Vanna provides an open‑source AI agent framework that connects large language models to any SQL‑compatible database, letting users ask natural‑language questions and receive results such as SQL queries, visualizations, or custom tool actions. The platform includes multi‑turn conversation, tool memory for reusing successful queries, and fine‑grained access control, and can be deployed as a managed cloud service or self‑hosted solution.
Funding
$20K raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Many organizations store critical data in relational databases but their non‑technical staff lack SQL skills, making it difficult to retrieve insights quickly. Traditional BI tools often require manual query building, complex setup, and rigid permission models, leading to bottlenecks and underutilized data.
Solution
Vanna 2.0 provides a user‑aware AI agent framework that connects large language models (LLMs) to any supported database, allowing users to ask natural‑language questions and receive results such as SQL queries, charts, or custom tool outputs. The platform records successful interactions in a tool‑memory, enabling the agent to reuse and adapt past queries for faster, more accurate responses. Built‑in access‑control propagates user permissions throughout the system, ensuring that each user can only retrieve data they are authorized to see. Vanna can be self‑hosted or run as a managed cloud service, and it integrates with any LLM provider and database via simple configuration code.
Target Audience
Primary customers are data‑driven enterprises—product teams, analysts, and business users—who need self‑service access to relational data without requiring SQL expertise, as well as developers building internal AI‑powered data assistants.
Features
- Open‑source framework that works with multiple LLM providers (e.g., Anthropic Claude) and any SQL‑compatible database
- Multi‑turn conversational interface that generates SQL, visualizations, or custom tool actions from natural language
- Tool Memory that stores successful question‑SQL (or tool) pairs and reuses them to improve accuracy over time
- Fine‑grained permission enforcement through user groups that control access to tools and data
- Extensible tool registry allowing developers to add custom functionalities such as chart generation or domain‑specific actions
- Ready‑to‑run FastAPI server for easy deployment as a cloud service or on‑premise solution
- Simple Python configuration with pip install, LLM setup, database connector, and optional authentication classes