I built a simple tool that shows what an LLM (OpenAI model) thinks about any product — both based on its internal knowledge and web search. It also shows which sources the model relied on. You can download the analysis as a JSON file or share it via link.
I’m using:
- openai/gpt-4.1-mini for queries that require search
- openai/gpt-4.1-nano for those that don’t
Since the product is completely free, I’m sticking with the cheaper models. I do have a cost cap :)
Tech-wise, it’s a Python backend with FastAPI and some Node.js. I’m tracking analytics with PostHog: mainly unique users and search queries.
Next steps I’m considering:
- Open-sourcing it and letting users provide their own OpenAI API keys
- Adding smarter models, deeper analysis, and customizable categories behind a paywall
Would be great to hear where you see this being useful.