The secure data, document, search, and AI back-end you build apps on.
Vectros gives your application a typed data store, a document pipeline, hybrid search (BM25 + vector with RRF fusion), and citation-grounded retrieval, with multi-tenant isolation, scoped access keys, and a tamper-evident audit trail, behind one coherent API.
import { VectrosClient } from "@vectros-ai/sdk";
const client = new VectrosClient({
environment: "https://api.vectros.ai",
token: process.env.VECTROS_API_KEY!,
});
// Hybrid (keyword + semantic) search over your indexed content, with citations.
const results = await client.search.content({ query: "annual leave policy" });- Docs & quickstart → https://docs.vectros.ai
- Website → https://vectros.ai
| Language | Install | Repository |
|---|---|---|
| TypeScript / Node.js | npm install @vectros-ai/sdk |
vectros-sdk-node |
| Python | pip install vectros |
vectros-sdk-python |
| Java | ai.vectros:vectros-sdk (Maven Central) |
vectros-sdk-java |
The full API surface is described in vectros-api-spec (OpenAPI). Agents can reach Vectros over MCP via vectros-mcp-server.
- Hybrid search over your documents and structured records, keyword + vector, fused.
- Document ingestion with schema-defined indexing, lookup/range queries, and folders.
- Grounded inference: chat, RAG, and document-ask, answered with citations.
- Multi-tenant by design: per-customer isolation, scoped keys, and app-contexts.
Per-customer fail-closed isolation, least-privilege scoped keys, and a tamper-evident audit and version history. Customer-facing surfaces are hardened through extensive adversarial security review. See the compliance & trust guide.
Building something with Vectros? Start at docs.vectros.ai.