From first scan
to a fix that holds.

Rank Sniper reports four disciplines, always in this order: SEO → AEO → GEO → AIO — from a crawler that indexes a product page, to an engine that quotes it, to a model that cites the store, to an agent that acts on the record. AIO here means machine readability — whether the record parses and can be acted on. It is not the umbrella definition that treats AIO as all four disciplines at once, and it is not the off-site trust definition.

Today: read the four pillars once, then the page for whichever pillar your store scores lowest in. Backlinks and link analysis are out of scope by choice, so a low number here is never a link problem.

Go deeper: answer-engine readiness for Shopify · agentic commerce for Shopify

Today: open the free scan and enter a storefront domain — yours or any other Shopify store. It reads only what the store serves to every visitor, so there is nothing to install, no account and no card.

The crawler announces itself as RankSniperBot and reads robots.txt first; if the store disallows it, the scan stops there. Otherwise it fetches /products.json, up to three product pages as raw HTML, /sitemap.xml and /llms.txt. That is the whole list. It never logs in, never writes and never runs JavaScript.

The result lands at its own address, /scan/your-domain, which you can send to a developer or an agency. Opening that address again within 24 hours shows the stored scan instead of crawling twice. Shopify notes that its sitemap cannot be read by crawlers while a store is password-protected, so scan the storefront your customers actually reach.

The composite. The number at the top is the legibility score, rubric v1.0: 9 catalogue checks, each applied to the share of products that pass it, with weights that sum to 95 and are published as they are. That is blended 80% with a structured-data sub-score at 20%, the crawler penalty comes off, and the result is rounded, never below 1. A coverage line beside it says whether the number rests on the full catalogue or a stratified sample, and with what margin.

Four pillar scores. Each pillar score runs out of 100 over its own checks — 29 in all, pillar layer v1.0. They are cut from the same audit and do not add up to the composite. The weakest is flagged “fix first”; AIO carries a transactability sub-score.

Not scored is not failed. When an input could not be fetched — a sitemap that timed out, product pages the edge refused — the check is listed as not scored and left out of that pillar’s denominator (an unscored check). A 404 is an answer and does count; a timeout is not.

The crawler penalty. Each of the 10 named AI agents your robots.txt disallows takes 6 points off the composite, capped at 30 — a capped penalty in the score, and in the world a crawler that does not read the store at all. Googlebot and Bingbot blocks cost the SEO pillar only, never the composite. On Shopify, the rules come from the default robots.txt or a robots.txt.liquid template in your theme.

Structured data is read from raw HTML. Product JSON-LD counts only if it is in the HTML the server sends. A theme or app that injects it with JavaScript can score nothing here while every product is unchanged.

Full report: request it from the result. It arrives as a sign-in link to HQ Merchant Command and lists, pillar by pillar, the failing checks, the penalty in points and the agents your robots.txt turns away.

Go deeper: structured data for Shopify · every check and weight on the rubric · the crawler index

Today: the free scan reports the two layers above, each labelled as what it is. Read them as a measure of what a public scan can reach, not as a full assessment of the store.

When it ships: Rank Preview gives SEO, AEO, GEO and AIO Preview scores and an overall Rank Preview, each with its coverage stated, so a public number never passes for the authenticated assessment. It belongs to the work on one canonical Rank Score, also In development.

Today: you cannot take this step. The app is not on the Shopify App Store, it is installed on no store, and nothing has been written to any merchant’s store. It is set out here so you know what it will ask for.

When it ships: install runs through Shopify’s own OAuth consent screen — no password to share, no staff account to create. The access scopes the app is being built against are read_products, write_products and read_inventory. There is no theme-edit permission, which is why the storefront structured-data app block will be your switch in the theme editor, not ours (Planned). Scopes can be revoked from the Shopify admin at any time, and uninstalling does the same. Shopify is the only billing authority for Shopify merchants. Rank Sniper subscriptions and add-ons will be billed through Shopify when the app is available; nothing is charged on ranksniperhq.com.

Today: the public scan reads /products.json and samples product pages. It cannot see what only the Admin API holds — GTIN or MPN per variant, inventory state, dimensions — so 18 registry checks wait on a connected scan.

When it ships: the connected scan reads the catalogue through the Admin API page by page rather than by sample, and feeds the connected-only checks into the same report.

Three words carry the report. A check is one test with a pass condition and a weight — “SKU coverage”, “Product type set”. A finding is a check your store fails. A fix class says how the fix would be written:

• Automatic. Published through Rank Sniper’s own metafield namespace, structured data or feed. Designed to need no approval, because it never touches a field you wrote.
• Assisted. Changes a native Shopify field — a title, description, tag, image or variant field. Drafted and staged; written only when you approve and apply it.
• Guided. Outside the app’s write scope — a theme template, a file at your domain root. You make the change from exact instructions; the next scan checks it.

Today: the full report names findings and nothing else — no weights, no fix classes. The rubric publishes the fix class of every pillar check, so look each finding up there.

Today: start with the pillar the public result flags weakest, then work its findings from the top — the full report lists them heaviest-weighted first, because the order is the advice. On the composite, the three heaviest catalogue checks are Description depth (16), Structured attributes (15), SKU coverage (12).

Read robots.txt before any of them. At up to 30 points, the crawler penalty is the largest single deduction the composite makes, and unblocking an agent is a one-file change. The order does not wait on the app; the fix pipeline changes who does the work, not what comes first.

Today: you make every fix. Rank Sniper writes nothing to any store. Most of the work lands in Shopify admin fields you already use — product type, tags, vendor, a SKU and a weight on every variant, real option names instead of “Default Title”, and descriptions long enough to state material and dimensions. Theme-level fixes — JSON-LD in the raw HTML, one canonical, one H1 — go to whoever edits your theme; llms.txt goes at the domain root.

As designed: 4 pillar checks are automatic, published in Rank Sniper’s own metafield namespace and structured data; 20 are assisted, drafted and staged, then written only when you apply them; 5 are guided, with exact steps for what sits outside the app’s write scope. Every write is designed to keep its prior value for per-product restore (Planned), shipping before the first automatic fix does.

Go deeper: Shopify product data for AI agents · llms.txt for Shopify · the safe-write contract

Today: re-scan the same domain after a fix. The re-scan option on the result runs a fresh crawl instead of showing the stored one, so the score it returns is measured against the changed store. For structured data, check the page source rather than the browser’s inspector: the scan reads what the server sends, not what a script adds later.

When it ships: post-fix verification re-fetches the delivered HTML raw and as each declared AI user agent, to confirm a change reached what agents receive; rollback restores any written field to its prior value.

Today: nothing runs on a schedule. Re-scan after anything that can strip markup or change crawler access — a theme update or switch, a new app that touches product pages, an edit to robots.txt.liquid, a bulk product import.

When it ships: scheduled re-scans, drift detection and alerts, with score and version history in HQ, so a number that moves can be traced to the scan that produced it.