SQLBot is a self-hosted conversational data-analysis application for turning natural-language questions into data results and visualizations through Text-to-SQL. It is intended for teams that can configure a data source and an LLM service.
Project overview
SQLBot is a self-hosted conversational data-analysis application for turning natural-language questions into data results and visualizations through Text-to-SQL. It is intended for teams that can configure a data source and an LLM service.
Project type
RAG · Data Processing
Use cases
Data Analysis
Deployment
Refer to project documentation
License
License pending
Best for
Data, enterprise, and development teams that need a self-hosted web interface for natural-language database questions, can run a Linux server, and can provide a compatible LLM service and data source.
Organizations that need workspace-level resource isolation and can calibrate prompts, terminology, and SQL examples for their business scenarios.
Key capabilities
Combines LLM natural-language understanding and SQL generation with RAG for Text-to-SQL conversion.
Supports ChatBI-style questions to retrieve data information and visualization charts, with further intelligent analysis support.
Provides workspace-level isolation to establish data boundaries and support secure, compliant data access.
Supports web embedding, pop-up embedding, and MCP calls for integration into other applications.
Supports custom prompt wording, terminology libraries, and maintainable SQL example calibration for business-specific scenarios.
Limitations and risks
The GPLv3 license includes additional restrictions that prohibit replacing or modifying the logo and copyright information.
Operation requires a data source and a Large Language Model service API that is OpenAI-compatible or native.
GPU requirements, telemetry behavior, and cost dependencies are not documented in the available information.
Getting started
Install Docker on a Linux server, run the documented Docker command, then open the web interface at port 8000 and sign in with the documented initial administrator credentials. Configure the LLM and data source before use.
Evidence and sources
GitHub project description: 🔥 基于大模型和 RAG 的智能问数系统，对话式数据分析神器。Text-to-SQL Generation via LLMs using RAG.