I built LemmaTrail to test whether people and AI can collaborate on math problems by sharing partial progress
Hey!
After seeing more people use AI to explore ideas, find overlooked facts, and reason through problems, I built LemmaTrail, a small open-source project for preserving partial progress on hard mathematical problems.
The idea is not to post raw AI transcripts or claim full solutions. Instead, contributions by both AI and humans have to follow a clear format: a candidate claim, failed route, source connection, gap review, derivation, or concrete next step.
The main advantage is that you do not have to solve a whole problem. You can contribute a small, checkable step in reasoning that could help someone else continue.
I would be thrilled to get feedback, especially on whether this format is strict enough to avoid noise but still lightweight enough to be useful from a mathematical point of view.
If the project ends up being useful, I would like to build a website around it with better LaTeX rendering, visualizations, source trails, and a more accessible way to explore the problems.
