Agentic-first PDF review for LaTeX papers: live web viewer + comment queue + MCP server for Claude Code
View on GitHub ↗Review a LaTeX paper from its rendered PDF while Claude Code or Codex edits the source.
You read the PDF in a browser and leave comments on selected text, sections, or the whole paper. The coding agent reads those comments through MCP and edits the LaTeX source. The running web server detects the edit, rebuilds the PDF, and displays it in the same browser.
tex-mcp-web requires Python 3.10 or newer and a supported document compiler on PATH. LaTeX projects use latexmk by default; pdflatex, xelatex, lualatex, and pandoc are also supported.
Register tex-mcp-web once for the agent you use.
tex-mcp is the process that the agent starts for the MCP connection. You do not run it separately after registration.
Each paper needs a .tex-mcp-web.yaml in its project root. This file identifies the paper, tells the web server what to compile and watch, and lets the MCP process find the same project when the agent is working in a subdirectory.
Without this file, tex-web falls back to main.tex on port 8765 in the current directory. That fallback does not mark the project root or assign a per-paper port, so the normal MCP workflow uses .tex-mcp-web.yaml.
Enter the paper directory and create the file if it does not already exist:
--main is the path to the top-level source file, relative to the project root. init refuses to overwrite an existing .tex-mcp-web.yaml.
Field Effect ------ main Top-level source file compiled into the PDF. watch File patterns that trigger recompilation after a change. Patterns are matched against filenames and project-relative paths. ignore Paths checked befor
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/MiiKiyoshi--tex-mcp-web.html)
Stop your AI from making things up — it proposes, deterministic tools decide, every claim checked against ground truth with evidence. Grounded facts and context survive resets. Reverse engineering is
zhuyansen/new-term-radarDaily radar for new AI terms: names of new models, products and repos, caught on X, Hugging Face, GitHub and sitemaps before keyword tools list them.
yzg1226/mcd-party-officer🍟 麦当劳门店派对成团调度 Skill — 把派对从「策划」推进到「办得成」：成团概率雷达 / 包场vs拼团 / 择日黄历 / 儿童营养指数 / 招募卡与邀请函
fellowgeek/mcp-memoryAn OKF-backed Model Context Protocol (MCP) server delivering persistent long-term memory and SQLite FTS5 search for AI agents.
hirotomasato/yowesGenerate realistic teacher documents (ID cards, licenses, letters) for 13 countries via MCP.
hujjun/mcd-saver-strategist麦当劳省钱点餐助手（麦门精算师）｜麦当劳优惠券组合优化 Skill：拉取麦当劳菜单与全部可用优惠券，算出最省的点餐组合，巨无霸套餐 ¥47 → ¥21.9（省 53%）。McDonalds MCP（麦当劳 MCP）coupon optimizer · McDonalds China MCP Server · Agent Skill for WorkBuddy · 麦当劳程序员创意开发大赛参赛作品
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.