DijitalPi
TREN
Contact →
Home/Products/PiScan
A DijitalPi product · Free · 3 scans a day

Is your site visible to AI?

To be picked up as a source in answers from ChatGPT, Claude, Perplexity and Google, your site has to be crawlable first, readable next, and quotable last. PiScan measures all three across five layers.

CAN THESE ASSISTANTS ACCESS YOUR SITE?

  • ChatGPT
  • Claude
  • Perplexity
  • Gemini
https://
Loading your allowance… · No sign-up, no card. Takes about 15 seconds.
What sets us apart Unlike tools that only read robots.txt and your sitemap, we send requests carrying crawler identities — measuring whether your server does what your rule file says.

We measured these brands’ sites

126 sites averaged 80/100. Scores come from PiScan’s own measurement; you can run the same scan on your own address.

  • StripeStripe88
  • HepsiburadaHepsiburada78
  • Amazon TRAmazon TR61
  • AkakçeAkakçe65
  • CimriCimri85
  • Arabam.comArabam.com65
  • MigrosMigros72
  • GetirGetir72
  • DeFactoDeFacto61
  • BeymenBeymen65
  • FloFlo65
  • GratisGratis92
  • EbebekEbebek89
  • KoçtaşKoçtaş87
  • Vatan BilgisayarVatan Bilgisayar88
  • MediaMarktMediaMarkt83
  • ArçelikArçelik86
  • VestelVestel88
  • Garanti BBVAGaranti BBVA90
  • İş Bankasıİş Bankası67
  • AkbankAkbank88
  • Yapı KrediYapı Kredi79
  • TurkcellTurkcell91
  • Vodafone TRVodafone TR92
  • Türk TelekomTürk Telekom90
  • THYTHY72
  • AppleApple95
  • MicrosoftMicrosoft69
  • GoogleGoogle71
  • AmazonAmazon62
  • NetflixNetflix78
  • SpotifySpotify55
  • AirbnbAirbnb65
  • ShopifyShopify94
  • NotionNotion86
  • FigmaFigma91
  • AdobeAdobe93
  • NikeNike73
  • ZaraZara72
  • IKEAIKEA86
  • BMWBMW64
  • A101A10197
  • Ziraat BankasıZiraat Bankası57
  • VakıfBankVakıfBank77
  • HalkbankHalkbank90
  • QNBQNB72
  • TEBTEB68
  • iyzicoiyzico93
  • ToggTogg90
  • TofaşTofaş69
  • Renault TRRenault TR88
  • ÜlkerÜlker90
  • PınarPınar52
  • SütaşSütaş97
  • TorkuTorku91
  • Coca-Cola TRCoca-Cola TR65
  • İstikbalİstikbal87
  • BellonaBellona86
  • VivenseVivense85
  • English HomeEnglish Home72
  • Madame CocoMadame Coco90
  • KaracaKaraca86
  • TekzenTekzen70
  • EnuygunEnuygun87
  • ObiletObilet92
  • ETS TurETS Tur94
  • TatilbudurTatilbudur89
  • SeturSetur92
  • AcıbademAcıbadem88
  • Medical ParkMedical Park94
  • HürriyetHürriyet65
  • MilliyetMilliyet65
  • NTVNTV65
  • HabertürkHabertürk90
  • WebrazziWebrazzi90
  • Anadolu AjansıAnadolu Ajansı71
  • MNG KargoMNG Kargo92
  • AygazAygaz80
  • OpetOpet93
  • Petrol OfisiPetrol Ofisi88
  • Koç HoldingKoç Holding82
  • SabancıSabancı55
  • EczacıbaşıEczacıbaşı85
  • ZorluZorlu87
  • KitapyurduKitapyurdu89
  • İdefixİdefix72
  • FlormarFlormar89
  • GitHubGitHub89
  • GitLabGitLab93
  • AtlassianAtlassian87
  • CloudflareCloudflare82
  • VercelVercel88
  • NetlifyNetlify92
  • DigitalOceanDigitalOcean88
  • IBMIBM84
  • IntelIntel72
  • NVIDIANVIDIA84
  • DropboxDropbox86
  • ZoomZoom96
  • SlackSlack96
  • SalesforceSalesforce85
  • HubSpotHubSpot91
  • MailchimpMailchimp88
  • LinkedInLinkedIn65
  • RedditReddit65
  • eBayeBay65
  • WalmartWalmart89
  • TargetTarget88
  • Best BuyBest Buy72
  • UniqloUniqlo64
  • PumaPuma65
  • New BalanceNew Balance89
  • Levi'sLevi's61
  • SonySony69
  • SamsungSamsung82
  • PhilipsPhilips80
  • Mercedes-BenzMercedes-Benz61
  • AudiAudi72
  • VolkswagenVolkswagen86
  • ToyotaToyota65
  • PorschePorsche71
  • VolvoVolvo64
  • Coca-ColaCoca-Cola83
  • McDonald'sMcDonald's93
  • UnileverUnilever87
  • L'OréalL'Oréal56

