Buy Velou if your catalog is fashion or lifestyle and the missing data is visible in your own photos and copy — its vertical model is genuinely good there. Choose Anglera when specs live in supplier documents and values need citations.
Both claim to enrich product data. This page is about where that claim stops.
A first pass is the easy part: this comparison is about day 31 — when a supplier swaps a fabric, a shopper asks for a number no photograph contains, and someone has to decide whether last quarter's generated value is still true.
Mine every spec from every source.
Every value traced to a document you can open. The catalog is only as honest as what it was built from.
Aim the catalog at the buyer who actually buys.
Grounded data still loses if it answers questions nobody asked. Alignment is what turns specs into conversion.
Product data is a practice, not a project.
Markets move, suppliers reissue, buyers change what they ask for. A catalog that is right in March is wrong by August unless something is watching.
Scored against public documentation. Grouped by the three acts — so you can see which ones Velou leaves on your desk.
Reads product images and existing text; no document mining found
PDFs, spec tables, drawings, manuals, images, sites
Vertical model adds new metafields; not derived from market signals
Proposes fields your schema never had
Generates standardized tags; no versioned governed pick lists documented
Normalizes and governs allowed values, versioned
Variant clustering and product graph; channel category mapping undocumented
Auto-classifies; channel and marketplace mapping
No value-level citation or lineage in docs or UI
Every value cites its source doc and page
Vertical-tuned for retail; no B2B/B2C persona variants found
B2B specifier and B2C shopper enriched differently
Tracking API captures searches and clicks; enrichment loopback undocumented
Reviews, search, competitor rails, social — fed back
SEO descriptions, FAQs, alt text, fit and styling notes
Original copy per persona and channel
Analyzes product images; no image generation or editing
Generates studio-grade imagery for photoless SKUs
Claims self-optimizing catalogs; Commerce-1 re-runs are operator-triggered
Re-enriches on its own after go-live
AI Readiness Audit scores catalog against agent requirements
Scored against your standards; nothing publishes below bar
Writes Shopify metafields directly; pushes to channels and feeds
Writes back to PIM, ERP, warehouse, commerce
REST indexing and search APIs plus webhooks; no MCP server found
API, webhooks, and MCP servers
Software; team reviews drafts and publishes to storefront
Anglera owns the work; review is a guardrail
“Buyer signals” is the emptiest phrase in this category, so here is the literal thing. Each of these is an observation from a live market, the gap it exposes, and the field that gets created as a result.
Four reviews on one counter stool, same complaint: "ordered a pair, the seat sits 26 in off the floor and our island overhang is 25 in — knees don't fit." The listing says "counter height" and the hero shot is styled against a table.
"Counter height" is a class you can infer from a photo. The number that decides the purchase — seat height, and the clearance it leaves under an overhang — lives in the supplier's dimension sheet, not in the image or the copy. Until the number is a field, the class is a guess wearing a label.
seat_height_in — numeric, sourced from the supplier dimension sheet; height_class derived from it, never the other way roundDining (17–19 in) | Counter (24–27 in) | Bar (29–32 in) | Extra-tall (33–36 in) | AdjustableRecurring queries with nothing behind them: "gold filled not plated", "nickel free hoops", "14k solid". The listings they should have matched say "gold-tone finish" and "18k gold look".
Composition is being described in the copy rather than governed as a field, so solid karat, gold-filled, vermeil and plated brass all collapse into the same marketing adjective — and nickel content, which is the reason a repeat buyer is searching at all, is never stated anywhere in the record.
metal_composition (enum) plus nickel_content_declared (enum, taken from the supplier's material declaration, not inferred)metal_composition: Solid 10k | Solid 14k | Solid 18k | Gold-filled (1/20 14k) | Vermeil (925 sterling base, ≥2.5 µm gold) | Gold-plated (brass base) | Gold-tone (base metal unspecified). nickel_content_declared: Declared nickel-free | Contains nickel | Not declaredThe outdoor cushion tech pack reads: "Face: 100% solution-dyed acrylic. Lightfastness 1,500 hrs (AATCC 16.3). Water repellency: 80 spray rating (AATCC 22)." The live page says "fade-resistant, water-repellent fabric."
