Running an audit
Most audit problems come down to reachability: our crawler has to fetch your live page the way an AI engine would, so anything that blocks bots or renders content only in JavaScript can leave it empty-handed. The fix is almost always the same — switch to Draft mode and paste your copy, and every content-level check still runs.
The crawler can't access my site / "blocked by bot protection"
Some sites (Cloudflare bot protection, WAFs, some Shopify and managed-WordPress setups) block automated requests. The fastest fix: switch the input to Draft mode and paste your page copy — the content analysis is identical; only the live technical checks and screenshots need the reachable page.
Worth knowing: if our crawler is blocked, several AI crawlers probably are too — the technical audit flags exactly which AI bots your robots.txt and firewall allow, and being reachable is itself a GEO ranking factor.
My site is built with React/Vue/Wix/Shopify and the scan found no content
We fall back to a full headless browser automatically when plain fetching yields nothing, which handles most JavaScript-rendered sites. If the audit still extracts nothing, the site is blocking headless browsers as well — paste your copy in Draft mode and the analysis runs normally.
Why were only some of my pages analyzed?
By design we crawl up to 8 same-domain pages per audit, discovered from your internal links, starting at the URL you enter. That keeps audits fast and focused. For a deep-dive on one specific page, run the audit on that page's URL directly, or paste just that page's copy.
Can you analyze a staging or password-protected site?
Not via URL — we only fetch publicly reachable pages (an AI engine can't see or cite your staging site either, so a staging score would mislead). To pressure-test content before launch, paste the draft copy in Draft mode: every content-level check runs the same.
How long does an audit take? Mine seems stuck.
Typically 60–90 seconds — most of that is live AI-answer testing across your prompt sample. Occasionally an upstream AI provider is slow and a run takes 2–3 minutes. If a run errors out or hangs well beyond that, just run it again: free previews don't cost you anything and a retry usually lands. Persistent failures are almost always an upstream provider incident — trying again later resolves it.
Will the audit affect my website?
No. An audit performs a one-off fetch of at most 8 pages — comparable to one visitor loading your site — and only when a user requests it. We don't crawl recurringly, and your content is never used to train models.
Understanding your results
Your report measures a neutral, logged-out baseline — the answer a brand-new buyer with no history and no personalization would get — across a sample of prompts, which is why it reads as a citation rate rather than a single yes/no and can differ from what you see in your own ChatGPT session. The answers below explain how to read each number.
Your report says I'm cited (or not cited), but when I ask ChatGPT myself I see something different
Both observations are real — they're different measurements. Your ChatGPT session is personalized: memory, chat history, location, and exact phrasing all change the answer you see. We query engines as a neutral, logged-out baseline — the answer a new buyer with no history gets. That baseline is what GEO can actually move.
AI answers are also probabilistic: the same engine asked the same question twice can answer differently. That's why we test a sample of prompts and report a citation rate rather than a single answer.
Why did my score change between runs when I changed nothing?
The structural and technical pillar scores are deterministic — they only move when your page moves. The live citation sample is measured against real AI engines, which are probabilistic and update constantly, so that component naturally fluctuates a few points run-to-run. Watch the trend across runs (signed-in re-runs of the same URL show the delta automatically) rather than any single run.
My citation rate is zero — is the tool broken?
Almost certainly not: a zero is the most common honest baseline for newer or niche brands, because AI engines cite established, corroborated sources. That gap is exactly what the report's optimization plans target. If your brand is well-known and still shows zero, check that the audited URL is your canonical domain, then contact support@findingyou.io — we can inspect the exact engine answers behind your sample.
Another AI-visibility tool gives me a different number — which is right?
Every tool tests a different prompt set, engine mix, and cadence, so numbers differ across tools the same way two SEO tools report different "site health" scores. What matters is methodology transparency and trend over time. Ours is documented end-to-end on the methodology page, and the per-engine table separates live-measured engines from estimates.
Are the test prompts real questions people ask?
Yes — they are objective-driven AI-search questions, not keywords. The prompt universe is generated from your page's actual entities, category, and buying intents and phrased the natural, conversational way real people and AI agents ask ChatGPT, Perplexity, Gemini, and Claude for a service like yours: problem statements, "best X for [my situation]", comparisons, alternatives, pricing and value, integrations, risk, and follow-up refinements — plus prompts where an AI agent is choosing a service to recommend on a user's behalf. Each prompt is labelled by who asks it (a human, an AI agent, or both). Per-prompt AI search volume doesn't exist publicly (no engine publishes it), which is why we measure citation share on a representative sample instead.
