Vector Documentation Skill
Quick Start
Vector is a high‑performance vector database for storing, querying, and managing vector embeddings.
Basic workflow:
- Install the Vector TS SDK.
- Connect to a Vector instance.
- Upsert vectors, query them, and manage namespaces.
Example (TypeScript):
import { Index } from "@upstash/vector";
const index = new Index({
url: process.env.UPSTASH_VECTOR_REST_URL!,
token: process.env.UPSTASH_VECTOR_REST_TOKEN!,
});
await index.upsert([{ id: "1", vector: [0.1, 0.2], metadata: { tag: "example" } }]);
const results = await index.query({
vector: [0.1, 0.2],
topK: 5,
});For full usage, refer to the linked skill files below.
Other Skill Files
TS SDK Reference
sdk-methods: Explains SDK commands: delete, fetch, info, query, range, reset, resumable-query, upsert
Features
features/namespaces: Explains namespaces and dataset organization.features/index-structure: Covers hybrid and sparse index structures.features/filtering-and-metadata: Details metadata storage and server-side filtering.
Use these files for deeper guidance on SDK usage, advanced configurations, algorithms, and integrations.