If you’re asking, “where do our sites actually stand, and is it getting better or worse?”, Tabwalk offers a self-hosted way to track automated accessibility findings over time. Its demo groups 66 failing elements into 13 rows—useful for seeing repeated issues more clearly, but not proof that every site will see the same reduction or that a scan can establish compliance.
What Tabwalk monitors
Tabwalk is a free, AGPL-3.0-licensed accessibility monitor that you run yourself. It discovers pages from sitemap.xml or sitemap_index.xml when available; otherwise, it follows links from the home page. It opens pages in headless Chromium using Playwright and runs axe-core checks against WCAG 2.2 A/AA rules. The project describes best-practice findings separately as recommendations. The Tabwalk repository documents its architecture and scope.
As an Amazon Associate I earn from qualifying purchases.
With recurring daily or weekly scans, Tabwalk compares a scan with the prior one to surface new and fixed findings and show count trends. That makes it a monitoring aid: it can help a team notice whether the issues its automated checks detect are changing, rather than treating one scan as a permanent verdict.
Why 66 failures become 13 rows
The numbers refer to Tabwalk’s included demo shop, not a general result. The demo reports 66 failing elements grouped into 13 unique-problem rows. Tabwalk fingerprints findings using the accessibility rule and normalized element markup, so repeated instances of the same issue can be represented by one row with a page count. For example, the same empty product-card link may recur across multiple pages.
#1 Best Overall
Grouping changes how findings are presented; it does not fix the underlying elements or reduce the number of failures. The demo is an illustration of deduplication, not an independent benchmark or a typical customer outcome. Artemy Andruschak’s project article describes the 66-to-13 example, and the project README calls it 13 unique problems on 66 elements.
How to run it and fit it into development
The documented deployment uses Docker Compose and includes Postgres, an API, a worker, and a web interface. Postgres also supports the job queue and schedule through pg-boss. A GitHub Action can run checks on pushes or pull requests; its documented defaults cap a run at 50 pages, and a configured severity threshold determines whether the action fails. Review the current README for exact setup steps and settings, since repository instructions can change.
Self-hosting gives your team responsibility for where and how the service runs. The repository documents the deployment components, but does not identify a required hosting provider. Choose an environment that can run the documented Compose setup and meet your organization’s operational and data-handling requirements.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →What automated findings can—and cannot—tell you
Automated checks catch only part of accessibility problems. The Tabwalk README warns that they catch less than half of WCAG problems and says the tool collects evidence rather than establishing legal compliance. Some checks need human judgment: axe-core can mark a result “incomplete” when it cannot decide automatically, such as a contrast issue involving a background image. Tabwalk retains these as findings that need a person to review.
Rank #3
Use scan results to prioritize investigation and track detected changes, not as a substitute for manual review by people who understand the interface and its users. A clean automated report does not demonstrate that a site is accessible or legally compliant. The project README puts it directly: “Tabwalk collects evidence. It is not a legal opinion and it does not make your site compliant. Automated checks catch less than half of WCAG problems; the rest needs a human. We show you which parts those are.”
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Check whether its current limits fit your site
The project documentation describes several constraints that can determine whether Tabwalk suits a particular workflow. These are source-time product details, so check the current repository before adopting it.
- Authenticated areas: The described version has no login support and does not scan pages behind a login.
- Reports: It does not provide report export in the described version.
- Scope: Its documented automated problem scope is WCAG 2.2 A/AA failures; recommendations are presented separately.
- Operation: The MVP is described as having no authentication and supporting a single organization.
- Remediation: It does not alter your site and does not add an overlay or widget.
Before relying on it, compare those capabilities with your needs: recurring whole-site coverage or selected URLs, CI checks and severity thresholds, human-review capacity, data control, authenticated-page coverage, and report export.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsWho should consider Tabwalk
Tabwalk is worth evaluating if your team can run Docker Compose, wants recurring automated scans, and can follow detected issues with human review and fixes. Its demo shows how deduplication can turn a long list of repeated elements into a smaller set of actionable patterns. It is a poor fit if you need built-in authentication, scans of logged-in pages, exported reports, or an automated declaration of compliance.
For setup details and the project’s current capabilities, consult the Tabwalk GitHub repository. The demo figures are described in Andruschak’s September 29, 2026 article; the repository README was accessed October 7, 2026.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