Brand names and icons are used only to identify which site each measurement belongs to. We have no commercial relationship with these brands, nor their endorsement.

WHY IT'S DIFFERENT

Everyone reads robots.txt. We knock on the door.

AI visibility isn't measured with a checklist. The only way to learn whether a crawler reaches your page is to send a request carrying its identity. Five things set PiScan apart:

01

Requests with crawler identities, not just rule reading

Most tools read robots.txt, say "GPTBot allowed" and stop there. We fetch the same page separately as GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and Googlebot, then compare the responses. Servers that welcome Google while refusing AI identities surface right here. We also state the limit plainly: this test shows identity-sensitive behaviour, it doesn’t prove real crawler access.

02

RAG chunking simulation

The page is split at heading boundaries the way a real answer-engine pipeline would, and every chunk is inspected individually. We don’t say "your page is good" — we say "5 of your 12 chunks open without context".

03

Token economics

Not how many bytes your page weighs, but how many tokens it occupies in a model’s context — and what share of that weight is genuinely content.

04

Schema-to-text consistency

We compare the FAQ questions declared in your structured data against the page body. Markup that doesn’t match isn’t merely useless — it can breach structured-data policies and cost you rich-result eligibility.

05

Scoring that avoids false alarms

When we see a block, we first separate "is this AI-specific or a general bot wall?". If Googlebot is refused too, we don’t report it as a critical finding — because it isn’t one.

MEASUREMENT FRAMEWORK

Five layers, one report card.

The layers are deliberately ordered: if a crawler can't get in, text quality is moot; if text can't be extracted, its structure is moot. The weights follow the same logic.

  1. 01

    Access

    28%

    Can the crawler actually get in?

    robots.txt is parsed separately for 16 AI crawlers; then the same page is requested carrying those crawlers’ identities. That is how we catch the case where the rule file says "come in" while the server looks at the identity and shuts the door.

  2. 02

    Extractability

    22%

    What does a crawler without JavaScript see?

    Most AI crawlers don’t execute JS. We take the raw HTML, strip the script and style scaffolding, and measure the real text left behind. The result: how many KB of your page is content, and how many tokens the page occupies in a model’s context.

  3. 03

    Chunkability

    20%

    Does it survive being split into chunks?

    An answer engine doesn’t quote your page — it quotes your CHUNK. So we perform the same split and ask each chunk two questions: does it open without context, and does it name the subject? A chunk that opens with "This reduces costs" tells a reader almost nothing on its own.

  4. 04

    Semantic layer

    16%

    Does your structured data build a graph?

    Not "is there JSON-LD" but: are the nodes linked by @id, is there an anchor to external authorities via sameAs, and do the declared FAQ questions actually appear on the page — and a mismatch can breach Google’s structured-data policies, which may cost you rich-result eligibility.

  5. 05

    Citability

    14%

    Is there a clean passage to quote?

    We measure headings that mirror real user questions, the short self-contained answer beneath them, verifiable figures, freshness declarations and boilerplate overlap with internal pages. When the same boilerplate dominates every page, retrieval systems struggle to tell one page from another.

