SQL Querying Agents With Tools

A LangGraph-powered AI agent that lets you query a SQLite database using plain English. No SQL knowledge required — just ask a question and the agent writes and executes the query for you.

---

How It Works

User Question
      |
      v
   Agent  <-----> Tools (get_schema, execute_sql)
      |
      v
  Final Answer

The agent first fetches the database schema, then writes and executes the appropriate SQL query, and returns the result in plain English.

---

Features

• Natural language to SQL — no query writing needed
• Live SQL execution against a real SQLite database
• Streaming step-by-step agent output
• Example questions in the sidebar to get started quickly
• Clean Streamlit web UI

---

Live Demo

Open the app

---

Local Setup

git clone https://github.com/Ishaan20072612/SQL-Querying-Agents-With-Tools.git
cd SQL-Querying-Agents-With-Tools
pip install -r requirements.txt
streamlit run sql_app.py

---

Sample Database

Users (5 rows): id, name, email, signup_date
Orders (4 rows): id, user_id, amount, status, order_date

Example questions you can ask:

• How many users have not placed any orders?
• What is the total revenue from completed orders?
• Which user has spent the most?
• Show all pending orders with user names.

---

Tech Stack

Library	Purpose
langchain-core	Tool definitions
langchain-openai	GPT-4o-mini
langgraph	Agent graph orchestration
streamlit	Web UI

---

Project Structure

├── sql_app.py          # Streamlit UI
├── sql_agents.py       # Core agent logic
├── requirements.txt    # Dependencies
└── README.md

---

Author

Ishaan Chowdhury · @Ishaan20072612

Built With

• python