Overview
DB-GPT is a framework that brings large language models together with structured databases. It offers Text2SQL capabilities, vector-based retrieval (RAG), and adapters for multiple vector stores and model backends.
Key features
- Text2SQL: translates natural language queries into SQL for structured data interaction.
- RAG integration: supports retrieval-augmented generation for context-rich responses.
- Multi-backend adapters: connectors for popular vector databases and model providers.
Use cases
- Data analysis assistants: query enterprise databases using natural language.
- Smart reporting and BI: generate SQL and visual queries from user prompts.
Technical details
- Implemented in Python with adapter patterns for backends, providing examples and deployment guides for integrating with existing data platforms.