🧠 TalkWithDB - AI Database Agent

Chat with any SQLite database using natural language. No setup required—just upload your .db file, add your Gemini API key, and start asking questions.

🔗 Try the Live Demo Here!

This project is an AI-powered agent that lets you have a conversation with your data. The agent uses Google Gemini and LangChain to understand your questions, inspect the database you provide, write SQL queries, and give you back answers in plain English.

---

🚀 Key Features

• Bring Your Own Database: Upload your own SQLite .db file directly in the browser.
• User-Provided API Key: Securely use your own Google Gemini API key.
• Live SQL Query Display: See the SQL queries the AI generates in real-time in the sidebar.
• Natural Language Q&A: Ask complex questions without writing a single line of SQL.
• Automatic Schema Detection: The agent automatically inspects table structures to write accurate queries.
• Built with: Streamlit + LangChain + Google Gemini.

---

💡 How to Use the Live App

Get started in seconds with the hosted Streamlit application.

1. Open the App: Navigate to TalkWithDB on Streamlit.
2. Configure: In the sidebar, enter your Google Gemini API Key.
3. Upload: Upload your SQLite (.db) database file.
4. Chat: Once configured, the chat window will activate. Start asking questions!

Example Questions:

• "Show me all employees in the Engineering department."
• "What are the salaries of managers?"
• "How many users signed up last week?"

---

🔧 Local Development Setup

Want to run the app locally or contribute? Follow these steps.

1️⃣ Clone the Repository

git clone https://github.com/your-username/TalkWithDB-AI-Agent.git
cd TalkWithDB-AI-Agent

2️⃣ Install Dependencies using uv

This project uses uv for fast dependency management. First, install uv:

pip install uv

Then, sync the project dependencies:

uv sync

3️⃣ Run the Streamlit App

uv run streamlit run app.py

This will launch the app in your browser, typically at http://localhost:8501.

---

📂 Project Structure

.
├── app.py                 # The main Streamlit application
├── utils/database.py      # Database utility functions (list_tables, get_schema, etc.)
├── temp/                  # Directory for temporarily storing uploaded .db files
├── pyproject.toml         # uv-managed project dependencies
├── uv.lock                # Lockfile for reproducible builds
└── README.md              # You are here!

---

📜 License

This project is licensed under the MIT License.