VerbaGPT

VerbaGPT is a data analysis platform that lets non-technical users query databases and files using natural language. It offers two modes: Virgo, a cloud-powered browser analytics tool at app.verbagpt.com, and Taurus, an optional local AI agent app for querying databases on a user's own computer or network and for agentic workflows. It generates Python code (including SQL queries) that runs in a sandboxed environment and returns results with charts and tables. The page also describes Data Notes for adding business context to databases, Prompt Libraries and Footnotes for sharing effective prompts, client access without per-seat licenses using credit caps, and security features such as encryption, role-based access control, and audit logging. Pricing plans are managed inside the app rather than on the landing page.

Key features

• Natural language to Python and SQL analysis with charts and tables
• Connections to PostgreSQL, MySQL, MSSQL, and Snowflake
• File support for CSV, Excel, JSON, PDF, and Google Drive import
• Multi-model support including GPT-4 and Claude via providers
• Smart schema selection using embeddings to find relevant columns
• Data Notes: contextual data dictionary for tables, columns, and business logic
• Prompt Libraries and Footnotes for sharing prompts across an organization
• Private Inference via Ollama, LM Studio, or OpenAI-compatible engines (with Taurus)
• Query databases on private networks through the Taurus connector without internet exposure
• Read-only database access enforced by Taurus
• Advanced analytics: regression, time-series forecasting, classification, clustering, and custom matplotlib visualizations
• Client access links with scoped datasets and credit caps
• Learning from corrections and thumbs-up ratings, with reviewable learned rules
• Reference facts with source links and dates, with opt-in web search
• Authentication with JWT, bcrypt-hashed passwords, and role-based permissions
• AES-256 encryption at rest and TLS 1.3 in transit
• Audit logging with admin dashboard and CSV export
• White-label branding for Team and Enterprise plans

Best suited for

• Non-technical business users who need data insights without SQL or Python
• Data teams sharing query context and prompts across an organization
• Consultancies and agencies sharing database access with clients
• Organizations with sensitive data that must stay on local networks
• Data science teams doing exploratory statistical modeling and machine learning

Pricing

Not stated on the website