ExtraltExtralt
Beta
products/03/ 04

Extend

Build product relationships across the web.

What is Extend?

Extend turns enriched Items into final analytical tables. It handles cross-seller product matching, connects the same exact variant across different sellers, and publishes Listings, Offers, Reviews, Stores, and Variants for analysis.

For a deeper breakdown, read the guide to product matching for ecommerce.

Why Extend?

Individual product listings only tell part of the story. Real market intelligence requires understanding how products relate to each other: where else is this exact product sold, at what price, and with which sellers, sizes, reviews, and availability?

Extend answers these questions by building relationships between listings and variants in your extracted data, creating a connected market dataset you can query and analyze.

How it works

Extend builds the final identity layer for product analytics:

Exact variant matching

Variants from different stores are matched by identifiers (GTIN, MPN, ASIN) and embedding similarity, creating deduplicated final variants for cross-seller analysis.

Alternate listings

Same product, different sellers. Find everywhere a product is sold across the web, with pricing and availability from each source.

Final offer history

Offers are published as append-only observations tied to final Listings and Variants, so price and availability history stay queryable without rewriting facts.

Current-state dimensions

Listings, Variants, Stores, and Reviews are maintained as latest-state analytical tables while historical observations remain append-only.

Features

  • Embedding generation for variant matching
  • Cross-seller product matching
  • Listing-to-variant mappings
  • Final offer, review, and store publication
  • Cross-seller price history queries
  • Deterministic conflict checks around identifiers and options

Use cases

Price comparison

Find the same product across sellers to compare prices, identify the best deals, and track price differences.

Recommendation engines

Build downstream recommendations on a cleaner identity layer: matched variants, alternate listings, offer history, and normalized review state.

See Extend in action

Pricing

Free

Extend requires enriched data. Once enriched, building relationships is free.