NLP to SQL: An Agentic AI Tool

Overview

Engineered a Streamlit web application that translates natural language queries into SQL, executes them on uploaded datasets, and generates relevant visualizations using the SmolAgents Framework powered by Google’s Gemini.

Key Features

• Natural Language to SQL Translation: Converts plain English questions into SQL queries
• Query Execution: Runs the generated SQL on uploaded datasets
• Automated Visualization: Creates appropriate charts and graphs based on query results
• Multi-Agent Architecture: Uses SmolAgents framework to coordinate specialized AI agents
• User-Friendly Interface: Simple Streamlit UI for easy interaction

Technical Details

This project leverages several cutting-edge technologies:

• Streamlit: For the web application framework
• SmolAgents: A lightweight multi-agent orchestration framework
• Gemini API: Google’s powerful language model for natural language understanding
• SQL Parser/Executor: For translating and running database queries
• Data Visualization Libraries: For creating meaningful visual representations of data

Workflow

1. User uploads a dataset
2. User enters a question in natural language
3. The NL agent translates the question to SQL
4. The SQL executor runs the query against the dataset
5. The visualization agent determines and generates the most appropriate visualization
6. Results are displayed to the user in an intuitive format

GitHub Repository

View the source code and documentation on GitHub.

Project Timeline

March 2025 - Present

Future Enhancements

• Support for more complex SQL queries
• Additional visualization options
• Query optimization suggestions
• Integration with database systems beyond uploaded files

Contact

For a demonstration or to discuss potential applications, please reach out via email.