LlamaIndex
Talk to us

LlamaIndex Sep 18, 2024

LlamaIndex Newsletter 2024-09-18

Hello, Llama Fans! 🦙

Welcome to this week’s edition of the LlamaIndex newsletter! We’re excited to introduce multimodal capabilities in LlamaCloud, now enabling you to develop RAG pipelines that efficiently handle both text and image data across diverse documents. This week also features integration of OpenAI's latest o1 model and a showcase of our new Code Generation Agent Template for the NeurIPS AI Hacker Cup. Plus, don't miss our upcoming tutorials and webinars designed to help you make the most of these new features

If you haven't explored LlamaCloud yet, make sure to sign up and get in touch with us to discuss your specific enterprise use case.

🤩 The highlights:

  • Multimodal Capabilities in LlamaCloud: Launched multimodal capabilities in LlamaCloud allows for the creation of end-to-end RAG pipelines for diverse data types like marketing decks and legal contracts. Blogpost, Notebook.
  • OpenAI o1 Model Integration: Day-0 integration of OpenAI's latest o1 model with our Python and Typescript LlamaIndex frameworks. Tweet.
  • Code Generation Agent Template Launch: In partnership with Weights & Biases for the NeurIPS AI Hacker Cup, we've launched a Code Generation Agent Template using MistralAI and LlamaIndex's event-driven workflows. Notebook, Tweet.
  • Code Generation Agent: Introduction of a code generation agent that utilizes the latest o1 model to develop a complete web app styled as an iPhone scientific calculator, powered by LlamaIndex.TS workflows, suitable for complex agentic workflows. Code, Tweet.

🗺️ LlamaCloud And LlamaParse:

  • We have launched multimodal capabilities in LlamaCloud, enabling you to quickly build end-to-end multimodal RAG pipelines for various unstructured data types like marketing decks, legal contracts, and financial reports. Simply enable 'multi-modal indexing' in LlamaCloud to index each page as both text and image chunks. Validate your pipeline via our chat interface with image citations or integrate it into your applications via API. Blogpost, Notebook.

✨ Framework:

  1. Code generation agent that uses the latest models like o1 to create a complete web app styled as an iPhone scientific calculator, powered by LlamaIndex.TS workflows, can serve as a standalone application for more complex agentic workflows. Code, Tweet.
  2. Guide to LlamaIndex Workflows: Build and orchestrate agentic applications using our framework that supports custom use cases, from simple LLM routing to complex ReAct agents. Our workflows are event-driven, composable, flexible, code-first, easily debuggable, observable, and ready for production deployment with llama-deploy. Examples, Tweet.
  3. We have launched a Code Generation Agent Template for the NeurIPS AI Hacker Cup in partnership with Weights & Biases, using MistralAI and LlamaIndex's event-driven workflows. Notebook, Tweet.
  4. We have made day-0 integration of OpenAI’s latest o1 model with Python and Typescript LlamaIndex frameworks. Tweet.

✍️ Community:

  • Marco Bertelli’s tutorial on Multimodal RAG with Product Manuals using LlamaIndex and LlamaParse.
  • Vivek Maskara's tutorial on how to build and test a LlamaIndex RAG app using CircleCI for continuous integration.
  • Ravi Theja’s tutorial on advanced Excel parsing abilities in LlamaParse, including multiple sheets and complex tables.
  • Multi document Agentic RAG tutorial using LlamaIndex and LanceDB.
  • Lauri Voss’s talk on Agentic RAG in 2024, discusses the essentials of LlamaIndex and RAG, why RAG alone isn't enough, and how to improve it with agentic strategies.

🎤Webinars:

  • Join Sourabh Desai in a webinar with Qdrant Engine on building advanced AI agents using LlamaIndex, set for September 26, 2024, at 10:00 AM PT / 1:00 PM ET / 7:00 PM CEST.
  • Join us at the AWS Loft on September 20th for a workshop on building an AI Scheduler using Zoom, LlamaIndex, and Qdrant.

👥 Recruitment: