Who it is forAn analyst or developer who knows what they need from the data but not the exact SQL.
The problemCopying SQL between a chat window and a database client is slow and error-prone.
The outcomeWorking SQL generated, explained or fixed inside the editor, checked before it runs.
What you need first
- macOS, Windows or Linux desktop, Docker for team access, or the web version
- An AI provider for the SQL assistant: Claude, OpenAI or a local model via Ollama
- Node.js (npx) for the MCP server
Steps
- Download DBX for macOS, Windows or Linux and connect a database
- Configure the AI SQL Assistant with Claude, OpenAI or a local model via Ollama
- Highlight a table and describe the result you want in plain language
- Ask it to explain or optimize the query, or fix an error message
- Run it with Cmd+Enter; destructive operations ask for confirmation
Prompt for your agent
In t8y2/dbx connected to <database>, use the AI SQL Assistant with <provider> to write a query that <goal>, explain it line by line, and optimize it before I run it.
Caveat.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for t8y2/dbx and "Write, explain and fix SQL in plain language with the DBX AI assistant"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer