# Parsing documents with AI agents just got a lot more seamless🚀 We've rebuilt the LlamaParse MCP se... Canonical URL: https://www.traeai.com/articles/172c754d-8283-40e2-8871-20fd38427a16 Original source: https://x.com/llama_index/status/2049519248490606809 Source name: LlamaIndex 🦙(@llama_index) Content type: tweet Language: 中文 Score: 7.2 Reading time: 1 分钟 Published: 2026-04-29T16:00:41+00:00 Tags: LlamaIndex, MCP, AI Agents, Document Parsing, LLM Infrastructure ## Summary LlamaIndex 重构 LlamaParse MCP 服务,支持文档解析、分类、分段与多方式上传,解决 OAuth 集成、文件上传缺失、可观测性等生产级挑战。 ## Key Takeaways - LlamaParse 现以 MCP 协议标准服务形式提供,兼容任意 MCP 客户端 - 新增文档结构化能力:Markdown 解析、自定义分类、语义分块与标签化切分 - 生产部署中攻克了身份认证对齐、无原生上传支持、速率限制与可观测性集成难题 ## Outline - 发布概览 — 宣布 LlamaParse MCP 服务完成重构并开放连接。 - 核心能力升级 — 支持 Markdown 解析、文件分类、语义分块和 URL/浏览器双上传通道。 - 生产挑战与解法 — 围绕 Auth(WorkOS)、上传(token 设计)、部署(Vercel)、可观测性(Axiom)展开工程实践。 - 技术落地细节 — 涵盖 OAuth 流程、token 化上传机制及关键权衡取舍分析。 - 资源入口 — 提供博客链接与 GitHub 仓库地址供深度查阅。 ## Highlights - > We've rebuilt the LlamaParse MCP server to handle your document processing workflows — 首句 - > Building a production MCP server surfaced some non-obvious challenges: getting auth to align with an existing platform identity system using @WorkOS — 中段 - > working around MCP's lack of built-in file upload support, and making deployments, rate limiting and observability feel native with @vercel and @AxiomFM — 中段 - > Parse documents into clean markdown / Classify files against your own categories / Split long documents into labelled sections — 功能列表 ## Citation Guidance When citing this item, prefer the canonical traeai article URL for the AI-readable summary and include the original source URL when discussing the underlying source material.