chrome://extensions
TOAD
Technical On-Page & Auditing Dashboard
Instantly audit SEO, accessibility, and performance — directly inside Chrome. A clean, real-time dashboard that helps developers, marketers, and SEO pros identify optimization opportunities faster.
Free · No account required
4
Audit Categories
100%
Local Processing
0
Data Sent Externally
Everything You Need for a Smarter Audit
A tool built by developers for developers.
Metadata Auditing
Checks title, description, canonical, hreflang, and social tags. Ensure accurate and complete metadata for every page.
Robots Directives
Reads robots.txt, sitemaps, and meta robots instructions. Confirm indexability and crawl rules instantly.
Schema Insights
Detects and validates JSON-LD and microdata markup. Verify eligibility for Google’s rich results.
On-Page Structure
Maps H1–H6 hierarchy, semantic tags, images, and videos. Strengthen accessibility and on-page SEO.
See TOAD in Action
Analyze any webpage in real time with a categorized popup interface that breaks down technical insights into clear, visual sections.
How It Works
Get started in seconds — no setup, no accounts, just insights.
2
Load Page
Visit any webpage you want to audit.
3
Click Icon
Click the TOAD icon to view your instant audit report.
Need Support
Found an issue or want to suggest a feature?
Trust & Privacy
TOAD is built to audit your pages without ever leaving your browser.
About the Publisher
TOAD is developed and maintained by Mktg.Tech, a leading provider of marketing technologies.
Limited-Use Disclosure
In compliance with Google’s User Data Policy, TOAD makes the following commitments:
- Allowed use: Any permissions or page content accessed are used solely to provide and improve the SEO audit feature.
- No external transfer: Audit data never leaves your device. The only network requests are to fetch the latest Schema.org vocabulary and to retrieve the audited page’s own robots.txt and HTTP headers.
- No advertising: No collected data is used or sold for personalized advertising or profiling.
- No human access: No developer or employee can view your audit data — all processing occurs locally in your browser.
Privacy & Data Handling
- TOAD does not send page content, URLs, or user data to any server.
- The only external request is for the Schema.org catalog, cached locally.
- All processing happens entirely within your device’s browser sandbox.
- You can remove all cached data at any time by uninstalling the extension.
Recent Updates & Change Log
Version 1.5.1
Latest
August 2026
New: View Comparison
- A button under the Rendered section opens a new tab with the raw HTML the server sent on the left and the rendered HTML on the right, additions in green and removals in red
- Content is aligned by matching lines, so a tag inserted mid-page does not push everything after it out of step
- A unified single-column mode, a changes-only mode, next and previous buttons that walk every change in order, and keyboard shortcuts
- The capture is taken once and never goes stale silently: reopening the tab after a reload says so rather than showing an old comparison
- Stays responsive on documents over 1MB with the whole page displayed
New Check: Favicon
- The 23rd check, weight 4, in the META section. Always judged on the homepage, because that is where crawlers look for it
- Reads the
<link>tag, fetches the file and decodes its real pixels (ICO, PNG, GIF, JPEG, SVG, WebP), and requires the icon to be square and at least 8×8 - Verifies that both the homepage and the icon are reachable and not blocked by robots.txt
- Flags invalid or unsupported markup, and unstable URLs: a cache-busting token, a hashed file name, or an address that changed since the previous audit of the same site
- The Apple touch icon and web manifest icons are validated whenever they are declared; ones that are absent do not fail
- Every failure names its own reason, and the fix report says what to change
JSON-LD: Double-Escaped Entities
- A structured data value escaped twice now fails the check.
Cats &amp; Dogssurvives one unescape pass with markup still in the visible text - The message names the schema block, the property path, and the offending text
- A single escaping pass stays legal, and entity names that do not exist are not flagged
Scoring Model Now 23 Checks
- Favicon enters at weight 4 and those points join the total rather than being taken from other checks, so the model now totals 104 across 23 checks
- No rebalance was needed. The score divides by the live total weight, added in 1.4.4 for exactly this case, so every share re-derived itself
Smaller Changes
- The Scoring Weights popup opens inside the TOAD window instead of covering the page
- The internal link threshold is raised to 200; external stays at 100
- A nofollow on a same-page
#link is no longer an error - Next Issue shows both numbers, which stop you are on and how many checks are failing, so it reads together with the toolbar badge
Verification
- 147 of 147 end-to-end tests passed
Version 1.4.4
August 2026
New Check: Rendered
- Re-fetches the page with caching disabled and compares the HTML the server sends against the page after JavaScript has run — across six dimensions in both directions: metadata, headings, visible text, images, links, and structured data
- Every added or removed item is listed individually, in the panel and in the fix report, rather than being reduced to a count
- A summary reports what share of the rendered page the server actually sent, counted in items, floored never rounded, and colored by pass or fail rather than by the size of the number
- A re-fetch that fails reads “could not compare” and stays out of the score, so a network problem can never be mistaken for a clean page
- The extension’s own panel is excluded from the comparison, so it never reports its own interface as page content
Scoring Model Now 22 Checks
- Rendered enters at a weight of 8, taken from checks whose influence on indexing is the most indirect: Meta Description 4→3, Open Graph / Twitter 4→3, X-Robots-Tag 6→5, Sitemap 6→5, JSON-LD 5→4, Images 4→3, Validation 4→3, Mobile 4→3
- The weights still total exactly 100, and that total is asserted at startup rather than trusted
- The score now divides by the live total weight instead of a hardcoded 100, so a future check can be added at any weight without another rebalance
- The toolbar badge and the Scoring Weights panel both count 22
Detection Corrections
- The URL check reads the page’s own URL only — uppercase letters, underscores or query parameters in the pages it links to no longer fail it
