SuggestAPISuggestAPIDocumentation
Contact usOpen app
Features

Search Capabilities

Explore SuggestAPI hybrid ranking, semantic and lexical search, typeahead, filtering, query enrichment, personalization, and result controls.

Copy page
Open in ChatGPTOpen with page contextOpen in ClaudeOpen with page context

SuggestAPI supports commerce and structured catalog search use cases.

The behaviors on this page apply when search is SuggestAPI hosted. They do not apply in search proxy mode (Manage Settings → Your search API). In that mode, /v1/autocomplete and /v1/typeahead forward to the connected search API, and that engine ranks and retrieves. See Search serving and Search Proxy.

Core search behavior

  • hybrid lexical plus semantic blending
  • FTS-only, vector-only, hybrid, and FTS-rerank modes
  • prefix-aware autocomplete with query normalization
  • spell correction before final retrieval
  • synonym expansion based on schema configuration
  • exact filters and numeric range filters

Result tuning

  • per-index ranking controls
  • lexical and semantic weight balancing
  • boost rules and optional scoring webhooks
  • pinned and curated merchandising results (Merchandising in the app)
  • display field mapping for title, subtitle, image, and URL output

Configure searchable fields, boosts, synonyms, and search mode in the app: Search index settings. Those controls change hosted retrieval only.

Typeahead path

/v1/typeahead is designed for fast prefix suggestions, making it a strong fit for:

  • nav search bars
  • mobile query assist
  • low-latency prefix suggestion UIs

In search proxy mode, typeahead is still accepted at the same path, but suggestions come from the connected search API rather than SuggestAPI prefix tables.

Intelligence features

  • query-time embeddings for semantic retrieval
  • knowledge graph scoring hooks
  • personalization scores using stored user profiles
  • query entity extraction and graph expansion

These intelligence features run on the hosted search path. They are not applied to proxied origin results.

Search proxy mode

When Search serving is Your search API (plugin):

Hosted behavior In search proxy mode
Hybrid / FTS / vector search mode Not used. The origin engine’s ranking applies. The mode query parameter does not change retrieval.
Schema searchable fields and boosts Not used for live search.
Configuration synonyms Not used. Expand queries on the origin engine instead.
Merchandising pins and campaign windows Not used. Pin on the origin engine if you need curated results.
SuggestAPI spell correction and hosted query rewriting Not used.
Personalization via user_id profiles and events Not applied to proxied hits.
Knowledge graph scoring and entity expansion Not applied.
Scoring webhooks Not invoked.
Reindex embeddings Does not change /v1/autocomplete results. Live search reads the origin API.

Filters are forwarded only when the plugin request template includes {filters} (or equivalent). Response cards follow the plugin field mapping, not hosted display-field mapping.

Keep using the same /v1/autocomplete and /v1/typeahead URLs and API keys. Switch back to SuggestAPI hosted when you need the hosted behaviors above.

When to use which endpoint

  • Use /v1/typeahead for fast label suggestions while the user is typing.
  • Use /v1/autocomplete when you need ranked product or content cards with richer payloads.
  • Use personalization and graph features when search is hosted and you have enough behavioral or catalog structure to benefit from them.