Show HN: Reverse engineering user queries for ChatGPT/Gemini

A new open-source tool lets brands audit how they appear in AI assistant recommendations from ChatGPT, Gemini, and Google AI Mode, revealing which competitors are cited and which queries they are missing. The tool, built for ecommerce but usable by any brand, generates client-ready PDF reports with KPIs, competitor share, and prioritized fixes.

See if AI recommends your store — or your competitor's. An open-source audit of how your brand and products show up when shoppers ask ChatGPT, Gemini, Google AI Mode & Google AI Overview "best X for Y" or "where to buy Z". Built for ecommerce — works for any brand SaaS, local, B2B . Shoppers increasingly ask AI assistants what to buy — and buy what gets recommended. AI referral traffic to online stores is growing triple-digits year over year. But when someone asks ChatGPT "best running shoes for flat feet" or "where to buy an espresso machine under $500", you can't see whether your store is recommended, ignored, or losing to a competitor. This tool shows you — across every major AI assistant, for every buyer query that matters in your category. Audit report — a client-ready PDF you can hand to a stakeholder: KPIs, competitor share, the queries you're missing, and concrete recommendations. - Which competitor AI recommends instead of you — and for which buyer queries - The high-intent questions "best…", "where to buy…" where your store is invisible - Which domains AI cites as sources and which of your pages, if any - Sentiment and how you're framed when your brand is mentioned - A prioritized list of what to fix first Showing the built-in sample dataset — load it in one click, no API keys required. Responses — every AI answer about your category: which store gets named, recommended, and cited. Dashboard — visibility, the GEO funnel, competitor share, and the buyer queries where AI recommends someone else. Topics — which product categories you win in AI answers, and which you're losing. Prompts — the buyer queries being tracked, each with how often AI names your store and the trend over time. git clone https://github.com/syntropicsignal-ai/ai-visibility-audit.git cd ai-visibility-audit docker compose up -d Then open http://localhost:8080 http://localhost:8080 — enter your API keys on first visit, or load the sample dataset and explore with no keys at all. Migrations run automatically on start. Point it at your store URL — Shopify, WooCommerce, PrestaShop, Magento, BigCommerce, Medusa, Saleor, Vendure, or PL platforms like Shoper, IdoSell & BaseLinker. The audit runs on any brand or domain, regardless of stack. - Docker Desktop or Docker Engine + Compose v2 - API keys: Gemini and Exa prompt generation + analysis , DataForSEO Google AI Overview + keyword signals . - Bright Data — unlocks the ChatGPT, Gemini, and Google AI Mode sources, the main AI assistants this tool measures. Without it you'll track Google AI Overview only via DataForSEO . api/ FastAPI backend Python, uv web/ Vue 3 frontend TypeScript, Vite docker/ Container build files