AccessSutra AI documentation
Beyond the scan. Test the journey. Prove what works. A local-first, evidence-based accessibility testing CLI.
v0.2.173 passed · 1 skipped · 0 failed0 runtime dependenciesApache-2.0Honest: never claims conformance; untested is not pass
Executive summaryAccessSutra AI is a local-first command-line tool and agent kit that helps teams test accessibility tasks with evidence, not just run a scanFunctional specification- detected: static evidence in the source (for example an image with no alt attribute).ArchitectureCLI (packages/cli/accesssutra.mjs)Technical referenceaccesssutra version | doctor | rulesUser guidenpx accesssutra-ai doctorSecurity and privacyLocal first, least privilege, no telemetry, evidence over claims.Rules and evidence model18 rules. Full table with severity, candidate WCAG criteria and remediation: see docs/RULES.md in the repository.Test reportRecorded 2026-11-21T20:13:16.345Z on Node v24.15.0, win32/x64. Version 0.2.1. Generated by npm run proof; nothing here is hand-edited.Operations and releaseThe assurance report from the Ubuntu Node 24 run is uploaded as an artifact.Legal and governanceAccessSutra AI is software. Its output is not legal advice, a compliance determination, a VPAT or Accessibility Conformance Report, or a WCARoadmap and FAQDoes it make my site WCAG compliant? No. It finds some candidates and organises human testing.GlossarySample assurance reportReal HTML output of
accesssutra audit (checks, accuracy, fuzzing, sealed)Raw test log (TAP)Unedited machine output