Back to projects
// professional
WorkNatural Language DB Query Tool
PythonStreamlitPostgreSQLSnowflakeLookerLLMsCastor
Developed at JWX to empower non-technical users to query complex data sources without writing SQL. The tool is a Streamlit-based web chat app that translates natural language prompts into valid database queries using an LLM backend.
Integrated with the Castor data dictionary to provide enriched schema context, improving query accuracy and relevance. LLM-driven cell validation catches data quality issues inline before results are surfaced to the user.
Key Highlights
- ▸Query PostgreSQL, Snowflake, and Looker in plain English
- ▸Castor data dictionary integration for enriched schema context
- ▸LLM-driven cell validation for instant data quality checks
- ▸Designed for non-technical business users