At Undermind, we're building a search engine that can handle extremely complex questions. It’s geared at experts, like research scientists and doctors, who need to find very specific resources to solve high-stakes problems.
We’ve rebuilt search from the ground up to address this. Our new approach employs high-quality LLMs to adaptively explore a database, mimicking how a human researcher carefully discovers information. This approach dramatically outperforms (by 10-50x) traditional keyword search and other modern AI-based retrieval methods.
Our first target users are the 50 million researchers searching for scientific literature on PubMed and Google Scholar every month. We’ve launched and have users paying us $200-$1000 per seat per year, from fields like medicine, ML, biotech, finance, and more. Our techniques in the future can port to even broader use cases and databases in law/consulting/finance.