Create a collection, load documents and run your first query over the API. Use REST, an SDK or MCP — all on top of one search platform.QuickstartAPI Reference
import { AACSearch } from "aacsearch";
const client = new AACSearch({ apiKey });
const { hits } = await client.search("products", {
q: "wireless headphones",
});Describe the document fields and the searchable configuration.
Load data through the API, an import or a connector.
Use keyword, hybrid or vector retrieval.
Report clicks and conversions for analytics, recommendations and personalization.
AACSearch publishes its API contract in a machine-readable format. That makes it usable for:
The documentation and the client examples have to match that same contract.
Use the generated client where one is available. Prefer no SDK? The REST API stays a first-class interface.
The CLI scaffolds a collection and indexes your first documents. It is generated with the SDKs; a package release is pending, so setup runs from the repository.
Connect AACSearch to MCP-compatible tools. The agent gets controlled search access to the same data your application uses.
Don't build a separate RAG index just for a coding agent.
The frontend must never receive an administrative key. Create restricted credentials for:
Query Trace lets you follow a request:
You see not just the result, but the path that produced it.
An API should be simple enough for the first request and transparent enough for production debugging.
import { AACSearchClient } from "aacsearch";
const client = new AACSearchClient({ apiKey });
const { hits } = await client.search("products", { q: "cordless drill" });Explore more
The best way to evaluate a search platform isn't reading one more comparison table.
Only then make the decision.