PaperFormalizer formalizes research papers in Lean with the author in the loop
Original title: PaperFormalizer: Author-in-the-Loop Autoformalization of Research Papers in Lean
It runs in Claude Code or Codex, pauses on ambiguous claims, and emits an annotated paper plus linked Lean code.
PaperFormalizer is a Lean pipeline that starts from a paper's LaTeX, builds a dependency graph of statements, audits them in natural language, and formalizes and proves them against Mathlib, optional domain libraries, and the author's toolkit. It stops whenever a statement has more than one reading and presents each repair as a proposal the author can accept, decline, or redirect. Output is a revised annotated paper, a change report, and Lean declarations linked to the statements they formalize. The project is on GitHub, and authors can share verified papers at paperformalizer.org.
- Only the author decides changes that could alter what the paper claims.
- The author's toolkit grows as more papers are formalized and can be reused.
- It is available at github.com/Jiyuan-Tan/PaperFormalizer.
- Shared Lean papers are listed at paperformalizer.org.
From the original article
1 min read at huggingface.co
Even a carefully proofread paper can leave small gaps in its proofs or its stated assumptions. We introduce PaperFormalizer, a Lean-based agentic pipeline that helps authors find and close such gaps in their own papers and turn each paper into a Lean-verified version.
By Jiyuan Tan, Vasilis Syrgkanis, Jose Blanchet · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.