XNews

Tencent WeChat Open-Sources WeVisDoc: End-to-End Page to Markdown Parser

Tencent WeChat Open-Sources WeVisDoc: End-to-End Page to Markdown Parser

Tencent WeChat vision team released WeVisDoc, an end-to-end document parser that converts page images to structured Markdown with text, LaTeX, tables, and reading order in one output. Based on Qwen3-VL, it uses a two-stage training strategy focusing on coverage first then targeted capability allocation via residual analysis.

Source posts