What problem does it solve?
It solves the difficulty of validating LaTeX layout quality using only compilation logs, by generating per-page visual evidence that downstream agents can use to catch real-world typesetting defects.
Core Features & Use Cases
- PDF page rendering for visual evidence: Converts a compiled PDF into high-resolution page images (PNG/JPG) with consistent naming.
- Optional cropped region rendering: Produces targeted, high-DPI crops for tables, formulas, and other small elements when deeper visual inspection is needed.
- Structured rendering results: Outputs a JSON render report including page counts, image dimensions, and any errors to drive deterministic verification loops.
Use it after each successful LaTeX compile to enable visual diagnosis (e.g., detecting double-column voids, overflow, missing figures/tables) and to compare before/after fixes.
Quick Start
Ask your agent to render the compiled PDF pages (for example main.pdf) into data/pages at 220 DPI and produce the corresponding render report for visual inspection.