Talk to your database in natural language — replace your enterprise web admin console.

👋 Join our WeChat

• Introduction
• Quick Start
  • Docker Deploy
  • Web & CLI
  • API Deploy
• Future Plans
• Contact Us

🤖 DB-GPT-X is an open-source data application development framework that uses Large Language Model (LLM) technology to interact with databases through natural language, replacing traditional web management backends.

🌐 Currently we provide query access only. Support for full CRUD operations (Create, Read, Update, Delete) is under internal testing and will be released soon.

🚀 In the Data 3.0 era, DB-GPT-X enables enterprises and developers to build custom applications with less code — letting developers focus on complex business logic instead of managing admin consoles.

git clone https://github.com/sql-agi/DB-GPT-X

1. Configure the .env file (reference: templates.env_temple)
2. Set OPENAI_API_KEY and OPENAI_API_BASE (official API key recommended)
3. Switch to the docker directory:

cd DB-GPT-X/docker

4. Follow the instructions in docker/README.md

Note: To customize the database schema or seed data, edit /docker/sql/init.sql.

Prerequisites: Python 3.9, Conda

conda create --name db-gpt python=3.9
conda activate db-gpt
pip install -r requirements.txt

Configure .env (reference: templates.env_temple):

🔥 We strongly recommend using the official OpenAI API key. Third-party relay keys may produce degraded results.

python web_demo.py

The program starts a local web server. Open the printed URL in your browser. The latest version features a typewriter effect for improved UX.

By default share=False. Set share=True in web_demo.py for public access via Gradio's relay server (note: this may reduce typewriter performance).

python cli_demo.py

Interactive conversation in the terminal. Type quit to exit.

python api.py

Starts a local server on port 8000. Call via POST:

curl -X POST "http://127.0.0.1:8000/chat/db" \
     -H "Content-Type: application/json" \
     -d '{"input": "hello"}'

Response:

{
    "reply": "Hello! How can I help you?"
}

• Frontend: Richer UI, support for more database types and open-source LLMs
• Backend: Deep testing of complex CRUD scenarios, environment switching, role-based permissions
• Community: More user feedback integration; contributors welcome

Join the DB-GPT-X WeChat community to discuss and share feedback.

Follow our official WeChat account for updates.

• LangChain

DB-GPT-X is an independently developed open-source project maintained by the sql-agi team.

This project has no affiliation, authorization, endorsement, or cooperative relationship with eosphoros-ai/DB-GPT.

The source code, architecture design, and core implementations of DB-GPT-X are independently developed and maintained by the sql-agi team.