som-shahlab/Clinfo.AI ? reverse-engineered prompt
Reverse engineered prompt
Build me a Clinfo.AI style demo that lets a user ask a specific medical question and gets back a grounded answer from scientific literature.
I want it to search PubMed, pull back the most relevant papers, filter out the ones that do not match the question, summarize the useful ones, and then combine those summaries into a clear final response with citations or paper links when possible. It should work with an OpenAI model using an API key, and also support a local open source model through a vLLM endpoint if I want to switch later. Please make the setup easy, with a simple config file for keys like OpenAI and NCBI, and include a notebook that walks through the full flow step by step so I can try it myself.
Keep it focused on evidence based medical questions, and make sure the output feels like a helpful research assistant, not a diagnosis tool. If you need current docs or examples, look them up online.
Are you gonna build this?
make sure you review the code using coderabbit