VerbaGPT

VerbaGPT is a tool designed to make data analysis with large language models (LLMs) simple and private. Users can ask questions about their SQL data, CSV files, or personal notes in natural language and receive answers. VerbaGPT runs locally to ensure data privacy, as it does not allow the LLM to access data directly; the LLM can only use the schema and explicitly provided information.

Upload data in CSV/TXT format, connect to a local or remote MSSQL, MySQL, or PostgreSQL database, and then ask questions in natural language. The system uses Text-to-Python to generate SQL queries and visualizations.

From Text to Python for Data Analysis
Local data processing to protect privacy
Supports SQL, CSV, and TXT data sources.
Advanced data visualization and modeling
Fully offline operation option (experimental)
Can block harmful SQL commands
Managing data descriptions across multiple databases
Retry button for recovery code errors.
Analysis button for commenting on code and results
Switch between different LLMs (OpenRouter, OpenAI, etc.)
Use successful question-SQL pairs to train the model.
Flag unsuccessful generation results