Features — Shopigent AI GEO

Give AI Systems Better Product Data

AI systems can interpret a catalog more reliably when product names, prices, availability, URLs, and categories are exposed in consistent formats. Shopigent AI GEO adds structured storefront data and measurement tools without promising rankings or recommendations.

🧩 Storefront JSON-LD

The problem: Product information is spread across page text, theme markup, and Shopify data.

The solution: Generate schema.org JSON-LD, store it in a product metafield, and inject it through the theme app extension.

Schema data	Delivery
Organization	Site-wide through the theme extension
WebSite + SearchAction	Site-wide through the theme extension
Product + nested Offer	Product pages; fallback Product data if no synced metafield exists

The merchant must enable the theme app block or embed for storefront delivery. Current catalog sync does not load review ratings or Shopify taxonomy categories. Settings does not expose FAQ, shipping, or return-policy schema configuration; these are not shipped merchant features.

🔍 Visibility Measurement

Scheduled and manual domain-citation checks: Growth and Pro include up to 4 scheduled runs per UTC month, plus 4 manual runs on Growth or 10 on Pro. Runs send up to 8 buyer questions from product titles through the selected, available engines: OpenAI web search and Perplexity Agent on Growth; Pro can also select Grok (xAI) web search; Claude (Anthropic) web search is available for manual runs only. Only provider-attached citations of your store domain enter the citation rate; retrieved results and text-only matches stay separate. This is not a chat ranking.

Weekly tracking is on by default for paid plans and can be turned off in Settings. A queued run still finishes. Scheduled runs use a separate counter: no catch-up and no fifth run in a five-Monday month. The schedule is idle until privacy deployment and owner approval.

Legacy web-search simulation (disabled): the DuckDuckGo-based simulation matched product titles, not your store, so it never measured AI visibility. It is disabled while measurement is rebuilt on real domain citations; past runs are kept only as history.

Plan	Manual citation runs/month	Scheduled citation runs/UTC month
Free	0	0
Growth	4	4
Pro	10	4

✍️ Advisory Description Optimizer

All plans: heuristic score across four practical dimensions, with recommendations.
Growth and Pro: optional GPT-4o-mini rewrite.
Merchant control: generated text remains advisory until the merchant explicitly saves it.

🔌 MCP Storefront — 8 Tools

The authenticated MCP endpoint exposes:

search_products
get_product_details
compare_products
get_promotions
get_store_info
get_popular_products
get_product_recommendations
get_category_tree

Monthly limits are 500 calls on Free, 10,000 on Growth, and 50,000 on Pro. The embedded Sandbox lets merchants test the same tools before configuring an external MCP client.

🏆 Competitive Comparison

Growth and Pro can compare a store with a supplied competitor URL through a manual provider measurement. A comparison uses one real-AI quota unit. Failed or unavailable providers do not create a successful result.

✓ Live Schema Validation

Fetch a product page and verify that JSON-LD appears in the delivered HTML. For password-protected development or preview storefronts, the merchant may optionally supply a password for a single check. It is sent only to that storefront, never stored in the database, logs, caches, or results, and cleared from the input on submission. Without a password, the check uses the public page. An unlocked check does not prove that public crawlers can access a locked store.

📄 Shopify-Native Agent Discovery

Shopify serves /agents.md natively on the storefront. Shopigent AI GEO does not publish or manage that file, and Settings has no llms.txt preview. For authenticated catalog access, connect a compatible client to the app's MCP endpoint.

Real provider measurement is disabled for test installs, including development/reviewer stores with test subscriptions. This includes manual runs, scheduled runs, and competitive measurement, even on a paid test plan.

📊 Reports and CSV

Review schema health, crawl history, and downloadable CSV data inside the app. Reports read domain-citation measurement runs only (citation rate, cited answers, seen-in-search-results); there is no legacy fallback.