// docs / changelog

Changelog

What's changed in the FixWeb scan engine — new checks, accuracy improvements, and false-positive fixes. Newest entries first.

July 17, 2026

IMPROVEDDesign and brand checks now read your pages visually. Our design, art-direction, logo, brand, and competitive-experience evaluations now score from rendered screenshots of your live pages — not just the raw HTML — so their results reflect what visitors actually see.

NEW31 expert evaluators join every scan. Scans now run a full panel of specialist evaluations — on-page and technical SEO, AI-search readiness, accessibility, design standards, content quality, editorial QA, performance, analytics readiness, brand and UX signals, and competitive posture — each scoring your site against a professional rubric and reporting evidence-backed findings.
NEWScorecards, charts, and a prioritized action plan. Completed reports now open with an executive diagnosis, per-evaluator scorecards with dimension-level detail, visual dashboards, and a 30/60/90-day action plan ranked by impact and effort. Live scans show progress, severity, and findings charts as they run.
NEWHonest coverage reporting. Evaluations that need data sources you haven't connected (Search Console, analytics, backlink or rank data) report exactly what's missing instead of guessing — dimensions are marked not-assessable rather than scored, and the report lists which data sources would unlock them.

NEWAI-search answerability scoring. A new GEO check scores each page for how well AI answer engines (ChatGPT search, Perplexity, Google AI Overviews) can extract it — heading structure, entity clarity, freshness, and authority — and flags low-scoring pages with the specific gaps to fix.
NEWStructured-data rich-result checks. A new GEO category validates your JSON-LD for Google rich-result eligibility — flagging missing required fields, invalid dates, URLs, and prices, and out-of-vocabulary values across Product, Article, Event, VideoObject, FAQ, and more.
IMPROVEDCrawlability checks expanded. FixWeb now flags a robots.txt that blocks all crawling (a site-wide Disallow: /), a robots.txt with no Sitemap: directive, and crawl-delay rules that can slow indexing — and no longer reports a missing sitemap when one is declared in robots.txt.

FIXEDSPA fallback false positives reduced. Path probes now compare suspicious responses against random-path baselines so client-side routers are not reported as broken pages by default.

NEWResearch backlog checks. FixWeb shipped additional website-quality modules for crawl/indexation, search presentation, semantic structure, structured data, media, performance, accessibility, forms, i18n/mobile/PWA, runtime, owner journeys, and repo quality.
NEWRepository website-quality check. GitHub repo scans can now flag template image issues and missing quality automation with evidence tied to source files.
NEWDeployment preview handling improved. Passive scans now treat generated deployment URLs as crawl targets and report website-quality evidence without legacy security scoring.

NEWPerformance-delivery module. Browser scans now estimate resource weight, script count, third-party pressure, and font volume.
NEWMobile/i18n/PWA module. Viewport, language, hreflang, and manifest basics now have dedicated findings.
IMPROVEDRendered crawl improved. Browser snapshots now capture forms, landmarks, images, headings, structured data, hreflang alternates, console errors, failed requests, and resource timing for every sampled page.

NEWRepo crawl-control check. GitHub repo scans can now flag missing robots.txt and sitemap files in common source locations.
NEWMedia and image-source checks. Repo and URL scans now connect rendered image issues back to likely template causes.
NEWWebsite-quality guardrail check. GitHub repo scans can now flag missing Lighthouse, axe, pa11y, web-vitals, or equivalent quality automation.

NEWRepo website-quality checklist. GitHub repo scans can flag template-level issues such as missing crawl-control files, image elements without alt or dimensions, and absent quality automation.
NEWRuntime-markup module. Browser scans now capture console errors, failed requests, and low-content renders so broken app shells show up in reports.

QUALITYCross-origin scope handling. Browser crawls now keep page snapshots scoped to the submitted origin and avoid carrying owner headers across unrelated hosts.
FIXEDOwner-depth scan path clarified. Verified-domain scans can carry a short-lived owner header into the browser crawl and purge it when the scan completes.
IMPROVEDSearch presentation scoring improved. Title, snippet, favicon, canonical, and Open Graph checks now grade rendered pages instead of raw HTML only.
IMPROVEDForm checks improved. Labels, placeholder-only fields, autocomplete hints, and mobile input metadata now produce clearer findings.
IMPROVEDMedia-quality checks improved. Image alt text, dimensions, and likely hero-image loading mistakes now appear in a dedicated media category.

FIXEDLive monitoring signal cleanup. External listing and DNS/certificate changes are now framed as website-monitoring events instead of scan defects.
NEWGitHub repo scans. Connect a repo and FixWeb checks source templates, crawl-control files, metadata patterns, image usage, and website-quality automation without ever loading your deployed site. See Scan types.
NEWTemplate pattern checks. Repo scans now look for repeated source-level causes of rendered page quality issues.
FIXEDFallback-route false positives reduced. SPA and edge fallback responses are compared against baselines before FixWeb reports broken rendered pages.
FIXEDRepo scan false positives reduced. Template checks now ignore comments, documentation snippets, and clearly server-only examples where possible.
FIXEDStructured-data checks added. Rendered JSON-LD syntax errors and missing homepage entity schema now surface as content findings.
FIXEDAccessibility baseline added. Language, main landmarks, skip links, unnamed buttons, and tap-target signals are now checked on rendered pages.
FIXEDRuntime-markup checks tightened. Browser console errors, failed requests, and low-content renders now produce clearer evidence and fewer false positives on SPA fallback routes.
FIXEDDomain verification handles apex ↔ www redirects correctly and is clearer about which value goes in the TXT-record Host field.

Each entry is tagged so you can skim:

Spot something that broke and isn't logged here? Email support@fixweb.app.