Hello Product Hunt 👋 — Luciano here, founder of Symphony. Over the past year, we noticed something strange: Founders are building better products than ever, but when people ask ChatGPT, Claude, Perplexity, or other AI systems to recommend tools in their category, many of those products simply do not appear. Not because the product is bad. Because AI search cannot clearly read, trust, or cite the website. The discovery layer is changing. It is no longer only about ranking links in Google. It is also about being understood, retrieved, and cited by AI systems. So we built Prelude by Symphony. 🚀 Prelude is an open-source, terminal-first CLI for AI search readiness. You can run: `npx symphony-prelude audit https://your-site.com` to get a fast local diagnostic of your page structure, metadata, schema, content chunking, trust signals, and AI crawler access. And with an OpenAI API key, you can run: `npx symphony-prelude trace "your discovery query" --domain your-site.com` to see what OpenAI web search actually does: → what it searches → what URLs it opens → what pages it cites → what pages are duplicated, weak, or hard to quote → what opportunities exist to improve AI visibility Prelude does not try to be a dashboard. It is the diagnostic layer: fast, local, open source, and built for developers. Prelude reveals the problem. Conservatory helps fix it with full AI search reports, downloadable PDFs, auto-fix, and GitHub PRs: https://conservatory.app I would love your honest feedback today: • Does the terminal workflow feel useful? • Does the trace change how you think about your website? • What would make this more actionable for founders, marketers, and developers? I’ll be here all day answering everything 🙌

Prelude by Symphony

Trace what OpenAI web search opens, cites, and misses

Trace what OpenAI web search opens, cites, and misses