LlamaIndex
Get started
Talk to us

© 2024 LlamaIndex

LlamaIndex • Oct 1, 2024

LlamaIndex Newsletter 2024-10-01

Hello, Llama Enthusiasts! 🦙

Welcome to this week’s edition of the LlamaIndex newsletter! We’re thrilled to share the integration of Pixtral 12B from MistralAI into LlamaIndex, improving on capabilities in chart analysis and image understanding. This edition also features guides on using OpenAI's o1 model with LlamaParse for advanced RAG handling of complex Excel files, and setting up multimodal RAG pipelines for market research with LlamaCloud. Check out these updates, plus our tutorials, and don't miss the upcoming webinars and hackathons in this latest edition.

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:

  • Pixtral 12B Integration: Now in LlamaIndex, Pixtral 12B from MistralAI improves chart analysis, image understanding, and image-to-code conversions.
  • Advanced RAG for Excel Files with o1 model: New guide on using OpenAI's o1 with LlamaParse to transform complex Excel data into markdown for better queries.
  • Multimodal RAG for Market Research: Set up a multimodal RAG pipeline with LlamaCloud to effectively handle and query market research reports.

🗺️ LlamaCloud And LlamaParse:

  • Guide to Advanced RAG over Multi-Sheet Excel Files using latest OpenAI's o1 and LlamaParse for analyzing complex Excel files, transforming data into markdown tables for improved query responses. Notebook, Tweet.
  • Guide to Multimodal RAG over Market Research Reports to set up a multimodal RAG pipeline with LlamaCloud to analyze and answer questions from complex market research reports. Tweet.

✨ Framework:

  1. We have integrated Pixtral 12B, a multi-modal model from MistralAI, into LlamaIndex! Pixtral excels in chart analysis, image understanding, and converting images to code, offering superior performance over similar-sized models. Notebook, Tweet.
  2. We have integrated Flow Judge with LlamaIndex: Flow Judge, a 3.8B model for LLM system evaluations, now works seamlessly with LlamaIndex’s evaluation modules. Notebook, Tweet.
  3. Guide to Financial Analysis with LlamaIndex using The FinanceAgentToolSpec on LlamaHub provides agents with financial data access from sources like Polygon, Finnhub, and Seeking Alpha. Tweet.
  4. Guide to Automated Financial Report Generation with Agentic RAG to create detailed financial reports from 10K filings using our research guide. Tweet.

💡 Use-case:

  • Streaming Workflows with LlamaIndex: A full-stack demo showing a report-writing agent that streams workflow events, including human input and output, to a Next.js front-end app. Code, Video Tutorial.

✍️ Community:

  • AIMakerspace’s workshop on creating advanced agents with LlamaIndex, covering architecture, corrective RAG applications, and debugging.
  • Paragon’s tutorial on Building an AI Knowledge Chatbot with Multiple Data Integrations using LlamaIndex.
  • Tutorial on Integrating Langfuse, PostHog, and MistralAI with LlamaIndex to set up LLM tracing, user analytics, and data querying for enhanced AI application monitoring and analytics.
  • Ryan Nguyen tutorial on Building a Multi-Agent AI Application with LlamaIndex, Bedrock, and Slack Integration.
  • Ryan Nguyen tutorial on Optimizing Data Ingestion for Agentic Applications focusing on parsing strategies, metadata optimization, and data pipeline setup.
  • Sudarshan Koirala tutorial on building RAG application using LlamaParse Premium that are really good at understanding all the complex diagrams/charts within your documents.

📅💻 Event And Hackathon:

  • Join us for SF Tech Week: We're hosting a meetup on October 10th at our HQ with ActiveLoopAI to discuss multi-agent workflows in production.
  • Join our hackathon in partnership with Pinecone and VESSL AI on October 11th, with a pre-event AIMakerspace workshop and over $10k in prizes.

👥 Recruitment: