninety places. They need to know what to change first and whether yesterday's change worked.
Siteimprove is built for the first person and is good at it. We are built for the second. Most of what follows is that one sentence, expanded.
Why we did not build a score
A score is a good summary and a bad instruction.
A team opening a report that says 218 issues and 71 out of 100 knows one thing — there is a lot of work — and nothing about where to start. What usually happens next is that somebody sorts by severity and works down, fixing critical faults one element at a time, which is close to the slowest possible order.
Our report opens with the ranked list of code changes, ordered by how many issues each one clears, with severity as the tie-break. One change usually clears every instance of a fault, so a moderate problem repeated forty times and a serious one on a single element are the same typing for very different results. Each row carries a running total, which turns a backlog into a budget: three fixes clear 47 of your 61 problems, and the fourth clears two.
There is a second reason, and it is less about workflow. A score is the artifact most likely to be forwarded to somebody who did not run the scan and read as a statement about conformance. It is not one. Fewer than half of the WCAG 2.2 AA criteria have an automated rule behind them at all, and our reports mark those as not checked rather than passed.
What a developer tool looks like instead
Four choices that follow from being bought by the team writing the code.
| Item | What it does |
|---|---|
| Ten addresses, chosen by you | A scan takes up to ten URLs, on every plan including the free one. That is a deliberately small number: the pages under change this sprint, not the estate. Nothing here discovers a site for you. |
| Did my change work? | Every run is compared with the one before it and reports fixed, introduced and unchanged, matched on the rule, the page and the element selector. A falling total is not the same claim as a fixed problem, and introducing three while clearing five is invisible to a score. |
| A real browser, and pages behind a login | Scans run in a desktop browser on the same machines that serve manual sessions, so a page that needs JavaScript is scanned after it renders. Authenticated scanning signs in with cookies, headers or a form, with credentials encrypted under a key held outside the database. |
| Beside the rest of the testing | The scanner shares a workspace with manual sessions on real devices, automated suites, visual baselines and test cases. An accessibility finding becomes a defect in the same tracker as everything else rather than in a separate system nobody opens. |
Where Siteimprove is ahead
Five things they do that we do not. If the job is governing an estate, that list is disqualifying and we would rather say so here than in month two.
- Site-wide crawling — whole properties, not a list of addressesNot ours
- PDF and media accessibility — often the larger half of the problemNot ours
- A score and executive dashboards — the artifact a programme office reports withNot ours
- Checks in design and publishing tools — so non-experts act where they already workNot ours
- Professional services — expert remediation alongside the platformNot ours
The rest of the category
Four answers to the same question, and what each of them is actually for.
| Item | What it is for |
|---|---|
| Siteimprove | The governance platform: estate-wide crawling including documents, a score and dashboards for reporting upward, role-based distribution into design and publishing tools, and analytics, search and content in the same subscription. |
| Level Access | Level Access occupies similar ground with a heavier services and training component, and owns an overlay product as well. |
| Deque | Deque is the deepest engineering tooling in the field, and its axe-core engine is what our own scanner runs. |
| TestKase | The developer end: the pages you name, fixes ranked by reach, a run-to-run comparison that names what you introduced, criteria nothing checked marked not checked, and no per-seat charge — inside the same workspace as the device testing. |
What this does not do
- No crawling.
- Up to ten addresses per scan, chosen by you. Nothing discovers pages, subdomains or orphaned content.
- No documents.
- PDFs, office files and media are out of scope entirely.
- No score, no trend line.
- A run is compared with the previous run only. There is no maturity index and no chart over time.
- No certification and no services.
- No VPAT, no audit letter, no expert remediation team, and nothing here is legal advice.
- No public link for a scan.
- Results are shared with teams inside the workspace. There is no address you can send to an external auditor.
- Automation reaches under half the standard.
- 23 of the 55 WCAG 2.2 AA criteria have an automated rule. The rest are reported as not checked, which is honest rather than reassuring.
Frequently asked questions
No. A scan takes up to ten addresses that you choose, on every plan including the free one. If your requirement is discovering every page across several domains and watching all of them, that is what Siteimprove is built for and we are the wrong shape. Ours is aimed at the pages a team is actually working on this fortnight.
No, deliberately. A score is a reasonable summary for an organisation reporting quarterly and a poor instruction for a developer with a fortnight. Our report opens with the ranked list of code changes and the running total of what each one clears, because that is the artifact somebody can act on. It is also the artifact least likely to be mistaken for a conformance claim.
We do not check them. Siteimprove covers PDFs and media as part of its estate scanning, and for a public-sector or financial organisation with thousands of documents that is often the larger part of the problem. Nothing here touches a document.
Siteimprove is generally bought centrally — by a digital governance, marketing or compliance function that owns the standard across many sites and needs to report progress upward. Ours is bought by the team writing the code, and it lives in the same workspace as their device testing and test cases. Those are different budgets and different conversations, which is usually what decides this comparison.
Siteimprove groups issues, attaches guidance and puts a score and a dashboard on top, then distributes work into the tools each role already uses. Ours ranks the fixes by how many issues each one clears, shows the selector the fault concentrates on so you can tell a one-file fix from a forty-file one, and compares every run with the one before it to show what you introduced rather than only what the total did.
With us, yes — every plan is published, the accessibility scanner has a free tier, and the device testing next to it is billed per parallel thread rather than per person. Siteimprove routes to a demo; we read their pricing page and it carried no figures. That is a normal enterprise model and it is also a real difference in how long an evaluation takes.
Neither, and no automated tool can. Fewer than half of the WCAG 2.2 AA criteria have any automated rule behind them. We report those as not checked rather than passed. Siteimprove sells professional services alongside the platform, which is a genuine route towards a defensible position that we do not offer.
Related
- TestKase vs Level AccessAccessibility scanning without a services engagement
- Accessibility overlaysWhy we do not sell one, and what the evidence says
- TestKase vs Sauce LabsReal and virtual devices, per-thread pricing, and one workspace
- All of Why TestKaseHow we compare with the tools you are probably already paying for
- PricingPer-seat test management and per-thread device testing, priced on the page
See what three changes would clear
Scan a public address and read the plan it produces. No demo, no sales call, no credit card.