DB-GPT

DB-GPT is an open-source, AI-native framework: it connects Large Language Models (LLMs) to diverse data sources for natural language interaction and Generative Business Intelligence (GBI).

DB-GPT functions as a robust, AI-native data app development framework: it bridges LLMs with databases, data warehouses, and Excel files. Its core architecture includes the Service-oriented Multi-model Management Framework (SMMF) and the Agentic Workflow Expression Language (AWEL), simplifying complex data application builds. Key capabilities include Multi-Source Retriever-Augmented Generation (MS-RAG) and automated Text2SQL fine-tuning; for example, it achieved an 82.5% accuracy on the Spider dataset. This framework ensures data privacy and security while enabling users to generate analytical reports and business insights through simple chat commands.

https://dbgpt.site/