GitHub
FreeChat with your SQL database — Natural language to SQL generation via LLMs using Agentic Retrieval
FreeFree tier
Inputs: textOutputs: text
About GitHub
Vanna is an open-source framework that converts natural language questions into SQL queries using large language models (LLMs) with agentic retrieval. It provides a modern web interface (vanna-chat) that streams tables, charts, and summaries in real time. Vanna supports any LLM (e.g., OpenAI, Anthropic, Ollama) and any SQL database (e.g., PostgreSQL, Snowflake, BigQuery). It includes enterprise features such as user-aware permissions, row-level security, audit logs, and rate limiting. The framework is extensible with custom tools, lifecycle hooks, and observability. Note: The project is archived and no longer actively maintained.
Key Features
Natural language to SQL generation using LLMs with agentic retrieval
Multi-LLM support: OpenAI, Anthropic, Ollama, Azure, Google Gemini, AWS Bedrock, Mistral, others
Multi-database support: PostgreSQL, MySQL, Snowflake, BigQuery, Redshift, SQLite, Oracle, SQL Server, DuckDB, ClickHouse, others
User-aware permissions and row-level security
Streaming responses with progress updates, SQL code blocks, interactive data tables, Plotly charts, and natural language summaries
Pre-built web component (vanna-chat) that works with React, Vue, or plain HTML
Enterprise features: audit logs, rate limiting, lifecycle hooks
FastAPI integration with built-in tracing and metrics
Extensible: custom tools, LLM middlewares, content filtering
Bring your own auth: cookies, JWTs, OAuth tokens
Pros & Cons
Pros
- Accurate text-to-SQL generation using agentic retrieval
- Supports a wide range of LLMs and databases
- Enterprise-grade security features out of the box
- Pre-built modern UI for quick deployment
- Real-time streaming of results including tables and charts
- Extensible and customizable with hooks and custom tools
- Open source and free to use
Cons
- Project is archived and no longer actively maintained (read-only as of 2026)
- Requires server setup and infrastructure to run
- Limited to SQL databases only
- Dependency on external LLM API services
- May require technical knowledge to configure and deploy
Best For
Business intelligence and data analyticsEnabling non-technical users to query databases using natural languageBuilding custom AI-powered data exploration toolsEnterprise reporting with natural language interface and security controlsEmbedding SQL chat capabilities into existing web applications
FAQ
What LLMs does Vanna support?
Vanna supports OpenAI, Anthropic, Ollama, Azure OpenAI, Google Gemini, AWS Bedrock, Mistral, and others.
What databases can Vanna connect to?
Vanna supports PostgreSQL, MySQL, Snowflake, BigQuery, Redshift, SQLite, Oracle, SQL Server, DuckDB, ClickHouse, and others.
Is Vanna free to use?
Yes, Vanna is open source and free to use. Note that the project is archived and read-only.
Does Vanna include a built-in user interface?
Yes, Vanna provides a pre-built web component called vanna-chat that streams results including tables and charts, and it works with React, Vue, or plain HTML.
What security features does Vanna offer?
Vanna 2.0 includes user-aware permissions, row-level security, audit logs, and rate limiting, with authentication via cookies, JWTs, or OAuth tokens.