The adjectives can be written from the description; the numbers cannot. They exist only in a PDF nobody parsed — and they are what answers the Phoenix-patio question, survives a warranty dispute, and lets a shopper choose between two cushions that both say "fade-resistant."
lightfastness_hours (integer) with test_method recorded beside it, plus face_fiber_content (enum) — extracted from the tech pack, with the source page kept as provenanceface_fiber_content: Solution-dyed acrylic | Solution-dyed olefin | Piece-dyed polyester | Cotton canvas | Unspecified. test_method: AATCC 16.3 | AATCC 16.1 | ISO 105-B02 | Not statedEnrichment models read what's in front of them. Velou's stated method is to analyze raw text and imagery to generate missing attributes and tags (velou.com) — a real answer to a thin catalog, and a first pass is where the work starts, not where it ends. A supplier swaps the fabric on a colorway and the description never changes, so whatever was inferred from that description stays unchanged with it. A merchant hand-corrects a seat height; something has to remember that decision, or the next pass relitigates it. Six months in, if nobody can say which values came from a spec sheet and which were generated, nobody publishes a filter on them. Velou's blog draws the line itself: "waterproof" and "water penetration resistance: class 2 (EN 343:2019)" are not the same claim. Completeness is a rate you hold, not a number you hit.
Values are normalized into a governed, versioned set of allowed values — so a filter works, and keeps working after the next import.
3/4 in0.75"3/4"19mm3/4 inchDN200.75 in (DN20)Six suppliers, six spellings, one physical size. Filters only work once they agree.
BlkblackBLACK MATTEMatte BlkRAL 9005Black — MatteFree text makes a colour filter useless. A governed value makes it a facet.
SS316316 StainlessStainless Steel 316A4 StainlessStainless Steel — 316 / A4Same alloy, four vocabularies, plus a trade name. Buyers search all of them.
Filling the fields you defined has an invisible ceiling: a catalog can hit 100% complete and still miss the attribute that loses the sale, because completeness is measured against a schema someone drew years ago. Schema Foundry reads competitor listings, buyer searches, review complaints and your supplier docs, and proposes the fields you never defined — which is where Velou stops.
How Schema Foundry works
Velou is a New York-based product data enrichment platform, founded in 2018, built around Commerce-1 — a retail-specific model it says was trained on 4 trillion tokens of retail text and imagery and refined with reinforcement learning from merchandisers. It reads existing product descriptions and product images to generate structured attributes, standardized tags, variant clusters, SEO copy, alt text, and FAQs, then writes them back into the customer's platform (Shopify metafields, feeds, channels). It also ships search and discovery APIs — search, suggestions, similarity, and Complete the Look — reflecting its origins as an AI site-search and catalog copilot product.
Pricing: Partially public. The Shopify app lists a free trial (10 products), Essentials at $99/month (500 products, 75K monthly visitors), and Growth at $250/month (3,000 products, 250K monthly visitors) — scaling on both SKU count and traffic. Beyond those tiers nothing is published: enterprise and non-Shopify deployments are quote-only via demo request, with no rate card or implementation fees disclosed.
Velou websiteShopify fashion, jewelry, and home brands wanting fast, accurate attribute and metafield fill from existing photos and descriptions — plus site search, similarity, and Complete the Look in one vendor.
We'd rather tell you here than in month three of an implementation.
Capability verdicts reviewed against Velou's public documentation on July 14, 2026. Vendors ship quickly — if something here is out of date, tell us and we'll correct it.
Bring one category and your supplier files. In 30 minutes you'll see it enriched — complete, structured, and consistent enough to launch on — plus the attributes your schema didn't have yet.
Book a demo