v32: SEO Batch 2 and the AI Category Begins

Two changes ship today. The SEO Suite grows from 19 to 25 tools with a second batch focused on technical-markup validation and page auditing. And a new AI category launches with three seed tools — a prompt template builder, an LLM comparison reference, and a curated system prompt library. Everything still runs in the browser. Still no uploads, no signups, no API keys.

The SEO additions close gaps the v31 batch identified but didn't address — rich-result eligibility under Google's 2026 rules, Schema.org structural validity (a different question from rich-result eligibility), breadcrumb markup that has to match its visible counterpart, link-level audit at the per-page granularity Lighthouse skips, mobile-readiness signals from static HTML, and a working reference for the Core Web Vitals thresholds now that INP has replaced FID. The AI additions are deliberately a starting line, not a finish — the heavier transformers.js work lands in v33.

Rich Results Tester

Rich Results Tester paste JSON-LD or a full page HTML and check eligibility against Google's 2026 rich-result rules. A schema generator from 2022 will happily emit FAQ markup that validates against Schema.org but earns no rich result anymore — FAQ rich results have been restricted to authoritative government and health domains since 2023, and HowTo rich results were retired entirely. The tester encodes those eligibility shifts, not just structural validity. Three layers of check per block: parse validity, structural completeness, current eligibility. Pairs naturally with the existing Schema Markup Generator for the build-then-test loop.

Schema Validator

Schema Validator answers the other question: is this valid Schema.org? Required fields, value-type checks (URL vs Text, Date vs DateTime, Number vs String), nested-object structure, recognized property names. It walks inheritance chains so a Recipe block can inherit HowTo properties correctly. Crucially separate from rich-result eligibility: a block can be perfect Schema.org and earn no rich result, and the two checkers are designed to be used together rather than collapsed into one ambiguous pass/fail.

Breadcrumb Builder

Breadcrumb Builder generates BreadcrumbList JSON-LD plus the matching visible HTML, with both kept in lockstep — same labels, same order, same position numbers. Google ignores breadcrumb schema when the visible markup disagrees in either direction, and the two halves are easy to drift apart when you maintain them separately. Three delimiter styles (chevron, slash, arrow), accessibility-aware HTML output (separators in CSS pseudo-elements so screen readers do not read them aloud), and explicit position integers in the schema.

Link Health Checker

Link Health Checker audits the outbound links on a single page at editorial depth. Anchor-text concentration (flagged at 30 percent — above which a phrase reads as keyword-stuffed), dofollow/nofollow distribution, rel-attribute hygiene (noopener for target=_blank, sponsored for paid links, ugc for user-generated content), and basic URL malformation. Most link audits crawl a whole site and surface volume metrics; this one inspects one page the way a careful editor would. Use before publishing, after a refresh, or whenever you inherit a site and need to verify the most-trafficked pages by hand.

Mobile-Friendly Tester

Mobile-Friendly Tester runs static heuristic checks on pasted HTML — viewport meta, tap-target dimensions, font sizes, layout widths likely to cause horizontal scroll on the smallest mainstream mobile viewport. Not a Lighthouse replacement; Lighthouse runs the page in a real headless browser with measured rendering and JavaScript execution, which static heuristics cannot match. But for catching template-level bugs in the editor before the slow simulation kicks in — missing viewport tag, an explicit 10px font in inline styles, links spaced four pixels apart that fail Google's tap-target rule — static is the right call.

Core Web Vitals Cheatsheet

Core Web Vitals Cheatsheet is a reference plus a classifier — paste a measured value and see the band. INP replaced FID in March 2024 and a surprising number of tutorials still teach the old metric, so the cheatsheet covers what the current three vitals actually measure, the active thresholds, and five-or-six common causes per metric with one-line fix patterns. Useful while reading a Lighthouse report and wanting a quick band check, or as the reference tab you keep open while debugging a regression.

The AI Category Opens

This category has been on the roadmap since v27 and was held until there were three coherent tools to launch with rather than one. The launch trio is intentionally light — no API calls, no transformers.js yet, no model weights to download. Everything stays browser-only and instant. The category sits at three tools, which is below the six-tool threshold for the Suite layout, so the category page is a clean card list rather than a workflow Suite. That changes in v33.

AI Prompt Template Builder

Templates with named variables and a fill-in form. Write `Summarize {{topic}} in {{paragraphs=3}} paragraphs for {{audience}}` once, then run it dozens of times by filling in three short fields. Default values supported via the `{{name=default}}` syntax, templates persist to browser local storage, export and import as JSON. The friction prompt templates solve is real — anyone using an LLM more than once a week ends up keeping a notes file of working prompts, and editing the variables by hand on each run is exactly where errors creep in.

LLM Comparison Table

Manually-curated reference of current Claude, GPT, Gemini, and Llama models. Context window, output cap, knowledge cutoff, input price per million tokens, output price per million tokens, capability tags (vision, code, agentic, long-context, fast, reasoning). Filterable by capability, sortable by any column. The footer date shows the last refresh. No live scraping — model providers do not all publish machine-readable change logs, so a manual curation cycle on a known cadence is more honest than a live table that drifts silently.

System Prompt Library

Nine system prompts for common roles — code reviewer (strict and collaborative variants), research assistant, copy editor (professional and conversational), brainstorm facilitator, debate partner, technical interviewer, explainer for difficult topics. Each entry has been tested across multiple model families and refined over real use. Not a 500-entry mega-library — depth beats count for system prompts, and a library is only as good as the worst entry in it. New entries land when they have earned their place.

Updated SEO Workflow

With 25 tools, the canonical on-page-launch sequence stays roughly the same shape but picks up two new validation passes. The SEO Suite page walks through the full chain. In short:

1. Headline Analyzer and URL Slug Optimizer for title and URL.
2. PAA Generator + Related Keywords Extractor for the brief.
3. Draft.
4. Readability, Density, Heading Check, Snippet Preview, Link Health on the body.
5. Mobile-Friendly Tester on the rendered HTML before deploy.
6. Meta Description Generator then SERP Preview for the snippet.
7. Canonical, plus Hreflang and Robots Meta if applicable.
8. Schema → Schema Validator → Rich Results Tester. The three together cover both Schema.org validity and Google eligibility.
9. Breadcrumb Builder if the page sits inside a category trail.
10. OG Debugger for the share card.

The Core Web Vitals Cheatsheet is the tab you keep open while looking at the Lighthouse report afterwards. Site-level tools — sitemap, robots.txt, internal-link visualizer — stay once-per-site, not per-page.

All 25 SEO tools and 3 AI tools are free, no signup, no quota, no API key. Browse SEO or AI, or jump straight into any of the nine new tools linked above.

Posted May 14, 2026 by Derek Giordano. Part of the UDT SEO Suite and the new AI category.