1. Normalize seller-provided text
Vacora normalizes the seller-provided title, bullets, description, and search terms into structured fields for consistent review.
Public Trust Center
Method, evidence, and limits
Vacora checks seller-provided Amazon US Listing wording for configured risk signals. The deterministic rule engine applies source-backed patterns, and Safe Rewrite is schema-validated and Rechecked before delivery. This page explains what current accepted evidence covers and what it does not cover.
Current accepted evidence snapshot
Marketplace
Amazon US only
Accepted release reference
REL-AMZ-LST-fdb9c29e7666b0c2
Deterministic rules
198 configured rules
Category families
31 category families
Risk segments
16 risk segments
Evidence last verified
2026-07-11
All figures above are an accepted evidence snapshot, not a live guarantee. Coverage may vary by specific product, category, and Listing content.
How a review works
2. Apply global and eligible rules
Global rules apply to every Listing. When a category or risk segment is identified, eligible family and segment-scoped rules are also applied.
3. Show matched wording and explanation
Each finding shows the matched wording, severity, a plain-language explanation, and the source posture behind the rule.
4. Safe Rewrite, validation, and Recheck
For paid Complete Review, Safe Rewrite creates a more conservative English rewrite. The output is schema-validated, then Rechecked against the deterministic audit boundary before delivery.
Coverage matrix
The 31 canonical Amazon US category families with accepted fixture-based deep coverage evidence:
View all 31 category families
- Amazon Devices AccessoriesAMAZON_DEVICES_ACCESSORIES
- AppliancesAPPLIANCES
- Arts Crafts SewingARTS_CRAFTS_SEWING
- AutomotiveAUTOMOTIVE
- BabyBABY
- Beauty Personal CareBEAUTY_PERSONAL_CARE
- BooksBOOKS
- Cds VinylCDS_VINYL
- Cell Phones AccessoriesCELL_PHONES_ACCESSORIES
- Clothing Shoes JewelryCLOTHING_SHOES_JEWELRY
- Collectibles Fine ArtCOLLECTIBLES_FINE_ART
- ComputersCOMPUTERS
- ElectronicsELECTRONICS
- Garden OutdoorGARDEN_OUTDOOR
- Gift CardsGIFT_CARDS
- Grocery Gourmet FoodGROCERY_GOURMET_FOOD
- HandmadeHANDMADE
- Health HouseholdHEALTH_HOUSEHOLD
- Home KitchenHOME_KITCHEN
- Industrial ScientificINDUSTRIAL_SCIENTIFIC
- Kindle StoreKINDLE_STORE
- Luggage Travel GearLUGGAGE_TRAVEL_GEAR
- Movies TvMOVIES_TV
- Musical InstrumentsMUSICAL_INSTRUMENTS
- Office ProductsOFFICE_PRODUCTS
- Pet SuppliesPET_SUPPLIES
- SoftwareSOFTWARE
- Sports OutdoorsSPORTS_OUTDOORS
- Tools Home ImprovementTOOLS_HOME_IMPROVEMENT
- Toys GamesTOYS_GAMES
- Video GamesVIDEO_GAMES
Fixture-based deep coverage
The 16 risk segments with accepted fixture-based deep coverage evidence:
View all 16 risk segments
- Supplement / vitamindietary_supplement
- Cosmetic / skincarecosmetic
- Skin lighteningskin_lightening
- Hair growthhair_growth
- Anti-acneanti_acne
- Oral careoral_care
- Medical-device-like productmedical_device_like
- Topical pain reliefpain_relief_topical
- Disinfectant / pesticide claimdisinfectant_pesticide
- UV sanitizeruv_sanitizer
- Children toy / child safetychildren_toy
- Baby carebaby_care
- Battery / rechargeable devicelithium_battery_device
- Charger / power adaptercharger_power_adapter
- Food with health claimfood_health_claim
- Pet health productpet_health_product
Fixture-based deep coverage
An unknown or unselected category can reduce the review to global rules and inferred compatible scope. Results may be less specific than when a recognized category is selected.
Non-US marketplaces are outside the current production claim. Only Amazon US marketplace Listings are covered by the accepted evidence.
Source posture
Representative, repository-backed sources that inform the deterministic rule set include official publications from:
Amazon Product Detail Page Rules
P4-SRC-AMAZON-PRODUCT-DETAIL-PAGE-RULESAmazon Seller CentralAmazon Seller Central links may require authenticated Seller Central access to view.
Open official source ↗Status: verified_authenticated · Last checked: 2026-07-04
Label Claims for Conventional Foods and Dietary Supplements
P4-SRC-FDA-LABEL-CLAIMSU.S. FDA
Open official source ↗Status: verified_public · Last checked: 2026-07-03
Advertising and Marketing Basics
P4-SRC-FTC-AD-SUBSTANTIATIONU.S. FTC
Open official source ↗Status: verified_public · Last checked: 2026-07-04
Pesticide Labeling Questions and Answers
P4-SRC-EPA-PESTICIDE-LABELINGU.S. EPA
Open official source ↗Status: verified_public · Last checked: 2026-07-03
Regulations, Laws and Standards
P4-SRC-CPSC-REGULATIONSU.S. CPSC
Open official source ↗Status: verified_public · Last checked: 2026-07-03
Public marketplace and community samples are used only for gap discovery and regression input. They are not policy authority.
What the evidence proves — and does not prove
What current evidence supports
- Deterministic known-expression regression for accepted fixtures.
- Scoped rule routing for accepted fixture coverage.
What it does not establish
- Perfect recall of every possible paraphrase or variation.
- Zero false positives across all possible Listing inputs.
- Coverage of every product, packaging claim, certificate, or lab evidence.
- Visual-image compliance review.
- Any specific Amazon marketplace outcome or decision.
Private real-seller validation (P0-T35) remains incomplete and is classified by the approved launch governance as post-launch research, not as a release gate. It is not represented as passed, and public or synthetic evidence does not replace it. Commercial launch still requires separate passing production evidence for technical delivery, privacy, payments, OCR, Report/PDF access, monitoring, and rollback controls.
AI and OCR boundaries
Safe Rewrite
Safe Rewrite is AI-assisted but schema-validated and Rechecked through the same deterministic audit boundary before delivery. Provider quality percentages from the small current evidence set are not published as live performance claims.
OCR Image Text Review
OCR reviews extracted visible text only, up to the existing paid-plan boundary of five images and 5 MB per image. OCR is text extraction — it is not visual compliance review and does not judge the image itself.