Pricing
Scan free. Get it fixed for $749.
In one sentence
The CrawlCheck scan, every number in the report and a single YouTube video check are free for any domain, with no account. What costs money is what the free report holds back — the named pages, listings and links behind each count — the record that says when each defect began, and whole YouTube channels checked upload by upload: the fix list implemented on one site for $749 once, Watch $29 a month for 3 domains, Agency $249 for 40 domains with the API and no daily scan cap.
The scan is free and tells you your grade, which stage fails and how many findings you have. The main thing sold here is the work: we implement the fix list on your site and re-measure it, once, for $749 — with thirty days of monitoring included. Keeping the record going afterwards is $29 a month, and agencies get their own plan with the API.
Own the business?
We fix it for you. One site, $749 once, re-measured afterwards, with thirty days of monitoring. On WordPress, one plugin puts the machine-readable fixes live the same day.
Get this fixed — $749Do the SEO?
Every finding named, with the evidence bytes and a dated record: 3 sites for $29 a month, or 40 for $249 with the API and the CI gate.
Compare plansNo card for free scans — no account for free scans — paying never changes a score
The free scan tells you what your site serves to 15 client identities today, and the grade, the score and the stage that fails stay free. A licence adds the four things a single fetch can never produce: which verified crawlers actually arrived at your server, a dated record of what the site was on every day, a proof of it anyone can check without trusting us, and which leads came from an AI assistant — on the same axis as the fixes, so the work and the number sit side by side.
| The loop | What it answers | Where |
|---|---|---|
| Measure | What each of 15 client identities is served, from outside and from your own address — the cause | Free |
| Observe | Which claimed GPTBot, ClaudeBot and PerplexityBot visits were the real crawler, verified against the operator’s published address ranges from inside the site — the arrival | Agency |
| Record | What the site was on every day, kept as a series, each day sealed and anchored to Bitcoin, with a change receipt that verifies on your own machine — the proof | Watch |
| Attribute | Which leads came from ChatGPT, Perplexity, Google, a campaign or direct, credited on first touch and filed beside the record — the outcome | Watch |
| One axis | Fixes and leads on the same dated line, so whether the work moved the number is read, not argued | Watch |
Every scanner on the market does the first row. The other four need something that was already watching before the question was asked, which is why they are what a licence buys and why they cannot be bought after the fact.
What a licence never buys: a better score. Nothing on this page changes a measurement, a finding or a grade. The moment a payment can move a result, every result on the site is worth less — including the ones you would be paying for.
The registry is free on both sides. Every domain this scanner measures is listed on the shelves it earned, and its owner can claim the record by DNS TXT or a meta tag at no charge. A licence moves no shelf — it adds the record kept behind the entry.
Plans
You do not pay per scan. You pay for the domains kept under record. Scan any domain for free. A paid plan stores its measurements over time, reveals the URLs and sources behind aggregate findings, and alerts you when crawler-visible behaviour changes.
Free scan · any domain
$0
The full report, every section and finding, JSON and CSV, a URL that regenerates — and any single YouTube video checked against the anchor fields. No account.
Scan a domainFix list implemented · one site · one-time
$749
The main product. We implement the fix list from your report — machine files, schema and NAP, the findings holding the grade down — then re-measure and hand you the before and after, sealed. Thirty days of monitoring is included in the price; after that it is $29 a month if you want the record to keep running.
Get this fixedWatch · 3 domains · 1 channel
$29/month
The scan told you what you serve. Watch keeps the record of it — what changed and when, sealed daily — puts every list behind every count in your hands, and adds a quote form that says whether the customer came from ChatGPT.
Start watchAgency · 40 domains · 10 channels
$249/month
The one plan for an agency: the API, the scanner without a daily cap, reports that carry your name, and forty domains under record.
Start agencyWhich plan
| If this describes you | Choose |
|---|---|
| You have findings on your site and you want them gone | The fix — $749 |
| You own or operate up to 3 sites and want the findings named and the record kept | Watch |
| You run an agency: client reports with your name on them, the API, and the scanner without a daily cap | Agency |
| You need more than 40 domains on one key | Agency, quoted |
The one-time service
One thing you buy once rather than subscribe to. It is delivered on your site and re-measured afterwards, so done is measured rather than asserted.
Try Watch on one domain for 30 days. No card.
You get a real Watch licence, limited to one domain and 30 days: every list the free report holds back, finding age on every defect, and a change receipt from the second measurement on. When it ends, monitoring stops and the record it built stays readable on every report for that domain — the days after it ends are the ones no tool can back-fill, which is the whole reason a trial of this product is worth having.
The two plans, in full
Watch — $29 / month
The scan told you what you serve. Watch keeps the record of it — what changed and when, sealed daily — puts every list behind every count in your hands, and adds a quote form that says whether the customer came from ChatGPT.
