Every directory promises you'll be seen. None of them can prove it.
Submit your product and your listing page carries a field no other directory has: your measured AI citation status — whether AI retrieval can reach you, whether it actually cites you, and where you get stuck.
listing is free · no account · no card
Three pages we measured. Read the middle one twice.
592 words, and it won every single run
592 words, 18 headings, an average sentence of 8.9 words. Clean structured data, a 170-character meta description, retrieval crawlers allowed in. It is not a big site — it is a page a machine can lift an answer out of.
JSON-LD: WebApplication, Offer, BreadcrumbList, ListItem
meta description: 170 characters
OAI-SearchBot: allowed
PerplexityBot: allowed
It was the number one search result. It was cited zero times.
This is the page a person would have clicked first: rank #1 for the query we built the pool from. 6,105 words and 32 headings — more than three times the length of the page that won its own query. Cited 0 out of 5. Ranking and being cited are two different contests, and only one of them decides whether AI ever says your name.
JSON-LD: none
meta description: none
summary block: none
search rank: #1
{ "@context": "https://schema.org",
"@type": "SoftwareApplication",
"name": "Your product",
"description": "one sentence a machine can quote" }
We are not claiming this alone flips a citation. It is what every page that did get cited here had, and what this one did not — so it is where we start, and then we re-test.
The crawler is allowed in, and there is nothing to read
OAI-SearchBot is allowed, so the page enters the pool. But the server sends 1 word of readable text to anything that is not running JavaScript. The crawler arrives, reads nothing, and moves on. Allowing the crawler is necessary, not sufficient.
$ curl -A "OAI-SearchBot" dokieai.com (1 word of readable text) $ curl -sI dokieai.com/llms.txt | head -1 HTTP/2 200 ← and 255 KB of SPA fallback
Both answers look fine. Neither one is. The page returns 200 and says one word; the llms.txt returns 200 and is a JavaScript fallback page, not an llms.txt. This is why we check content, not status codes — and why a green checkmark from another tool means very little.
One robots.txt, two different jobs. Blocking training crawlers such as CCBot affects model training. Blocking retrieval crawlers such as OAI-SearchBot or PerplexityBot removes your page from the pool of sources AI is allowed to fetch and quote. Most sites treat the two as one decision. They are not.
What a listing page carries
The product itself
Name, one-line description, category and a link straight to your site. The link is a normal dofollow link — it is the entry ticket, not the product.
EVERY DIRECTORY HAS THISYour AI citation status
Can retrieval reach you. Did it cite you, and how many times out of how many runs. Where you get stuck, named step by step. Re-measured every month.
NOBODY ELSE HAS THIS See one →Structured data built for it
Every listing ships JSON-LD in which the citation numbers are named, typed and dated — carried as additionalProperty values, not dropped into a generic Product block. A machine is told what the numbers are.
OURSFree gives you the measurement. Paid does the work.
Everything you can do yourself
- Your listing page, live
- Crawler reachability check
- One citation test — when it opens
- Every fix as copy-ready code
The part you don't want to do
- Re-measured every month
- Before/after comparison, kept over time
- We make the fixes for you
- Priority review and a labelled sponsored slot
we do not promise a citation number
What we promise, and what we don't. We promise that we test every month, that we report what we find — including a zero — and that we make the fixes. We do not promise your citation count will rise. It depends on your content, the platform algorithms and your competitors, and we only measure; we do not control. A report that says zero is a complete delivery, not a failure.
Where this is right now
The measurement is real and reproducible. 48 sources, 5 prompts, 6 runs each (28 of 30 usable), every source's status published. The methodology page shows the exact conditions behind every figure, so you can reproduce any of them or show us where one is wrong.
Submissions go through their own page. Send the domain and the query you think you should be cited for; we run it through the same method as every listing here and publish what comes out.
Until then, the free tools on this site work today and need nothing from us. Run the readability check, generate the fixes, keep them.
Questions about this directory
How is this different from the other directories I already submitted to?
The others sell you a position: you pay, you get a page and a link. We sell you a measurement: after you are listed, we test whether AI retrieval can reach your page and whether it actually cites you, and we publish that status on your listing.
Do you promise my citation count will go up?
No. We cannot control your content, the platform algorithms or your competitors, so we do not promise a number. What we promise is that we test every month, that we report what we find including a zero, and that we make the fixes. The number is a finding, not a guarantee.
Does paying change where I rank in your listings?
No. Sponsored placements are labelled and carry rel=sponsored, and they never enter a ranking that is ordered by citation count. Rankings are produced by measurement, and payment cannot move them.
Can I get a listing without paying?
Yes. Listing is free. You get the listing page, the crawler check and one citation test, once that tool opens. What is paid is the monthly re-test and the work of fixing what the test finds.