Is the testing live or simulated?
Live where it counts: every audit tests your prompt sample against Gemini with real Google Search grounding, plus live Perplexity, ChatGPT (web search), and Claude (web search). The per-engine table in your report marks which rates were measured live; anything estimated is labeled as such. The methodology page documents the distinction openly.
Payments & billing
Checkout runs through Stripe and unlocking is normally instant — the report you were viewing simply opens in full on the spot. Your unlocked report also lives permanently under My Reports, so you never need to pay twice. The answers below cover declines, receipts, duplicate charges, and refunds.
I paid but the report still shows locked
Unlocking is normally instant. After Stripe confirms payment, the report page checks automatically and shows a "payment received — unlocking" notice while it completes — don't pay again. Your report is also always available under My Reports once unlocked.
Still locked 10 minutes after paying? Email support@findingyou.io with the report link or the email you paid with — we can see exactly what happened to your payment and unlock it manually within hours, not days.
When do I get the report after paying?
Immediately — it's the same report you were viewing, with every section unlocked on the spot. You also get an email with a permanent link, and the report stays in My Reports for the life of your account.
I need a receipt or a company invoice (VAT)
Stripe emails a receipt to your checkout email after payment (check spam and Gmail's Updates tab). If you need an invoice with your company name, address, or VAT number on it, email support@findingyou.io with those details and your checkout email — we'll send a proper invoice for the purchase.
What will the charge look like on my card statement?
The charge is processed by Stripe on behalf of FindingYou.io. If you see a charge you don't recognize, email support@findingyou.io with the date and amount before disputing with your bank — we resolve genuine errors with a refund faster than a chargeback does.
My card was declined or the bank verification (3D Secure) failed
Declines come from your bank, not from us — the usual fixes are completing the 3D Secure popup (check it wasn't blocked), trying again, or using a different card. Apple Pay and Google Pay appear automatically in checkout on supported devices and often sidestep verification friction.
I think I was charged twice
A "pending" entry next to a settled charge is usually your bank showing the authorization and the capture — it drops off within a few days. If you genuinely have two settled charges for one report, email support@findingyou.io and we'll refund the duplicate immediately.
I fixed the issues — do I pay again to re-run the audit?
Free previews are unlimited — re-run as often as you like to watch your Health Index move, and signed-in re-runs of the same URL show the score delta against your previous run automatically. Each $29/$49 unlock buys the full report for that snapshot; unlock again only when you want a fresh complete action plan and deliverables for the updated page.
What payment methods do you accept?
All major cards via Stripe Checkout, plus Apple Pay, Google Pay, and Link where your device supports them. We can't take PayPal or bank transfer today. The MCP subscription is billed monthly through Stripe with the same card options.
How do refunds work?
Full terms are in the Refund Policy: in short, we refund in full if a paid report failed to unlock or deliver due to a fault on our side, if you were double-charged, or where consumer law grants the right. Email support@findingyou.io within 14 days with your checkout email.
How do I cancel or change my MCP subscription?
Open the MCP API Keys page and choose "Manage billing" to reach the Stripe billing portal, where you can upgrade, downgrade, or cancel. Cancelling stops future monthly renewals; your plan stays active until the end of the current paid period, so you keep access to the tools you've paid for until then.
What happens to my API keys when the MCP subscription ends?
Your account drops back to the Free plan at the end of the paid period. Your keys and any OAuth connections stay listed but every paid tool stops working (Free can connect and list tools only). Nothing is deleted — resubscribe from the MCP API Keys page and the same keys work again.
Do you refund MCP subscription charges?
A charge for a subscription period that has already started is generally non-refundable, since you had access for it — cancel before the next renewal to avoid the following month. We do refund a renewal charged in error, a duplicate charge, or a period where the service was faulty or unavailable. See the Refund Policy and email support@findingyou.io with your account email.
Account & sign-in
A report belongs to the account you were signed into when you ran it, not to the email you typed at Stripe checkout — so sign in with your original account to find a purchase. Most sign-in trouble is a confirmation email caught by spam or a corporate filter; the checklist below clears nearly every case.
The password-reset or confirmation email never arrived
Work through this checklist — it resolves nearly every case:
- Check spam/junk, and in Gmail also the Promotions and Updates tabs.
