apconw/Aix-DB

What it solves

Aix-DB is a conversational data analysis system (ChatBI) that allows users to extract insights and create visualizations from databases using natural language, removing the need for manual SQL writing.

How it works

Built on LangChain and LangGraph with an MCP (Model Context Protocol) multi-agent architecture, the system follows a specific pipeline to convert text to data:

1. Intent Understanding: An LLM parses the user's natural language query to extract entities and conditions.
2. Knowledge Retrieval: A RAG engine uses a hybrid of Embedding, BM25, and Neo4j graph knowledge to find relevant table structures and business context.
3. SQL Generation: A Text2SQL engine generates, validates, and optimizes the SQL query.
4. Execution: The SQL is executed against the target data source.
5. Visualization: Results are automatically rendered using ECharts or AntV charts.

Who it’s for

It is designed for users who need to perform data analysis and reporting but may not be proficient in SQL, as well as developers looking for an easy-to-integrate AI data assistant.

Highlights

• Broad Database Support: Compatible with MySQL, PostgreSQL, Oracle, SQL Server, ClickHouse, DM, Apache Doris, and StarRocks, as well as CSV and Excel files.
• Multi-Agent Architecture: Utilizes MCP Skills for collaborative agent-based data processing.
• End-to-End Pipeline: Handles everything from natural language input to visual chart output.
• CLI Tool: Includes a command-line interface for querying data directly from the terminal.