Frequently asked questions
Permissions and safety
Does the app change anything in my store?
No. It holds four read-only permissions and has no write access to anything — not your theme, not your products, not your content. This is not a policy we apply, it is what the app is technically capable of.
Instead of editing, it produces a fix list with file:line
references you can export as Markdown and hand to whoever maintains your
theme.
Why does it need each permission?
read_themes— to read your published theme's Liquid, settings and assets so a finding can name the file and line.read_products— to let you pick a product page to track.read_contentandread_online_store_pages— to let you pick a blog article or a custom page to track. Both are needed; without the second, the pickers come back empty.
It asks for nothing touching customers, orders, payments or checkout.
Do you read my theme by crawling my storefront?
No. Theme files come through Shopify's Theme Asset Admin API, which is why the app can quote a line number from a Liquid file rather than guessing from rendered HTML.
AI features
Is my store's data sent to an AI provider?
Only if you switch AI analysis on, and it is off when you install. Before the first run the app shows exactly what would be sent and asks.
What is sent is public storefront text — titles, meta descriptions, headings, body copy — the same words any visitor or search engine reads. Theme source code, store settings and support messages are never sent. Switching it off stops the calls immediately.
What is "AI search readiness"?
Two different things. The first scores how well a page answers the kind of question people put to an assistant rather than to a search box — whether it states an answer plainly, covers the obvious follow-ups, and is structured so a model can quote it.
The second actually asks. You choose questions a customer might ask about your category; the app puts them to an AI assistant with web search on and records whether your store is mentioned, where it ranks among the sources cited, and which competitors appear instead. Questions are re-asked monthly — assistant answers do not turn over week to week, and re-measuring before you have had time to change anything would report noise.
Could the AI be wrong?
Yes, and the app labels those results as suggestions rather than findings. Everything else it reports is a measurement. Check anything AI-generated before acting on it.
Scores and measurements
Why does my score differ from running PageSpeed Insights myself?
The app calls the same Google API you would, so the difference is not in the tool. Lighthouse is a simulation run on a shared machine, and repeated runs of the same page routinely differ by several points — Google says so themselves. Time of day, current server load and which third-party scripts happen to respond slowly all move it.
This is why the app keeps a history per page: a single number is noisy, a trend is not.
Why is there no Core Web Vitals data for my page?
Because Google does not have enough real Chrome traffic to that page to publish it. This is normal for newer or lower-traffic stores and is not a fault. The app says the data is absent rather than showing zeros, which would read as a catastrophic score.
The Lighthouse numbers still work in the meantime — they are a lab simulation, whereas Core Web Vitals are what Google actually ranks on.
Why do phone and desktop numbers differ so much?
They are separate measurements of separate experiences, and Google keeps them separate too. Mobile is simulated on a slower CPU and network, so it is normally the lower of the two and the one worth fixing first.
My audit failed. Why?
Almost always because Google's PageSpeed API returned nothing for both phone and desktop — usually because it was rate-limiting or momentarily overloaded. The app deliberately fails the audit in that case instead of completing with blank sections, so you can retry rather than trust an empty result.
A password-protected store cannot be audited at all: Google has to be able to reach the page.
Plans and billing
What counts as one page check?
One audit of one page. That is the only unit — there is no separate per-page allowance inside a check. Running an audit of your homepage consumes one; running it again next week consumes another.
Does one subscription cover several stores?
No, and it cannot. Shopify bills per installation, so each store has its own subscription. Enterprise is a volume tier for one busy store, not a multi-store licence.
How often can audits run on their own?
Weekly or monthly from Starter, daily from Pro. Every plan with schedules also re-audits automatically when you publish a theme, since that is the change most likely to move the numbers.
Quotas are sized so a tier can actually run every page it lets you track, on the schedule it sells.
How am I charged?
Through Shopify's Billing API, on your Shopify invoice. The app never handles a card and never takes payment directly. Cancel by changing plan in the app or by uninstalling.
Your data
What happens when I uninstall?
Shopify sends a shop redaction request 48 hours later. On receiving it the app deletes the store record and everything attached to it — audits, issues, snapshots, tracked pages, schedules, support requests, sessions — in one cascading delete. Backups roll off after 30 days.
Do you store anything about my customers?
No, and the app could not obtain it: the permissions it holds give no access to customers, orders or checkout. When Shopify sends a customer data request or erasure webhook, the app answers that it holds nothing.
Getting help
How do I reach you?
There is a support form inside the app. Send a question and the answer comes back by email — replies go to the address you give, so the thread stays in your inbox.
You can optionally include your Shopify collaborator request code if the question needs someone to look inside the store. That only lets us ask for access; you still approve the request and choose what it may see.
Will you do the fixes for me?
No — the app is a diagnostic tool and does not sell implementation work or connect you with agencies. What it gives you is a precise fix list your own developer can work from without rediscovering anything.