- On a company address? Corporate filters (Mimecast, Proofpoint, Outlook Safe Links) often quarantine automated mail — ask IT, or try a personal address.
- Confirm you're checking the exact address you signed up with — a different or mistyped address is the most common cause. (Reset emails only send if an account exists for that address.)
- Wait a few minutes: first-time delivery to strict providers (Outlook/Hotmail) can be delayed.
- Use the "resend" option in the sign-in dialog before requesting again.
Still nothing? Email support@findingyou.io from any address, tell us the account email, and we'll sort it manually.
I paid with a different email than my account — where is my report?
The report belongs to the account that was signed in when you ran the audit — the email you typed at Stripe checkout only receives the payment receipt. Sign in with your original account and open My Reports. Can't remember which account? Email support@findingyou.io with the checkout email and we'll locate the purchase.
How do I change my account email or delete my account?
Email support@findingyou.io from your account address. Deletion is complete and permanent: your account, reports, and the diagnostic logs behind them are removed together.
Report access, email & sharing
Once unlocked, a report lives permanently under My Reports and at its own long, unguessable link — the "report unlocked" email is just a convenience copy you never have to rely on. The answers below cover re-sending it, PDF export, and exactly who can see a shared report.
I never got the "report unlocked" email
You don't need it — the email is a convenience copy. Your unlocked report is always available under My Reports and at its permanent link. To re-send it yourself, open My Reports and click "Email me" on the report — it goes to your account address. Check spam and Gmail's Promotions tab first; that's where it usually is. Need it sent to a different address? Email support@findingyou.io.
How long do I keep access to a purchased report?
For the life of your account. "Permanent link" means exactly that — the report doesn't expire, and losing the email costs you nothing because My Reports always lists every report you've run while signed in.
The PDF won't download or looks broken
PDF generation renders the full report server-side and can take 10–20 seconds for large reports — give it a moment after clicking. If nothing arrives, check your browser's download/pop-up blocker and try once more. Still stuck? Email support@findingyou.io with the report link — we'll generate it on our side and send you the file directly.
How do share links work — who can see what?
Every report has a share link with a long unguessable token. For an unpaid report it shows the gated preview; for an unlocked report it shows the full report. Treat the link like a confidential URL — only people you give it to can find it. Need a link disabled? Email support@findingyou.io and we'll deactivate it.
Agent Readiness audits
Agent Readiness audits score how discoverable and selectable an MCP server or agent is to AI clients. A reachable streamable-HTTP endpoint unlocks the deepest checks (tool metadata, retrieval simulation), but registry presence, packaging, and documentation are scored from public evidence even without one. The answers below cover connection failures and what's scored.
The audit says it couldn't connect to my MCP server
Three distinct causes, and the report tells you which one it hit:
- Unreachable — DNS, timeout, or wrong path. Most remote servers mount at
/mcp or /sse; give us the exact streamable-HTTP endpoint URL. - Wrong transport — stdio-only (local) servers can't be probed over HTTP. Publish a remote endpoint or audit the package/repo target instead.
- OAuth required — we record the auth requirement as a positive compliance signal and skip tool-level checks; everything else is still scored.
A quick way to self-verify: connect with MCP Inspector. If it can't connect either, the issue is on the server side, not the audit.
Can I get a score without a public MCP endpoint?
Yes — registry presence, syndication, documentation, packaging, and trust surfaces are all scored from public evidence (npm/PyPI/GitHub/registries). Tool-metadata checks and retrieval simulation need a reachable endpoint, so providing one upgrades the audit substantially.
For website owners
If you found FindingYouBot in your server logs, it fetched your site once because a person requested an audit of a specific URL — there is no recurring or autonomous crawling, and your content is never used to train models. You can block it permanently any time with a single robots.txt rule, shown below.
I saw your crawler in my server logs — why was my site fetched?
Audits run only when a user requests one for a URL — there is no recurring or autonomous crawling. A single audit fetches at most 8 public pages once, comparable to one visitor session, and content is used solely to generate that report (never for model training). The crawler identifies itself as FindingYouBot — see the bot documentation page for the full user-agent string and the robots.txt opt-out. Questions about a specific fetch? Email support@findingyou.io.
How do I block FindingYou audits of my site?
Add this to your robots.txt — it is checked before any page is fetched:
User-agent: FindingYouBot
Disallow: /
Full details on the bot page.