auto-paper-improvement-loop

Automate LaTeX paper improvement by running two independent GPT-based review cycles that convert critique into targeted fixes, then recompile the document with latexmk to ensure it builds cleanly.

Fix writing inconsistencies and overclaims in a generated manuscript by applying a post-write quality hardening loop, where an LLM reviewer identifies issues and applies targeted LaTeX edits across two rounds.

Recompile LaTeX automatically after applying review fixes by using latexmk execution-grounded quality control, which verifies successful compilation and reduces the risk of broken references after each edit round.

Yes, you can restrict edits to specific files for camera-ready or text-only microedits by applying an optional edit whitelist, enforcing constraints on which files and edit types are allowed during the review loop.

Yes, you need a compiled PDF to run the automated LaTeX paper review loop, as the process requires a paper directory containing main.pdf and LaTeX sources to begin the improvement cycles.

Limitations of using LLM review for academic writing quality assurance include dependency on successful latexmk recompilation after each round and the need for fresh context-naive reviewer rounds to avoid compounding errors.