pgvector Embeddings

Generate vector embeddings using Ollama and store them in PostgreSQL with pgvector. This skill covers the ingestion phase of RAG pipelines.

When to Apply

Use this skill when:

• Generating embeddings for documents or text
• Storing embeddings in PostgreSQL
• Building the ingestion pipeline for RAG
• Converting text to vectors for semantic search
• Chunking documents for better retrieval

Embedding Models

Recommended: nomic-embed-text

The nomic-embed-text model provides 768-dimensional embeddings with good quality and performance: