vanna-ai/vanna: Chat with your SQL database
The most-starred open-source text-to-SQL framework; MIT-licensed and works with any LLM including local ones.
Open github.com →Three working patterns: connect an MCP server (like Postgres MCP) so Claude or Cursor can query your database conversationally, deploy an open-source text-to-SQL layer like Vanna or a BI tool with a built-in AI analyst like Metabase's Metabot, or upload exports to an AI data analyst like Julius. Accuracy jumps when the tool can see your real schema, so prefer connected tools over pasting CSVs into a chatbot. Keep the connection read-only and define key metrics once, or 'revenue' will quietly mean something different every week.
A quick orientation. The real value is below: resources worth your time, from people who've actually done it.
The most-starred open-source text-to-SQL framework; MIT-licensed and works with any LLM including local ones.
Open github.com →A hands-on walkthrough of training Vanna on your own schema.
Open medium.com →Built for exactly this user: the CEO or support lead who needs answers without a BI seat.
Open askyourdatabase.com →A database-tooling company's sober shortlist of what actually generates runnable SQL.
Open bytebase.com →Configurable read/write access so your AI can query the database without being able to wreck it.
Open github.com →Explains how schema introspection lets Claude reason about your data model, not just guess.
Open pgedge.com →The shortest path from zero to chatting with your production database.
Open datamcp.app →For the slightly technical founder who wants to understand what the connector actually does.
Open dev.to →Self-serve BI with Metabot AI: teammates ask questions, it writes and explains the SQL.
Open metabase.com →Context on the fast-growing category of upload-and-ask AI data analysts.
Open techcrunch.com →Honest about the failure mode founders must know: no semantic layer means inconsistent metric definitions.
Open upsolve.ai →A concrete build of a chat-with-your-database app, demystified in under an hour.
Watch on YouTube youtube.com →The upgraded version showing how fast this pattern matured in one year.
Watch on YouTube youtube.com →Moda made company data queryable from Slack; the blueprint for a data-access culture.
Open lennysnewsletter.com →Weekly operator demos of AI workflows, including the Amplitude internal-data episode.
Listen on Apple Podcasts podcasts.apple.com →For the half-technical founder: SQL clients where AI assists rather than replaces you.
Open dbgate.io →Connects Postgres plus HubSpot, Sheets, and Shopify, matching a founder's actual scattered stack.
Open talktoyourdata.app →Why text-to-SQL goes wrong and the guardrails that keep answers trustworthy.
Open medium.com →