LlamaIndex Newsletter 2024-12-24

Greetings, LlamaIndex Community! 🦙

Welcome to this week’s edition of the LlamaIndex newsletter! Highlights include the LlamaReport Preview launch, enabling quick and smart document-to-report conversions, Audio Parsing with LlamaParse, now supporting audio-to-text and handling PDFs and spreadsheets, and a series of Agentic Document Workflows Notebooks showcasing use cases like simplifying domain-specific logic, processing insurance claims, and standardizing invoice data. Plus, don’t miss out to check our latest tutorials from the community.

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.

As we kick off this week's edition, we want to share our year-end review for LlamaIndex, celebrating milestones like tens of millions of pages parsed, millions of framework downloads, and a growing base of happy users. Check it out here.

🤩 The Highlights

• LlamaReport Preview Launched: Quickly turn document databases into reports with smart features, easy templates, and tools for developers.
• Audio Parsing with LlamaParse : Now supports converting audio to text, along with handling PDFs and spreadsheets.
• Agentic Document Workflows Notebooks Released: Explore LlamaCloud and LlamaParse powered use cases:
  • Simplifying domain-specific business logic: Extract and match product details. Notebook, Tweet.
  • Simpler Insurance Claims: Process claims and suggest settlements. Notebook, Tweet.
  • Standardize Invoice Units: Align data from different vendors. Notebook, Tweet.

🗺️ LlamaCloud And LlamaParse:

• We’ve launched LlamaReport in preview, turning document databases into customizable reports in minutes with intelligent processing, flexible templates, LLM-powered editing, and a developer-friendly API. Blogpost, Tweet.
• We have launched audio parsing in LlamaParse! Alongside handling complex document formats like PDFs and spreadsheets, it now converts speech into text effortlessly. Upload an audio file, and let LlamaParse do the rest. Tweet.
• We have released a series of notebooks on building agentic document workflows:
  1. Simplifying domain-specific business logic: Learn to process invoices, extract data into schemas, and match SKUs/products using vector retrieval with LlamaParse. Notebook, Tweet
  2. Automating auto-insurance claims: Parse claims, retrieve policies, and generate settlement recommendations efficiently using agents.Notebook, Tweet
  3. Standardixe units across invoices: Standardize units and measurements across vendors and integrate extracted data into workflows with LlamaIndex. Notebook, Tweet

✨ Framework:

• We have introduced AI-powered IT support for web apps—no coding, no call centers, using Gemini 2.0 Flash and LlamaIndex RAG to revolutionize enterprise support. Tweet.

💡 Tutorials:

• MistralAI’s guide to build an ArXiv research agent for search, retrieval, and PDF management using tools like RAG and ArXiv integration.
• Guide to building multi-agent systems with LlamaIndex, showcasing the Orchestrator pattern, error handling, and data validation through an Anki flashcard generator example.
• Vectara's guide on showcasing its RAG capabilities, covering data loading, streaming, reranking, chat, and agentic RAG applications.
• Tarun Jain’s crash course on building agents, covering function calling, agentic RAG, ReAct agents, and a real-time search assistant.