Skip to main content

3 posts tagged with "Search"

Catalog Search: synonyms, stopwords, typo tolerance, and richer indexing controls

MINOR Synonym Sets for Catalog Search

You can now define groups of equivalent or related terms so shoppers find products regardless of the exact wording they use.

What's New?

  • Manage synonym sets - Create, list, retrieve, update, and delete named synonym sets, scoped to an organization or a store
  • Multi-way and one-way synonyms - Treat terms as fully equivalent (sneakers, trainers, running shoes) or map a root term (fruit) to its synonyms (apple, mango, peach) in one direction only
  • Attach to search profiles - Reference synonym sets from a search profile's synonym_sets attribute so they apply at search time, with locale and symbol-indexing options per item

For details see:

MINOR Stopword Sets for Catalog Search

Define sets of common, low-value words (such as "the", "a", "an", "is") to ignore during search so results stay relevant and free of noise.

What's New?

  • Manage stopword sets - Create, list, retrieve, update, and delete stopword sets, scoped to an organization or a store
  • Per-locale sets - Maintain one stopword set per locale, with a default set for unspecified locales
  • Asynchronous sync - Track each set's sync status (pending_sync, synced, sync_failed) directly in API responses

For details see:

MINOR Handling typographical errors (misspellings) and prefix matching for search profiles

Search profiles now expose fine-grained control over how forgiving and predictive matching is, both at the profile level and per field.

What's New?

  • Typo tolerance - Configure typo correction with number of typos (misspellings), length thresholds, token-dropping behavior, split/join handling, and synonym typo resolution
  • Prefix matching - Treat the last query token as a prefix so results appear as shoppers type (for example, "red app" matches "red apple"), with a profile-level default and per-field overrides
  • Per-field overrides - Override num_typos and prefix on individual fields in your field weight configuration

For details see:

MINOR Stemming and tokenization controls for indexable fields

Indexable Fields can now fine-tune how text is analyzed, improving recall for descriptive content and giving you control over how terms containing special characters are indexed.

What's New?

  • Stemming - Enable Snowball stemming per field so a search for "running" also matches "run" or "runs" on both custom and core fields
  • Token separators and symbols to index - Control how text is split into tokens and which special characters are preserved, at the collection level or per field
  • Core field overrides - Apply stemming, tokenization, and sorting settings to built-in fields like name, description, and sku using core_field_overrides

For details see:

Introducing Catalog Search

MAJOR Catalog Search - Built-in Search for Your Storefront

Elastic Path now includes Catalog Search, a fully integrated search service that provides fast, relevant product discovery for your storefront. Search is built directly into Elastic Path Commerce Cloud and your catalogs, eliminating the need for third-party search vendors and complex integration work.

What's New?

  • Store search enablement - Enable search for your store or organization and go live without configuring a separate search provider.
  • Automatic catalog indexing - When you publish a catalog, products are automatically indexed and become searchable. Multiple catalogs (e.g., live and pre-live) can be indexed simultaneously, with catalog rules and shopper context controlling which catalog a shopper searches.
  • Configurable searchable attributes - Control which product fields drive search results, including product name, description, SKU, and custom extended attributes from your templates.
  • Faceted search and filtering - Shoppers can filter results by categories, attributes, price ranges, and more. Configure which attributes are facetable to power your filtering UI.
  • Sorting - Sort search results by relevance, price, or custom attributes.
  • Typeahead and keyword search - Provide instant suggestions as shoppers type, guiding them toward the right products and categories.
  • Search profiles - Create different search behaviors for different use cases. Weight fields differently (e.g., emphasize size and material for rugs, finish and style for lighting), apply filters, and configure boost rules.
  • Query boosting - Promote or demote products based on categories, tags, or attribute values. Boost seasonal collections, featured products, or sale items to the top of results.
  • InstantSearch.js adapter - Integrate search into your storefront using the popular InstantSearch.js library with our adapter package.

For details see:

Catalog Delta Files for OpenAI Product Feeds

MINOR Catalog Delta Files for AI Product Feeds

Elastic Path now supports using Catalog Delta Files as a direct data source for OpenAI Product Feeds, enabling near real-time synchronization of product changes into AI-powered search, discovery, and agent workflows.

What's New?

  • Delta files as AI ingestion source - Catalog Delta Files can now be used as an official feed input for OpenAI Product Feeds to support AI search, recommendations, and shopper agents.
  • Change-based updates only - Delta files include only created, updated, and deleted products, eliminating the need for full catalog re-ingestion.
  • Automatic generation on publish - Delta files are produced automatically during full and incremental catalog publish events.
  • Optimized for AI pipelines - Reduced ingestion cost, faster update cycles, and improved data freshness for AI-driven experiences.
  • Supports full product lifecycle changes - Includes updates to attributes, availability, pricing references, and product deletions.

For details see:

Ask External AI