- robots.txt rules are read inside their own
User-agentgroup, so aDisallowwritten for one named crawler no longer blocks the page for every crawler;Allowrules count too, with the longest matching rule winning - Semantic HTML counts landmarks rather than tags: headers and footers inside cards and sections are valid, and a hidden
<main>(an inactive route in a single-page app) is not a second main - Custom elements injected into
<head>by tag managers and testing tools no longer fail Validation, while real HTML elements that belong in the body still do
Errors Reported on the Element That Failed
- The H1 and H2 summary rows are gone. Each heading error now shows in red on the offending heading itself, which stays clickable to locate it on the page; only a missing H1 or H2 keeps a standalone card, since there is no heading row to attach it to
- Image EXIF is neutral when absent: a page whose images carry no EXIF metadata, or that has no images, keeps its full weight and shows gray with “not counted toward the score”
Finding Issues on the Page
- Next Issue visits every issue strictly in display order, stopping once per offending element even when it trips several conditions, and expands Schema sections so the highlighted card is visible
- Clicking a heading scrolls the live page to it and flashes a highlight; headings are matched by tag and text rather than by position, so they survive pages that change after analysis, and a hidden heading highlights its nearest visible container
- The floating icon and panel are always pulled back into view and can no longer sit off screen
Verification
- 128 of 128 end-to-end tests passed, covering every check toggle on the QA page and each of the twelve Rendered fail modes
Version 1.3.4
August 2026
Weighted 100-Point Scoring
- Every page now gets one score out of 100, split across 21 checks with fixed weights that reflect how much each signal influences indexing and rankings — full weight on pass, zero on fail
- Neutral states (no pagination tags, no hreflang, no meta robots tag, no video) count as passing — a signal that’s simply absent is never a failure
- Bundled checks (Semantic HTML, Validation, Links, X-Robots-Tag, Sitemap) pass only when every one of their sub-signals is clean
- The toolbar badge now counts failing checks at a glance, and a new Scoring Weights panel lists all 21 checks with their weights
Expanded Check Coverage
- New URL category: multiple slashes (path-only — a protocol’s // never counts), spaces, non-ASCII characters, uppercase, repetitive paths, query parameters, underscores, and URLs over 115 characters
- New Validation category: documents over 2MB, individual resources over 2MB, multiple
<head>/<body>tags, and invalid elements inside<head> - Full hreflang cluster testing: self-reference, x-default, wrong language mappings, invalid codes, and links that don’t resolve
- Expanded heading checks across H1 and H2: duplicates, skipped levels, out-of-order headings, and over-length text
- Link quality checks: nofollow links present, nofollowed internal links, empty anchor text, and 100+ internal or external link volumes
Real Measurements & Specific Messages
- Meta description width is measured in real rendered pixels — the way Google actually truncates — not a character-count guess
- Document weight is measured on the live page, hreflang URLs are actually fetched, and EXIF is read from the image’s binary data
- Every failing check shows its own specific message (for example: “Multiple conflicting canonical tags found”, “Description renders wider than 985px”, “A heading level is skipped”)
- Image issues now highlight the offending image’s own card with the message, instead of a callout at the top
Downloadable Fix Report
- A new download icon in the bottom-left corner saves a fix list for the current page: every failing check, what exactly failed, and what to change on the website — grouped by tab, with passing checks left out
Refreshed Look
- Section emojis replaced with clean single-color SVG icons in the TOAD theme colors, on every tab
Version 1.2.5
July 2026
Performance & Experience Improvements
- Instant page rendering
- Caching — analyzed pages are cached for an hour so reopening/navigating is instant; a refresh button forces a fresh run anytime
- Reorganized the health bar settings to match the tab structure for easier navigation
In-Page Widget
- Floating icon on every page opens the auditor directly on the page (same content as the popup)
- Stays open as you navigate and refresh
Markdown for Agents
- Markdown is broken out by LLMs.txt and readme.md files
Social Image Improvements
- Added an image dimension overlay on the featured image
- Added the full info line (Example: Image: 1080 × 500px · 2.16:1 · Recommended 1200 × 630) under the OG Image / Twitter Image fields where the URLs are
Version 1.1.0
February 2026
Performance & Experience Improvements
- Overall performance enhancements for faster processing and responsiveness
- Prioritized loading of First tab content for improved initial load speed
- Added health processing progress bar for clearer status visibility
- Error button hidden until health processing is complete
- Deferred image EXIF extraction until after health processing
- Deferred image downloads until after health processing
Sitemap Enhancements
- New sitemap scanner for improved sitemap discovery
- Added sitemap submission functionality
- Automatic checks for common sitemap paths during initial health processing
Image Handling Improvements
- Increased default image file size threshold from 100KB to 200KB
- Updated image file size error messaging to reflect 200KB threshold
- Images sorted from largest to smallest for better prioritization
Markdown for Agents
- Detects whether a page supports clean markdown output
- Displays token size comparison between initial HTML and generated markdown
- Shows percentage token reduction to illustrate efficiency gains
- Provides content signal indicators: ai-train = yes, search = yes, ai-input = yes
- View full generated markdown in an expandable dropdown
- Download complete markdown file (.md)
Next Up: Analytics Tab with Google Search Console integration
Version 1.0.0
January 2026
Initial Release
- Core audits: Meta, Directives, Schema, and On-Page
- Schema.org catalog caching
- Robots.txt and sitemap parsing
- H1–H6 and semantic structure analysis
- Local caching and privacy-first design
Next up:
- Optional (opt-in) anonymous analytics
- Core Web Vitals audit
Enhanced visual health scoring