Read-only Shopify access
The Shopify app requests read_products only in v1. It reads catalog fields for the audit and does not request product-write, order, customer, checkout, or payment scopes.
Trust & safety
Built to audit product data without taking control of the merchant store.
FeedSnapshot is designed as a read-only Shopify app. It turns catalog data into a score, product worklist, and exports, while leaving final store changes under the merchant's control.
Launch-safe operating boundaries
No automatic store edits
Reports recommend fixes for titles, descriptions, attributes, identifiers, images, and policy context. Merchants choose what to apply inside Shopify admin.
Shopify billing for App Store
Shopify App Store purchases are approved through Shopify billing. Stripe is only for a separate standalone channel if that channel is enabled later.
Exportable, reviewable reports
Paid plans unlock Markdown and CSV exports so merchants can review, assign, and apply fixes without giving the app write access.
Stored only for product delivery
Stored data is limited to app installs, report results, billing state, monitoring targets, lead requests, email events, and privacy-minimized product-usage events needed to deliver and improve the service.
Uninstall and privacy webhooks
The app includes Shopify uninstall, customers/data_request, customers/redact, and shop/redact webhooks. Uninstall and shop redact delete the installation and linked scans, findings, snapshots, usage events, and email records.
Scoring methodology
Repeatable catalog evidence, not a prediction.
The readiness score summarizes observable catalog completeness. It is not a forecast of Google approval, rankings, AI mentions, traffic, or revenue. The dashboard exposes the score drivers and the rule version used for every scan.
01
Read catalog evidence
The app reads product, variant, identifier, image, category, price, availability, inventory, and content fields through Shopify's read-only product scope.
02
Apply versioned rules
Deterministic checks evaluate the same stored fields against the same rule version. An AI model does not invent the numeric score.
03
Name the affected work
Each paid finding records severity, field, observed evidence, affected product, confidence, source, and recommended correction.
04
Rescan to verify
A later scan compares stable rule and product fingerprints and separates resolved, new, unchanged, and regressed findings.
Signals included in the evidence model
Identity
Title clarity, vendor, product type, taxonomy category, SKU, and GTIN state
Buyer context
Description depth, attributes, FAQ signals, shipping, returns, and trust context
Images
Image presence, known dimensions, images below 500 x 500, and alt-text coverage
Variants
Variant sampling, price, availability, inventory, identifier, and option coverage
Delivery evidence
Catalog coverage, rule version, source links, and before/after rescan differences
What the app stores
Only what is needed to deliver reports and monitoring.
The app is not a customer-data, order-data, or analytics warehouse. Report storage exists so paid buyers can reopen their audit, download exports, and track changes after rescans.
Shop domain and install state
Needed for OAuth, billing state, app access, uninstall handling, and support.
Read-only product catalog fields
Needed to score product-data readiness and generate product-level fix suggestions.
Generated report results
Free product-level data expires after 24 hours. Paid history remains available only while the app stays installed.
Product-usage events
Limited scan, export, upgrade, billing-start, and review-prompt outcomes used to understand activation without storing customer or order data.
Email and support events
Needed for requested launch notes, report delivery, monitoring alerts, and support follow-up.
Billing state
Needed to unlock the correct Shopify plan after Shopify confirms payment or subscription status.
Claims we avoid
- No guaranteed rankings, traffic, AI citations, Google Merchant Center approval, conversions, or revenue.
- No private scraping, private admin crawling, customer-data scanning, or automatic product edits.
- No fake testimonials, fake screenshots, or customer stories in the launch assets.
- No product claims should be published without the merchant verifying their own policies, warranties, ingredients, safety notes, and regulated-category rules.
Operator: AGENTIC SYSTEMS S.R.L.
nacevgabriel@gmail.com