What they are
Three free checks anyone can run on any public website at chatti.sh/tools, without an account:
- Broken links: links that end in an error, a redirect loop or a chain of redirects, grouped by the page they're on.
- Sitemap & robots: whether your sitemap exists and is valid, which linked pages it leaves out, robots.txt rules that block pages you link to, and noindex or canonical tags that drop pages from search.
- AI-readiness score: a score from 0 to 100 for how well AI assistants can read and answer from your site, with every point explained.
They are the same checks Chattish runs before Wilma goes live on a site, because the things that confuse search engines also stop a site guide from finding answers.
How a check works
When you paste an address, our crawler (ChattishBot) reads your site the way it does for Map any site:
- It reads
robots.txtfirst and never fetches anything it disallows. - It reads your sitemap, if there is one.
- It reads up to 25 pages, starting from the homepage and spreading across your site's sections, about three pages a second.
- It checks links on those pages that it didn't read itself with a single lightweight request each (up to 30 on your site), plus a sample of up to 50 links to other sites, at most two per site.
One check runs all three tools, so a result page links to the other two for the same site at no extra cost. Results are kept for 7 days at a link you can share, and a site is checked at most once a day: running a tool again within 24 hours shows the same result.
The AI-readiness score
The score adds up seven factors. Each one says what we saw and lists example pages.
| Factor | Points | What earns them |
|---|---|---|
| Content readable without JavaScript | 20 | Text is in the HTML the server sends. Many AI crawlers don't run JavaScript. |
| Clear headings | 15 | One main heading per page, and subheadings on pages over 300 words. |
| FAQ and structured data | 15 | schema.org markup for your business, markup on most pages, and an FAQ page. |
| Contact, hours, pricing and location | 20 | Each fact is stated in plain text on a page we read. |
| Pages with enough to say | 10 | Pages with at least 150 words. |
| Unique page titles | 10 | Every page has its own title. |
| AI crawlers allowed | 10 | robots.txt lets GPTBot, ClaudeBot, PerplexityBot and similar crawlers in. |
Not every factor applies to every site: a web app may have no opening hours. Treat the score as a checklist, not a grade from a search engine.
Questions your site can't answer
Under the score we list common visitor questions (price, hours, contact, location, refunds, booking, timing) that none of the pages we read answers. We use the same rule as the gap report: a page has to contain a sentence on the topic and the kind of answer the question asks for, such as a price for "How much does it cost?".
- No page: no page we read mentions the topic.
- Partly covered: the topic comes up, but without a concrete answer.
Limits
- It's a preview. Sites with more than 25 pages can have problems we don't reach; after you sign up, Chattish reads every page your plan allows.
- Some sites refuse automated requests (status 401, 403 or 429). We don't count those links as broken; the result says how many we couldn't verify.
- If your site blocks bots entirely, the check stops and says so. You can still map it from your browser after signing up.
- Each network can start a few new checks per hour, shared with Map any site. Results that already exist are always free to open.
Results about your site
A result describes a site as our crawler saw it; it isn't endorsed by that site. Result pages for sites that haven't been claimed in Chattish tell search engines not to index them. If a result is being used to mislead, use the Report abuse link on the page or write to hello@chatti.sh.