LlamaIndex 周报：LlamaExtract 公测、Corrective RAG 教程与智能体课程

Hi there, Llama Friends! 🦙

Welcome to this week's edition of the LlamaIndex newsletter! We have some exciting updates and resources to share, including a new course on building agents, a step-by-step tutorial on corrective RAG, and the launch of LlamaExtract in public beta. Dive in to discover how you can enhance your projects with these innovative tools and insights!

Explore our free and paid plans today.

🤩 The Highlights:

• LlamaExtract Public Beta: LlamaExtract is now in public beta! It allows you to define a schema and automatically extract structured data from your documents. Read the full demo post here and sign up today.
• Step-by-Step Tutorial on Corrective RAG: Check out this fantastic tutorial by @akshay_pachaar on building an agentic reasoning system for search and retrieval (specifically, corrective RAG) from scratch using LlamaIndex workflows.
• Free Courses on Building Agents: In case you missed it, we launched some high-profile free courses on building agentic systems with LlamaIndex, one with DeepLearning.AI and another with HuggingFace! Don't miss them.

🗺️ LlamaParse & LlamaParse:

• AI Voice Assistant Pipeline: Build an AI voice assistant pipeline with ROCm, LlamaIndex, and RAG on @AMD GPUs. This tutorial covers transcription, context-aware responses, and text-to-speech. Get started here.
• Large-Scale Structured Extraction with LlamaExtract: Learn how to extract 50-100 fields from complex documents using LlamaExtract. This guide covers defining schemas, using optional fields, and verifying accuracy. Explore the notebook here and sign up for LlamaExtract.

✨ Framework:

• Gen AI Toolbox for Databases: We've teamed up with @googlecloud to simplify building AI agents using the Gen AI Toolbox for Databases. Learn more about this integration and check out the quickstart.
• Unified Google AI Integration: Now there's a unified integration that supports both Google Gemini and Google Vertex AI, handling everything from streaming to structured prediction. Explore the new integration here.