What problem does it solve? Reading and translating full academic papers across PDF, DOI, arXiv, HTML, or pasted text is slow, and typical tools degrade into summary-only output that loses paragraph alignment, figure context, and source traceability. ## Core Features & Use Cases - Full-paper bilingual reader: Produces a paper.md with paragraph-level Original/中文 pairs, stable page and block anchors, and a terminology table for consistent technical terms. - Figure and table extraction: Crops figures and tables into assets/ and places them near their first substantive mention with bilingual captions and reading notes. - Source-grounded follow-ups: Maintains a source_map.json so later questions can be answered with exact page and block citations. - Use Case: A researcher provides a scanned conference paper and receives a complete OCR-based Chinese-English Markdown reader with cropped figures, uncertainty labels, and traceable anchors instead of a plain summary. ## Quick Start Ask the agent to turn the attached paper PDF into a full Chinese-English side-by-side Markdown reader with figures extracted and placed near the relevant text.
