45 files, metadata, headers and metafields AI agents read on your store. Verity Score keeps them in sync.
One source in your Shopify admin, 45 derived formats, auto-synced on every product or collection change. ChatGPT, Perplexity, Gemini and Claude always read the same truth.
Get early access
30 days free · personal onboarding
Matrix: 45 files, metadata, headers and metafields, 9 families
No AI agent reads the same files. Full coverage demands nine complementary families, read by different engines.
Family H : HTTP headers: X-Robots-Tag, Content-Language, Cache-Control, Link rel="alternate" and Vary. AI agents read these 5 HTTP headers before parsing HTML to decide whether to index, cache and route locale.
Family I : Vertical-aware product metafields: 12 new Shopify signals (INCI, ingredients list, dosage instructions, allergens, size chart, care instructions, fabric composition, country of origin, certifications, safety warnings, expert endorsement, compatibility). These metafields only surface for the verticals they apply to: a beauty store sees INCI and allergens, a jeweller sees size chart and certifications, a supplements store sees dosage and ingredients. Applicability is computed server-side via the _verticalApplicability matrix : leather goods are never asked for INCI.
| Family | Files | Read by | Why it matters |
|---|---|---|---|
| Discovery and crawlers | robots.txt, sitemap.xml, ai.txt | All AI crawlers | Grants the crawl permit and the catalog map. Without these, no agent knows what to read. |
| Well-known endpoints | agent-card.json, security.txt, llms.txt, UCP manifest, ACP manifest | ChatGPT Search, Claude, A2A agents | Store identity card. First touchpoint read by A2A agents and OpenAI crawlers. |
| Agentic commerce feeds | Product feed JSON, Offer feed, Atom feed | ChatGPT Shopping, Perplexity Buy, Google AI Mode | Format consumed by recommendation and agentic-buying pipelines. |
| Inline JSON-LD | Product, Offer, AggregateRating, Organization, BreadcrumbList | Google AI Overviews, Perplexity, Bing Copilot | Structured data read on each PDP. Verified source for AI citations. |
| Schema commerce subtypes | MerchantReturnPolicy, OfferShippingDetails, ItemList | Google Shopping Graph, AI Mode | Reduces agent questions on return policy, shipping and stock. |
| Multi-format syndication | Markdown for Agents (.md), RSS, hreflang alternates | Plain-text agents, AI translators | Compact format for fast ingestion. Multi-locale mandatory for non-English markets. |
Why 45 signals and not just one
Maintaining one file (llms.txt, sitemap, or JSON-LD) leaves 80% of AI agents without signal. Coverage demands the full matrix.
| Approach | Files maintained | Single source | Auto webhook update | Multi-locale |
|---|---|---|---|---|
| Traditional SEO | 1 (sitemap) | , | Partial | , |
| Early GEO tools | 1 (llms.txt) | , | , | , |
| ChatGPT-only tools | 1 (JSON-LD) | , | Partial | , |
| Verity Score | 33 | Yes | Yes | Yes |
How it works
Three steps. Zero manual file editing.
- 1
Connect Shopify in one click
Standard OAuth. Verity reads your catalog, collections, metafields and policies. No modification at this stage.
- 2
Audit the 45 files, metadata, headers and metafields (coverage in percentage)
Each file is analyzed: present, absent, stale, inconsistent or unknown. You get a detailed per-file report with observed sources and Shopify-native recommendations.
- 3
Auto-maintain via Theme Assets, Metafields and App Proxy
Verity writes missing files, fixes stale ones, and re-syncs on every product or collection webhook. No theme override, no file overwritten without authorization.
Included in your plan
The file audit is free on every plan. Automatic maintenance and check frequency depend on the plan you choose.
See pricing →Frequently asked questions
Which files do you maintain?
45 signals in 9 families: discovery and crawlers (robots.txt, sitemap, ai.txt), well-known endpoints (agent-card.json, security.txt, llms.txt, UCP manifest, ACP manifest), agentic commerce feeds, inline Product and Offer JSON-LD, Schema commerce subtypes (MerchantReturnPolicy, OfferShippingDetails, ItemList), Markdown for Agents alternates.
Why 33 and not just llms.txt?
No AI agent reads the same file. ChatGPT Search prefers agent-card.json and JSON-LD. Perplexity requires sitemap and schema. Gemini consumes UCP and Markdown. Maintaining one file leaves four agents out of five without signal. Full coverage demands all 45 files, metadata, headers and metafields, synced from the same source.
How does Verity avoid conflicts with my Shopify theme?
Three non-destructive surfaces: Theme Assets in scoped read-write mode, product and shop Metafields for structural content, App Proxy to serve /.well-known/ endpoints without touching the theme. No existing Liquid template is modified. Every patch is reversible via the audit timeline.
What happens if I cancel my subscription?
Already deployed files stay in place: Theme Assets, Metafields and App Proxy are not deleted. Automatic webhook updates stop. You keep the final versions of the 45 files, metadata, headers and metafields, but they no longer reflect future catalog changes.
Do competitors do the same thing?
No. SEO tools maintain the sitemap. Early GEO tools maintain llms.txt alone. ChatGPT-only tools generate Product JSON-LD. Verity Score is the only one maintaining all 45 files, metadata, headers and metafields from a single source, multi-locale coherent, on product and collection webhooks.
Ready to see your 45 signals?
Free audit in 60 seconds. Per-file coverage, observed sources, Shopify-native recommendations.
Get early access
30 days free · personal onboarding