vectordb

Semantic search with vector embeddings works by converting text into high-dimensional vectors, storing them in a vector database, and retrieving similar documents by comparing vector proximity. This Skill provides embedding generation, storage across Pinecone, Chroma, and pgvector, and retrieval logic to enable similarity-based queries at scale.

This Skill integrates with Pinecone, Weaviate, Chroma, and pgvector (PostgreSQL), providing Python and SQL interfaces for embedding storage, upsert, querying, and metadata filtering across all supported stores for production RAG implementations.

Generate embeddings using OpenAI models, then use this Skill's vector store integration to upsert embeddings with metadata into Pinecone, Chroma, pgvector, or Weaviate. The Skill handles index provisioning, storage, and retrieval operations automatically.

Yes, pgvector is a PostgreSQL extension supported by this Skill. It enables you to store and query vector embeddings directly within PostgreSQL using SQL interfaces, combining relational data with semantic search capabilities.

Pinecone is a managed vector database optimized for large-scale similarity search, while pgvector embeds vectors within PostgreSQL for integrated relational and vector queries. Choose Pinecone for pure vector workloads at scale; pgvector for co-located structured and vector data.

Yes, document chunking is part of the RAG pipeline this Skill supports. Break large documents into manageable chunks, generate embeddings for each chunk, and store them with metadata for effective retrieval of relevant context in RAG workflows.