Curie - Assistant for scientific literature and document analysis.
by•
Curie is an AI research assistant built for real scientific work, not a generalist chatbot. It searches PubMed, arXiv, Europe PMC, OpenAlex and Semantic Scholar in parallel, then verifies every claim against the source text, so you see what's backed and what isn't. Point it at a document to extract structured data with the quote behind each value, or run a full systematic review, protocol, screening, PRISMA diagram, with you approving every decision. Projects and a library keep it organized.
Replies
Hey everyone!
I'm the developer behind Curie, and I'm excited to share that it's my first solo launch on Product Hunt, and one of the first real products I've built.
Curie has been over a year in the making, built together with the University of Barcelona and in collaboration with several research institutions across Europe. We opened it up for everyone to try this September, and I wanted to share where it actually came from.
Working alongside with Biost3 (a bioinformatics, data science, and statistics research team) and watching the researchers day in and day out, we realized one of the most exhausting parts of their job was reviewing dense scientific papers. Drawing real conclusions meant holding entire articles in their heads at once, a massive cognitive overload.
On top of that, what they desperately needed was information they could actually trust: answers checked and verified against real scientific databases, not just whatever plausible hallucination a generalist AI churned out. Most tools aren't even connected to actual scientific literature, but research demands rigorous biological interpretations grounded in peer-reviewed evidence. Seeing the researchers in my unit struggle with this exact friction every single day made it clear that generic AI wasn't going to cut it. That struggle is the exact reason why Curie was born.
From there it kept evolving: searching across multiple scientific databases at once, verifying every claim against its source, extracting structured data with the exact quote behind each value...
Would love to hear what you think. Any questions, feedback or suggestions to improve Curie are genuinely welcome!
A hug,
Víctor