Every layer of the search system — from indexing through curation and analytics — has an API for developers and an admin screen for operators.
How each capability works, the request that drives it, and when to reach for it.
One request runs keyword and vector retrieval, merges them with reciprocal rank fusion and can rerank the top hits. Hosted API — no cluster to operate.
Nearest-neighbour search over your own or server-generated embeddings: query by vector, by document or by image, with the same filters and API keys as keyword search.
Send plain text, get the nearest documents by meaning. The query is embedded server-side; filters keep the results inside your real inventory.
Search several collections in a single HTTP request and get the results grouped per collection or merged into one ranked list.
Brand, price, size and category filters with counts computed for the current query — from the same request that returns the hits.
Rerank the leading hits with a cross-encoder or a language model using one parameter. If the model is slow or fails, the search still answers.
Retrieval, prompt assembly, the model call and citations behind one endpoint. Every answer returns the documents it was based on.
REST API and typed SDKs (TypeScript, Python, Go, PHP). Webhooks and real-time sync. MCP server for AI agents.
Embedded admin console. Synonyms, curation, analytics and experiments. No code required to tune search.
Every tenant is isolated at the database level. API keys scope by collection, field, filter and IP. OIDC, Google and GitHub SSO authenticate users. SCIM provisions and deprovisions teams. Every operation logs to an audit trail.
API Resource
/tenantsAdmin Screen
Tenants & Users
Developer
Scoped keys restrict what an API key can query and modify
Operator
Manage user roles, SSO, and per-key restrictions from the admin console
Documents sync in milliseconds. Push via the REST API, webhooks ingest from your infrastructure, or scheduled crawls keep catalogues synchronized without rebuilding or redeploying.
API Resource
/documentsAdmin Screen
Collections & Indexing
Developer
POST /v1/documents indexes a batch; webhooks keep your data fresh
Operator
Schedule crawls, monitor indexing status, and manage document retention from the console
One-way (sneakers → trainers) and multi-way (fully interchangeable) synonyms, transliteration ladders and typo rescue apply on every query without reindexing.
API Resource
/synonymsAdmin Screen
Synonyms
Developer
Programmatically manage synonym sets via the synonyms API
Operator
Build and test synonym rules from the UI; see query coverage for each rule
Pin products to the top, boost categories for the season, run sponsored slots for partners, and exclude competitors — all without reindexing. Changes take effect in milliseconds.
API Resource
/search-rulesAdmin Screen
Search Rules
Developer
Rules apply automatically; pass search_profile to switch between curated sets
Operator
Create pins, boosts, exclusions, redirects and scheduled campaigns from the UI
Track what users search for, what they click, and how they interact. A/B test ranking changes with confidence intervals. Golden query sets score search quality with nDCG, MRR and MAP.
API Resource
/analyticsAdmin Screen
Analytics & Experiments
Developer
Log click events via the tracking pixel or the API; query the analytics REST endpoint
Operator
Monitor search volume, create experiments, and benchmark ranking quality from the console
Behavioral profiles feed 16 recommendation models. Ranking adjusts per-visitor without a separate ML infrastructure. Profiles refresh with every interaction.
API Resource
/recommendationsAdmin Screen
Personalization
Developer
Pass session or user ID; get personalized results and recommendations automatically
Operator
Tune recommendation models, view behavioral profiles, and manage personalization settings
The best way to evaluate a search platform isn't reading one more comparison table.
Only then make the decision.