Hi HN, I'm Gonzalo, a CS student in Argentina. QueryPilot is a Chrome extension I built because general-purpose chatbots kept inventing column names when I asked them for SQL. How it works: - You run one information_schema query in your console (Supabase, Neon, pgAdmin) and paste the result. The extension parses JSON, CSV, psql output or CREATE TABLE DDL and stores the tables and columns locally. - That schema goes into the system prompt, so generated PostgreSQL uses your real names. Results can be inserted straight into the page's editor (Monaco / CodeMirror) via a content script, so undo still works. - You can paste EXPLAIN ANALYZE output and it points at the expensive nodes (seq scans with selective filters, big nested loops, bad row estimates) and suggests CREATE INDEX CONCURRENTLY statements. Some design decisions: - No backend at all. It's bring-your-own-key (Gemini, Claude or OpenAI) and requests go from the browser directly to the provider. Keys and schemas live in chrome.storage.local. - It never connects to your database or executes anything. You review and run the SQL yourself. - Built with WXT (MV3), React and TypeScript. Licenses are validated with Polar. There's a free tier (5 queries a day, 1 project). The index optimizer and unlimited usage are a one-time $29. Chrome Web Store: https://chromewebstore.google.co... I'd especially like feedback on the schema onboarding (is pasting a query result too much friction?) and on cases where the index suggestions are wrong.

QueryPilot

AI SQL copilot that knows your real database schema

AI SQL copilot that knows your real database schema