upstashupstashOfficialยท1 Agent Skills Included

vector-js

Serverless vector database queries, upserts, and metadata filtering

Stores, queries, and manages vector embeddings in a serverless database over plain HTTP. Handles dense, sparse, and hybrid indexes with namespaces, metadata filtering, and resumable queries. Removes connection management and TCP setup so serverless functions and edge workers run similarity search instantly.
npx skills add upstash/vector-js --all -g -y

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install vector-js?โ–ผ

Run `npx skills add upstash/vector-js --all -g -y` in your terminal to install all skills in this suite globally.

How to query a vector database from serverless functions?โ–ผ

Create an Index with your REST URL and token, then call query with a vector or plain text. The HTTP-based client works in AWS Lambda, Cloudflare Workers, and Next.js without persistent connections.

How to filter vector search results by metadata?โ–ผ

Attach JSON metadata when upserting, then pass a SQL-like filter string such as "genre = 'fantasy'" to your query. It supports nested fields, arrays, glob patterns, and boolean logic.

What are namespaces in Upstash Vector?โ–ผ

Namespaces partition one index into isolated subsets so reads and writes stay scoped to a single tenant or dataset. They are created automatically on first upsert.

Does vector-js support hybrid and sparse search?โ–ผ

Yes. You can upsert and query dense, sparse, or combined hybrid vectors, with RRF or DBSF fusion algorithms for ranking hybrid results.

Related Repositories in Data & Analytics

View All in Data & Analyticsโ†’