RAG (Retrieval-Augmented Generation)

RAG enhances LLM responses by retrieving relevant context from your own data before generating an answer. DB-GPT provides a comprehensive RAG framework supporting multiple retrieval strategies.

How RAG works

1. User asks a question
2. Retriever searches your knowledge base for relevant documents
3. Context is assembled from retrieved chunks
4. LLM generates an answer grounded in the retrieved context

Knowledge base types

DB-GPT supports multiple knowledge base types out of the box:

Type	Storage	Best for
Vector	ChromaDB, Milvus, OceanBase	Semantic similarity search
Knowledge Graph	TuGraph, Neo4j	Entity relationships, structured Q&A
Keyword (BM25)	Built-in	Exact keyword matching
Hybrid	Combines multiple	Best-of-both retrieval

Supported file formats

Upload and process a wide variety of document formats:

• Documents: PDF, Word (.docx), Markdown, TXT
• Spreadsheets: Excel (.xlsx), CSV
• Web: HTML, URLs
• Code: Python, Java, and other source files

RAG pipeline

The full RAG pipeline in DB-GPT:

Key steps

1. Parsing — Extract text from uploaded documents
2. Chunking — Split text into manageable segments
3. Embedding — Convert chunks into vector representations
4. Storage — Store vectors in a vector database
5. Retrieval — Find relevant chunks for a given query
6. Re-ranking — Optionally re-rank results for better relevance
7. Generation — Feed context + query to the LLM

Quick start with RAG

1. Open the DB-GPT Web UI
2. Navigate to Knowledge Base in the sidebar
3. Create a new knowledge base
4. Upload your documents
5. Wait for processing to complete
6. Start chatting with your knowledge base

For programmatic access, see the RAG Cookbook.

What's next

• Knowledge Base UI — Manage knowledge bases in the Web UI
• Graph RAG — Knowledge graph-based retrieval
• RAG Module — Deep dive into the RAG framework
• RAG Development Guide — Build RAG apps programmatically