LlamaIndex 周报：赞助 Agents & MCP 黑客松，LlamaParse 与 LlamaExtract 更新

Hi there, Llama Lovers! 🦙

Welcome to this week's edition of the LlamaIndex newsletter! We have some exciting updates, including our recent sponsorship of the Agents & MCP Hackathon, new features in LlamaParse, and insights into the Model Context Protocol (MCP). Dive into this week's highlights and community contributions!

Explore our free and paid plans today.

🤩 The Highlights:

• Build a Multi-Agent System with MCP: Check out @microsoft's new AI Travel Agents demo showcasing how to coordinate multiple AI agents using the Model Context Protocol, LlamaIndex.TS, and @Azure AI Foundry for complex travel planning scenarios. Read the full technical breakdown.
• Does MCP Kill the Need for Vector Search? Explore how the MCP protocol creates new possibilities for agents to connect directly to data sources while still requiring preprocessing and indexing for unstructured data. Read our full analysis.
• Build Agent-Powered Frontends with AG-UI Integration: Discover how to create agent-powered frontends with zero boilerplate using our new AG-UI integration with @copilotkit. Get started here.

🗺️ LlamaParse & LlamaParse:

• LlamaParse in Action: Watch LlamaParse in action with MCP to Claude Desktop, where we connected LlamaExtract as a local MCP tool to generate reports from financial documents. Learn more.
• New Features in LlamaExtract: We’ve launched an Auto-Schema Generation Agent in LlamaExtract, addressing friction points in automating document extraction. Discover the new feature.
• LlamaParse Presets: Our recent "preset-modes" on LlamaParse allow you to parse complex figures and charts within research reports into formatted results. Check out the details.

✨ Framework:

• Flexible Memory Blocks in LlamaIndex: We’ve introduced flexible Memory Blocks to LlamaIndex for various agent memory approaches, including fact extraction and vector memory. Read more about it.

✍️ Community:

• Hanane D's Notebook on Multi-Agent Financial Analysis: Check out Hanane D's excellent notebook about building a multi-agent financial analysis system using LlamaIndex. Explore the notebook.
• Insights from the Data + AI Summit 2025: If you missed our CEO Jerry Liu's talk, catch up on the emerging landscape of agentic document workflows. Learn more about LlamaParse.
• MCP Server Design Playbook: Block's engineering team shares their systematic approach to creating MCP servers that integrate seamlessly with Claude and other AI systems. Read the playbook.

Thank you for being part of our community! Stay tuned for more updates and insights next week.