The short answer
We publish 32 free tools at /tools. No account, no card, no email. They are free because none of them costs us anything per use: each one is either our own fetch and parse, or pure JavaScript running in your browser.
That rule is also why there is no free AI Overview checker on the list, however good a demo it would make. Every one of those calls costs us real money at a provider, and an open endpoint that spends money is a bill that anyone can run up.
Below is the full list, grouped by the job it does, with a straight note on which ones are genuinely useful and which are commodity utilities we ship because you would otherwise open a worse one.
AI and answer engine readiness
The five that exist because nothing else answers these questions cheaply.
| Tool | What it does | Verdict |
|---|---|---|
| AI crawler checker | Resolves real robots.txt group precedence for eleven AI crawler tokens, says what each one governs, and flags whether the page needs JavaScript | Run this first |
| Agent readiness check | Scores crawler rules, sitemap, llms.txt, Link headers and the well-known descriptors an agent looks for | Useful |
| Answer readiness score | How quotable a page is inside an AI answer: opening paragraph, question-shaped headings, lists, tables, sources, date | Useful, opinionated |
| AI content extractor | Shows a page as a text-only crawler sees it, navigation and scripts stripped | Useful |
| Structured data for AI | Which structured data an answer engine can use on this page | Useful |
The AI crawler checker is the one worth your time. Most "does this bot allow" tools search robots.txt for a string, which gets the answer wrong whenever a specific group and the wildcard group disagree. Google documents the actual precedence rules in its robots.txt specification, and this implements them: the most specific matching group wins, longest pattern wins inside it, and Allow takes a tie. It also separates training crawlers from search crawlers, which is the single most repeated error in this niche. More on that in does blocking GPTBot remove you from ChatGPT.
llms.txt and robots.txt
| Tool | What it does | Verdict |
|---|---|---|
| Robots.txt tester | Is this path crawlable by this bot, with real precedence | Useful |
| Robots.txt generator | Builds a robots.txt with AI crawler rules, spelling out what each block costs you | Useful |
| llms.txt generator | Reads your sitemap and homepage and drafts an llms.txt | Optional, see below |
| llms.txt validator | Checks an existing file against the spec and fetches its links | Optional, see below |
We ship both llms.txt tools with a caveat printed on the pages themselves. No major AI search product has confirmed that it reads llms.txt. The spec at llmstxt.org is a proposal, not a standard anyone honours, and Google's own AI features documentation asks for crawlable HTML rather than a new file. If you have twenty minutes, spend it on the crawler checker, not on llms.txt. We explain the whole position in llms.txt, what it is and who actually reads it.
Structured data
| Tool | What it does | Verdict |
|---|---|---|
| Schema validator | Reports which types are present and which required fields are missing | Useful |
| Article schema generator | Article, BlogPosting or NewsArticle JSON-LD | Useful |
| Product schema generator | Product JSON-LD with offers and ratings | Useful |
| Breadcrumb schema generator | BreadcrumbList JSON-LD with positions numbered | Commodity |
| FAQ schema generator | FAQPage JSON-LD | Ships with a warning |
| HowTo schema generator | HowTo JSON-LD with ISO 8601 durations | Ships with a warning |
The FAQ and HowTo generators carry a note on the page saying the thing that hurts our own conversion rate: neither earns a Google rich result any more. Google removed the HowTo rich result in September 2023 and retired the FAQ rich result for all sites on 7 May 2026. The markup is still valid entity data and some answer engines still parse it, which is why the tools exist. It will not put stars or steps in your search listing. What to do after FAQ schema was deprecated covers the replacement, and structured data that still matters covers what Google still rewards.
Crawling, indexing and the plumbing
| Tool | What it does | Verdict |
|---|---|---|
| Redirect checker | Traces the whole 301, 302, 307 and 308 chain, flagging loops and https downgrades | Useful |
| Canonical tag checker | Does the canonical survive redirects, point at itself, and name a real page | Useful |
| HTTP header checker | Every response header, scored on transport, safety, caching and crawlability | Useful |
| Sitemap checker | Finds your sitemap the way a crawler does, then checks what is inside it | Useful |
| Broken link checker | Samples up to 25 links on one page and finds the dead ones | Honest sample, not a crawl |
| Title and meta preview | Your page as a Google result and as a shared link | Commodity |
The canonical checker is the underrated one. A canonical that points at a URL which then redirects is a common and quiet way to lose a page, and no browser tab will tell you about it.
Domain, DNS and mail
| Tool | What it does | Verdict |
|---|---|---|
| SPF, DKIM and DMARC checker | The records that decide whether your mail arrives, with your SPF lookup cost counted | Useful |
| DNSSEC checker | Is DNSSEC on, and does the chain still validate. Two different questions | Useful |
| DNS lookup | Every record a resolver can see, with TTLs | Commodity |
| SSL certificate checker | Issuer, expiry, days left, chain trust | Commodity |
| IP lookup and reverse DNS | Hostname to addresses, and whether each points back | Commodity |
Four of these five are commodities and we will say so. The SPF checker earns its place because it counts your record's real DNS lookup cost against the limit of ten, which is the failure nobody notices until mail starts bouncing.
Writing, URLs and text
| Tool | What it does | Verdict |
|---|---|---|
| Title tag generator | Measures real pixel width against Google's roughly 600px desktop cutoff by rendering the text | Useful |
| Word counter and readability | Counts plus a real Flesch-Kincaid grade with the syllable method explained | Commodity |
| Keyword density checker | Term frequency for one, two and three word phrases, stopword list published | Commodity, and mostly obsolete |
| URL slug generator | Titles to slugs, handling accented letters plain Unicode stripping would delete | Commodity |
| HTML to markdown | Clean markdown from pasted HTML, using your browser's own parser | Commodity |
| UTM campaign URL builder | Tagged URLs, catching the capitals and spaces that split your reports | Commodity |
The title tag generator is the only one here we would defend. Every other title tool counts characters, and Google truncates on pixels, so a title full of capital W characters gets cut while the counter says you had room.
Keyword density deserves a sentence of honesty. It has not been a meaningful ranking input for a long time, and we ship it because people search for it. The stopword list is published in full so you can at least see what the number is made of.
The rule behind all of this
A free tool has to cost nothing per use, or it is not free, it is a bill with a delay on it. That constraint decides which tools exist:
- Anything reachable without a session is our own fetch and parse, or pure client-side code. No paid data provider sits behind an open endpoint.
- Every tool that fetches a URL you supply goes through a guarded fetch. One of our routes once read cloud metadata at 169.254.169.254 because a user-supplied URL was passed straight to fetch. That class of bug is why the guard exists.
- Every URL-fetching tool is rate limited by IP.
The cost of that rule is the tool we cannot build. A free AI Overview checker would be the best demo on the site and we will not ship it, because each of those questions costs real money at a provider and an unauthenticated endpoint spending money is a bill anyone can run up. AI visibility scans need an account instead, and the free account gets one scan of ten questions, ever.
Start with the AI crawler checker. If it says a crawler you care about is blocked, fix that before anything else on this page, because the method for measuring AI visibility assumes the answer engine can read you in the first place.