- The record, kept for you: every measurement as a dated series — finding age, first-seen and closed dates, whether a defect came back — each day sealed and anchored to Bitcoin
- Change receipts — the before and after of any two dates, an artifact a third party verifies on their own machine without trusting us
- The CrawlCheck form on every domain: a one-tag quote form that records first- and last-touch source (ChatGPT, Perplexity, Google, a campaign, direct), names AI sources as such, and files each lead beside the record on the same axis as your fixes
- Which pages, across up to 60 declared pages: not-200s, thin, no JSON-LD, canonical elsewhere, noindex, orphans, undeclared and duplicate titles — named, not counted, up to 25 URLs each and 10 for duplicate titles
- The 300-page walk with the link graph, on demand, per domain
- Which directories list you and under what name, and who links to you
- Webhook to Slack, Zapier or your own endpoint the moment a crawler-visible result moves. Silence otherwise
- YouTube: one channel checked whole — every upload against the video anchor fields (chapters, captions, description, topics), tallied and sorted by views — and whether your site carries your videos with VideoObject schema
- 3 domains. We re-measure twice a day so the record has no gaps; that is the mechanism, not the product
Or pay for the year and pay for ten months. Same measurement either way — the longer term is worth more only because the record is.
Agency — $249 / month
The one plan for an agency: the API, the scanner without a daily cap, reports that carry your name, and forty domains under record.
- Scan as much as you like. The licensed lane is outside the public daily allowance entirely — 60 scans per ten minutes on your key, on any domain, whether or not it is one of your forty
- The API on that key: POST /api/scan/licensed, the full record back as JSON, and the CI workflow that fails a pull request above a severity you set (/docs/ci)
- Reports branded as yours — your name, logo and link at the same URL, timestamp proof intact — and a client-safe link with the findings and none of the internal notes
- Identity Guard: which verified AI crawlers actually reached the server, read from the site itself, checked against the operator’s published ranges. Three outcomes, never two
- 40 domains under record, everything Watch names on every one of them
- Compare pages saved to the record: your client against their competitor, dated
- The licensed API lane: POST /api/scan/licensed on your key, 60 scans per ten minutes, outside the anonymous limit — and the CI workflow that fails a pull request above a severity you set (/docs/ci)
- Portfolio view: every domain, every open defect, one page
- Regression alerts across the portfolio — a fixed defect that returns is invisible to any tool without a prior record
- YouTube on 10 channels, with the change record per video: chapters added, captions uploaded, descriptions rewritten, dated
- Priority on new checks, and a direct line when a finding needs arguing about
Sold with a person rather than with a button. Forty domains is a portfolio move — we scope it against your sites first, then invoice or send a checkout link. Leave an address below.
What you get, and what costs money
Every measurement is free, on any domain, as often as you like. The free report gives you the grade, the score, which stage fails and how many findings you have. A licence names them — which findings, the bytes behind each one, which pages, which listings, who links — and the record that says when each defect began. A record cannot be bought after the fact at any price — it exists only if something was already watching.
So the unit is not the scan, and it is not the site. It is how many domains you want the lists and the record for. Whoever owns forty sites has one habit, not forty subscriptions, and should pay for it once.
| What the free report holds back | Free | Watch $29/mo | Agency $249/mo |
|---|---|---|---|
| Domains held under record — the unit everything else is priced by | none | 3 | 40 |
| Which pages — the free report counts twelve whole-site problems and three medians; this names the URLs behind eight of them, up to 25 each and 10 for duplicate titles | — | ✓ | ✓ |
| The 300-page walk with the link graph — up to 300 declared URLs, orphans against your own internal links, undeclared pages, per-page status and text ratio, on demand | — | ✓ | ✓ |
| Which directories list you, and who links to you — listings found by your phone number, the name each one prints, and the pages that link to you | — | ✓ | ✓ |
| The record itself — every measurement kept as a dated series, so you can answer "when did this start", not just "what is wrong today" | — | ✓ | ✓ |
| Change receipt — an artifact your client or your boss can verify themselves, proving what moved between two dates and that neither reading was edited afterwards | — | ✓ | ✓ |
| Finding age — not "your llms.txt is missing" but "missing for 47 days, on every one of 47 measurements". No tool that started looking today can print that sentence | — | ✓ | ✓ |
| Told when it breaks — the moment a crawler-visible result changes it lands in your Slack. Silence when nothing changed | — | ✓ | ✓ |
| Which crawlers actually reached your server — Identity Guard posts your own access log under the licence; each address is verified against the operator’s published range here. The outcome, from inside | — | — | ✓ |
| Reports branded as yours — your name and logo, at the URL you already share, with the timestamp proof intact so a client can still verify it | — | — | ✓ |
| The licensed API lane and the portfolio view — 60 scans per ten minutes on your key, and every domain with every open defect on one page | — | — | ✓ |
| In every scan, free, and staying free | Free | Watch | Agency |
|---|---|---|---|
| Unlimited scans of any domain — no account, no card, no cap | ✓ | ✓ | ✓ |
| AI visibility score, with the three stages behind it | ✓ | ✓ | ✓ |
| Every answer engine probed by name — GPTBot, ClaudeBot, PerplexityBot and the rest, each sent and each answer recorded | ✓ | ✓ | ✓ |
| Timestamp proof, anchored to Bitcoin | ✓ | ✓ | ✓ |
| Embeddable badge that updates with your score | ✓ | ✓ | ✓ |
| Name, phone and address a machine can read | ✓ | ✓ | ✓ |
| Directory coverage — which of ten sources you declare | ✓ | ✓ | ✓ |
| Where you sit against everyone else — your percentile against every site this scanner has measured | ✓ | ✓ | ✓ |
| Whole-site counts — fifteen counts across up to 60 declared pages: not-200s, thin, mostly code, no JSON-LD, canonical elsewhere, noindex, missing and duplicate h1, orphans, undeclared, near-duplicates, duplicate titles, and the medians for words, text share and response time — on every scan | ✓ | ✓ | ✓ |
| Compare against any competitor — two sites, one scan, side by side | ✓ | ✓ | ✓ |
Nothing above the line can be back-dated. Start today and the first change receipt is possible within two days; start in March and March is gone. Nothing a licence unlocks changes a measurement, a finding or a grade — paying buys more measurement, never a better number. Pay for a year and you pay for ten months; the record is worth more the longer it runs, which is the only honest reason to prefer the longer commitment.
