Connect commercetools with Algolia hosted search and discovery to deliver instant autocomplete, smart ranking rules, facets, synonyms, personalization, merchandising analytics, and faster product findability.
• Product projections from Commercetools (master data, variants, attributes, categories) are transformed into Algolia records, typically one record per variant, with stable objectIDs to support updates.
• Delta-based indexing publishes only changed products to Algolia using Commercetools subscriptions or message queues, while full reindex runs handle schema changes and recovery scenarios.
• Locale-specific fields are mapped to Algolia indices or attributes per language, and category trees are flattened into hierarchical facets for navigation and filtering.
• Pricing, inventory availability, and channel-specific data are mapped as separate attributes, with index ownership defined per use case (catalog index vs. availability index).
• Ranking signals are routed into Algolia custom ranking and rules, while synonyms, query rules, and facet configuration live in Algolia’s dashboard or via API.
• Event logs capture indexing outcomes, rejected records, and API errors, with retry logic and dead-letter handling for failed batches.
.png)
We sync products, categories, and key attributes from commercetools into Algolia indices via APIs and event-driven updates, then map facets, ranking signals, and synonyms. You keep commercetools as the system of record, while Algolia handles query-time speed and relevance.
Typically: SKU, name, descriptions, images, category paths, attributes for facets, and any relevance signals you want to boost. We also model variants and availability so results stay usable when catalogs get large.
Yes, we translate commercetools product types, attributes, and taxonomy into Algolia facetable fields and filterable rules. This keeps navigation consistent across PLPs and search without duplicating catalog logic in the UI.
Synonyms and ranking rules live in Algolia, so merchandisers can tune relevance without waiting on backend releases. We set guardrails so rule changes don’t break indexing, filters, or reporting.
Yes, we usually separate indices by locale and store, or use a shared index with locale-aware fields, depending on query patterns. It scales cleanly when each market needs different facets, synonyms, and boosts.




