Constructor
visit →The search engine reasoning engine powering commerce discovery
Constructor replaces traditional keyword-matching search and category navigation with a behavior-trained ranking engine that optimizes results against revenue and conversion rather than text relevance. The platform handles site search, browse, recommendations, collections, and autosuggest for mid-market and enterprise retailers. Merchandising teams get attribution-aware controls, while data teams get an API-first stack that ingests clickstream signals continuously. Buyers typically evaluate it against Algolia, Bloomreach, and Coveo when replacing legacy Endeca or Solr deployments.
> what it does for ecommerce
- Ranks results against revenue lift, not text-match relevance scores
- Trains continuously on clickstream behavior across search, browse, recommendations
- Unifies search, category, autosuggest, and collections under one model
- Gives merchandisers attribution data tied to every pinning or boost rule
- Deploys via API-first architecture suited to headless commerce stacks