One-time work, if you would rather not do it yourself
| Fix everything in your report Every finding corrected on your site and re-measured afterwards. You get a change receipt: the before and after, each sealed and independently verifiable on your own machine, so the fix is evidenced rather than asserted.
| $749 one-time · one local-business site Start the fix — $749 |
The $749 price is one local-business site: the findings in its own report, corrected and re-measured. A four-hundred-page site, several domains at once, or a stack we have not measured yet is not the same job — those are quoted against the report before anything starts. Scan the site first: the report is what a quote is written against.
The machine layer, inside the fix
The scan tells you the files are missing. This is the alternative to reading that list and doing it yourself: an entity map, an agents.md and an llms.txt authored for the specific business — not a template with the name swapped, which is the thing this scanner catches on other people’s sites every day.
Ordered deliberately. The entity map is the piece with the most behind it — it is what lets a retrieval system tell your business apart from one with a similar name, and almost nobody publishes one. agents.md is next. llms.txt is included because it is cheap, some agents fetch it, and a malformed one is worth catching — but we will not sell it as a visibility lever: a study across roughly 300,000 domains found no correlation between publishing one and being cited, and removing the variable made the model more accurate. Our own scan stopped grading it for the same reason. If someone is pricing an llms.txt as an AI-visibility product, ask them for the measurement.
This used to be sold separately at $349 a domain. It is not any more: it is part of the $749 fix, because a fix that corrects the findings and leaves the business unidentifiable to a retrieval system has done half the job.
- Every URL cited in the files is fetched and confirmed to resolve before delivery, including redirects — citing a 301 in a machine-readable instruction file is the sloppiness we criticise in others
- Written from your own published facts: services, hours, service area, contact route. Nothing invented, no price claims, no credentials we cannot verify
- Deployed on your stack and verified live from outside, not from your admin panel
- Five domains or more at once is an Agency conversation, at $249 a domain — ask for a quote
A one-time build, not a subscription. If you never speak to us again the files keep working.
Citation repair — quoted per site
The free scan lists the directories a site declares and the ones it does not. Two things can follow from that list:
- The list, and you do it. Included with an agency licence — which directory, why it matters for that trade, and what the listing has to match to agree with the site.
- The list, and we do it. Quoted per directory once we have seen the evaluation, because the work is different for a business with three inconsistent listings than for one with none. Corrections to existing listings are slower than new submissions, and we say which is which before you agree to anything.
No checkout button here on purpose: this is hand work by a person, and a price set before seeing the site would be a guess. Leave your domain below and the scan does the rest.
Ask for a licence or a service
Prefer an invoice, need a plan that is not on this page, or want one of the one-time services? Leave your details and a person replies — nothing is charged and nothing starts automatically.
What a number on this page can and cannot say
The claims above are only worth the gates that bound them, so the gates are stated here rather than in a footnote.
- A verified-arrival figure is published only once at least five sites are reporting, and a crawler whose operator publishes no address ranges is counted as unverifiable, never as forged.
- A cohort figure — how many were broken at least once — needs at least two measured days and twenty domains; below that the page says so instead of printing a percentage.
- A change receipt proves that two measurements existed, were not altered, and differ. It does not say who caused the change or whether it mattered.
- A lead is attributed to ChatGPT or Perplexity only when the visitor’s own browser carried that source. Siri and Google AI Overviews carry nothing that separates them from direct or Google, and are reported as such.
- The free whole-site summary reads up to 60 declared pages; a licence adds the URL lists, capped at 25 per category.
- Nothing on this page changes a measurement, a finding or a grade.
How a licence is used
A licence is a key, not an account. Open any report URL once with ?lic=<your key>: you are signed in for 12 hours and the key is removed from the address bar, so it never sits in history, a screenshot or a Referer. To show a report to someone else, send a share link, not the key: POST /api/share {"rid":"<report id>","days":30} with the key in x-crawlcheck-key returns a link that opens that one report as licensed for up to 90 days, and DELETE /api/share?s=<token> revokes it; a signed-in report page has a button that does the same. Give the key to the monitoring form and the webhook starts. Nothing to install.