TOAD: Page Auditor

AI assistants are answering questions about your industry right now. Whether your page is part of those answers depends on what an AI crawler can actually fetch, parse, and trust, and almost no one is checking. TOAD (Technical On-Page & Auditing Dashboard) shows you. One click on any page reveals what AI crawlers and search engines see, scored across 23 weighted checks, entirely inside your browser. No account, no limits, no page data leaving your machine. AI SEARCH VISIBILITY ★ llms.txt and readme.md detection, the files AI systems look for to understand your site ★ Markdown export built for AI ingestion, with a token count comparison showing how much cheaper a clean page is to process ★ Content signals for AI training, search, and input surfaced directly from the page ★ Rendered versus raw HTML diff. Most AI crawlers do not execute JavaScript, and this is where pages quietly disappear ★ Structured data validation, the vocabulary AI systems use to resolve entities and attribute facts to a source Generative engines reward the same things classic crawlers do: clean markup, resolvable entities, content that exists in the HTML. They are just far less forgiving when those things are missing. TOAD audits both in the same pass. THE FULL TECHNICAL AUDIT ★ Metadata: title, meta description, canonical, hreflang, and social sharing tags, normalized and flagged when missing or misconfigured ★ Robots and indexing: page level directives, response headers, robots.txt rules, and sitemap discovery, so you know what is genuinely crawlable ★ Schema and structured data: JSON-LD and microdata detection, validated against a locally cached schema.org catalog ★ Page structure: heading hierarchy, semantic tags, images, videos, and favicon validation ★ Image EXIF data extracted inline for every image on the page WHAT YOU GET ★ A weighted score across 23 checks that places the page at a glance ★ Findings grouped into SEO, Social, Robots, and Schema so nothing gets buried ★ A downloadable fix report you can pass straight to a developer or client ★ A floating widget that audits without opening the popup ★ Audits that work on staging, localhost, and IP restricted environments no cloud crawler can reach BUILT FOR SEO and GEO specialists auditing client sites, agencies proving technical work to clients, frontend developers shipping template changes, QA testers validating CMS output, and marketers who need to know why a page is not being cited. HOW THE SCORING WORKS TOAD runs 23 checks and assigns each one a fixed weight. A check earns full weight on pass and zero on fail. Neutral states, where a signal does not apply to the page, count as passing, so a page is never penalized for something it does not need. The current scale totals 104 points. The heaviest weights sit on rendered HTML comparison, indexing headers, and sitemap discovery, because those are the failures that cost the most traffic when they go unnoticed. Schema and favicon checks carry mid weight, and the remaining checks divide the balance of the scale. WHAT EACH MODULE INSPECTS Metadata Values are normalized before comparison, so a trailing space or inconsistent casing does not read as a false positive. Anything missing or self contradictory is flagged alongside the value TOAD actually found on the page. Robots and indexing The page's own robots.txt is fetched live and parsed against the current URL, then reconciled against directives in the document and in the HTTP response headers. Sitemap references are discovered and reported. The result tells you whether this specific URL is crawlable and indexable, not whether the site generally is. Schema and structured data Markup is detected, parsed, and validated against a schema.org vocabulary catalog cached locally on your machine. Output covers rich result eligibility and flags double escaped entities, where a value still contains markup after one unescape pass, a failure mode that passes casual inspection and breaks parsing downstream. Page structure Heading hierarchy is mapped and checked for duplicates, skipped levels, headings out of order, and heading text running past a usable length. Semantic HTML is measured by counting landmarks rather than raw tag counts, so a page wrapped in divs does not score the same as one with real structure. Images are inventoried with EXIF data read from the binary. Link quality covers empty anchor text and nofollow applied to internal links. Rendered versus server HTML TOAD compares the HTML the server sent against the DOM after JavaScript has run, across six dimensions: metadata, headings, visible text, images, links, and structured data. The output reports what share of the rendered page the server actually delivered. This is the check that catches client rendered content invisible to crawlers that do not execute JavaScript. Favicon Evaluated on the homepage. Checks that the icon loads in a standard image format, is square, and meets a minimum size of 8 by 8 pixels. Flags cache busting tokens and unstable URLs that prevent reliable caching. Markdown for agents Detects whether a markdown version of the page is available, along with llms.txt and readme.md. Reports the token reduction against the full HTML, provides a downloadable file, and surfaces the page's AI content signals. Fix report Every failing check exported as a document listing what failed, why it failed, and what to change on the site. Built to hand to a developer or attach to a client deliverable without rewriting it first. PRIVACY AND DATA HANDLING All processing happens locally in your browser. The only outbound requests TOAD makes are for the schema.org vocabulary, which is cached on your machine after the first fetch, and for the audited page's own robots.txt and HTTP headers. URLs and page content are never sent to a server. There is no account, no tracking, and no mechanism by which anyone can view your audit data.