ResearchConjectureIllustrative scenario
How a research group could use Conjecture to find weak steps before reviewers do.
The challenge
Long proofs go to review with gaps that take months to surface.
The approach
- 01Conjecture splits each draft proof into claims with Lemma.
- 02Steps below 0.9 confidence are checked in Lean.
- 03Reviewers receive the proof with its checked steps marked.
The outcome
Gaps are found before submission, and reviewers spend their time on the ideas rather than the arithmetic.
Why the reasoning trace matters
In a long proof, one unjustified step can invalidate everything after it. Splitting a proof into claims and checking each one moves the search for gaps from reviewers to the authors, before the paper is submitted.
How a team could start
A group could start with one recent paper, compare the steps Conjecture flags with the points reviewers raised, and decide from that whether to use it on new drafts. Academic teams can apply for a Research Grant.
This is an illustrative scenario, not a customer account. It shows how a team could use Conjecture; we will publish real stories, with permission, as they happen.