Neura Market
Web logo

Web

Free

The SQL Agent Open Source – Multi-database, Multi-turn, Access controlled

FreeFree tier
Inputs: textOutputs: text
Type
Open Source
Visit Website GitHub (40 stars)

About Web

Vanna 2.0 is an open-source SQL agent framework that enables users to interact with databases using natural language. It supports multiple databases (SQLite, PostgreSQL, MySQL, Snowflake, BigQuery, and others) and works with any major LLM provider (Anthropic, OpenAI, Gemini, Ollama). Built for multi-turn conversations and access-controlled environments, Vanna allows organizations to let users query company systems of record without SQL expertise. The framework offers optional hosted admin features including access control, observability metrics, agent memory via vector DB, secure file storage, audit logs, and data retention. It can be self-hosted or used via Vanna Cloud with tiered pricing plans.

Key Features

Natural language querying of SQL databases
Supports multiple databases: SQLite, PostgreSQL, MySQL, Snowflake, BigQuery, and others
Works with any LLM provider: Anthropic, OpenAI, Gemini, Ollama
Multi-turn conversations for complex queries
Access control to manage who sees what data
Observability with real-time metrics and distributed tracing
Agent memory via hosted vector database for learning
Secure file storage and retrieval
Comprehensive audit logs for activity tracking and compliance
Data retention for long-term storage of conversations and logs

Pros & Cons

Pros
  • Open source with no vendor lock-in; works with any LLM and database
  • Supports multi-turn conversational SQL generation
  • Built-in admin features for enterprise needs: access control, audit logs, observability
  • Flexible deployment: self-hosted or managed cloud with low-cost starter plans
  • Easy to get started: 1-minute Colab demo or pip install
Cons
  • Advanced admin features (access control, observability, etc.) require paid cloud plans
  • Daily usage limits on paid plans (20 questions/day on Explorer tier)
  • Dependency on external LLM APIs may introduce cost and latency

Best For

Empowering non-technical users to query databases using natural languageReducing time spent writing SQL queries for analysts and data scientistsGenerating data insights across entire organizationsEnabling SQL agents in internal tools with access control and complianceBuilding conversational data interfaces for customer-facing products

FAQ

What happens if I exceed my daily limits?
Daily throughput limits reset every 24 hours. If you exceed your limit, requests will be rate-limited until the next reset. You can upgrade to a higher tier at any time for immediate access to higher limits.
Can I start with the open source version and upgrade later?
Yes. All tiers use the same core framework. You can start self-hosting with the open source version, then seamlessly upgrade to the managed cloud platform or enterprise deployment without rewriting your code.
How does Enterprise pricing work?
Enterprise pricing is custom based on deployment needs (on-prem, VPC, or private cloud), desired SLAs, and support requirements. Contact the sales team to discuss specific requirements.