RAG with Claude
RAG lets Claude answer questions using your own documents and data — not just its training knowledge. This section covers the full RAG stack from ingestion to evaluation, all using the Claude API and Anthropic-compatible tooling.
In This Section
What is RAG
Why RAG exists — the gap between static model knowledge and live retrieval, and when RAG is the right answer.
RAG vs Fine-Tuning
When to use RAG, when to fine-tune, and when to combine both — a decision framework.
RAG Pipeline Anatomy
Ingest, chunk, embed, store, retrieve, generate — every stage of the pipeline explained.
Embedding Models
How embeddings work, which models pair well with Claude, and how to choose by task and cost.
Vector Databases
Pinecone, Weaviate, Chroma, pgvector — tradeoffs, when to use each, and integration patterns.
Advanced RAG
Hybrid search, HyDE, reranking, multi-hop retrieval — techniques that improve RAG quality significantly.
RAG Evaluation
Measuring faithfulness, relevance, and groundedness — metrics and tools for assessing RAG quality.