@baokui

pdf-ocr-layout extracts tables and charts from documents via glm-ocr visual parsing, then routes extracted content to glm-4.7 (text) or glm-4.6v (images) for semantic analysis. returns json with bounding boxes and interpreted meaning.

wan-text2image wraps alibaba dashscope wan2.6-t2i model for text-to-image generation via bash script. supports prompt extension, negative prompts, and three fixed image dimensions. requires pre-configured api key.