HOW IT WORKS

A full report in fifteen seconds, a roadmap on one page.

01

Enter the address

Give a real content page rather than the homepage; the measurement becomes far more meaningful.

02

The scan streams live

About 10 requests go out: the page, robots.txt, llms.txt, the sitemap, 5 crawler identities and 2 sampled internal pages. Total time is capped at 55 seconds.

03

You get a full report

A separate score for each of the five layers, the crawler parity matrix, and a "what to do" note beside every finding.

04

Talk it through if you like

If you want to go through the report with a specialist, you can get in touch in one click. Entirely optional.

Scan Now — Free →

MONTHLY SCAN

Let us scan your site every month and send you the report.

AI crawler behaviour isn't fixed: robots rules change, CDN settings get updated, new pages appear. A one-off scan shows you that day; a monthly scan shows you the change.

  • The same address is scanned every month, with scores and findings compared
  • You are told when crawler access changes — a new block, a new rule
  • Your technical inventory is refreshed: new pixels, removed tags, changed infrastructure
  • Free — you can unsubscribe in one click whenever you like

Your name, email, phone, company and site details are processed by DijitalPi as data controller for the sole purpose of preparing and sending your monthly scan report; they are not shared with third parties. Your record is deleted when you end the subscription. Read the full privacy notice

You can leave any time — every report carries a one-click unsubscribe link.

FAQ

Frequently asked questions about PiScan.

What exactly does PiScan measure?+

It measures how accessible, readable and quotable your site is for AI crawlers. There are five layers: access (can a crawler get in), extractability (how much text survives without JavaScript), chunkability (do retrieval chunks stay meaningful), the semantic layer (does structured data build an entity graph) and citability (can an answer engine find a passage to quote).

Is this an SEO tool?+

No. Classic SEO tools look at ranking signals; PiScan measures readiness for answer engines (ChatGPT, Claude, Perplexity, Google AI Overviews). The robots.txt, sitemap and llms.txt checks are only the entrance — the real measurements are the crawler parity test, raw-HTML efficiency and chunk health.

How many scans can I run per day?+

Three free scans a day, resetting each night at midnight (Europe/Istanbul). Your allowance is tracked server-side: refreshing the page, opening a new tab or closing the browser and coming back does not reset it. Rescanning the same address within 10 minutes serves a cached result and uses no allowance. If a scan fails for technical reasons, the allowance is refunded.

What data do you keep for the allowance?+

Two things: a salted hash of your IP address (the raw IP is never written anywhere) and a random identifier cookie (pis_uid). Both may count as personal data under Türkiye’s Personal Data Protection Law (KVKK) and, where applicable, the GDPR; they are processed solely to count daily scans and prevent abuse, never for advertising or profiling, never shared with third parties, and deleted after three days.

Does the scan change anything on my site?+

No. PiScan only reads: it downloads your page and rule files like an ordinary visitor. No form is submitted and no data is sent. Roughly 10 requests are made per scan, and the same address can only be scanned a few times per minute.

Why do the results differ from what I see in my browser?+

Because we measure the page a crawler sees, not the page you see. PiScan doesn’t run JavaScript — just like GPTBot, ClaudeBot and PerplexityBot. If the gap is large, that gap is precisely the point.

Is the bot test done with real crawlers?+

It is done by imitating the crawler identity (User-Agent). Real crawlers additionally pass IP verification, so where we see a block the real crawler may get through. PiScan states this openly in the report: if Googlebot is refused as well, it does not present that as "AI is blocked" but marks it as a general bot wall.

I got a low score — what now?+

Start with the critical findings; those are the items that directly cut visibility. Every finding carries a "what to do" note. If you like, we can go through the results together and turn them into a prioritised plan.

CONTACT

Let's read the report together.

Phone+46 76 254 29 47 WhatsApp+46 76 254 29 47 Emailinfo@dijitalpi.com FormContact page

ASK AI ABOUT DIJITALPI

Let an AI explain what DijitalPi does.

Opens your chosen assistant with a ready research prompt. It reads the site live and answers.

ChatGPT Claude Gemini Grok Perplexity