LlamaIndex Newsletter 2024–02–20: introducing LlamaCloud

Hi there, LlamaIndex Enthusiasts 🦙,

Today marks a milestone for the LlamaIndex ecosystem with the introduction of LlamaCloud, a next-generation suite of managed parsing, ingestion, and retrieval services tailored for production-grade context augmentation in your LLM and RAG applications.

As an enterprise AI engineer using LlamaCloud, you can concentrate on crafting the business logic, leaving the heavy lifting of data management to us. Process vast amounts of production data effortlessly, enhancing response quality instantly. LlamaCloud debuts with:

• LlamaParse: A specialized parsing service for complex documents, including tables and figures, seamlessly integrated with LlamaIndex for handling semi-structured documents. This enables answering intricate queries previously out of reach.
• Managed Ingestion and Retrieval API: Simplify data loading, processing, and storage for your RAG applications, supported by over 150 data sources via LlamaHub, including LlamaParse, and more than 40 data storage solutions.

LlamaParse is now in public preview, with a current focus on PDFs and a usage cap for public users; contact us for commercial terms. The managed API is in private preview, and available to a select group of enterprise partners. Interested? Reach out for more details.

Your inventive spirit is our driving force! We look forward to discovering the projects, articles, or videos that excite you. Send your outstanding contributions to news@llamaindex.ai. If you haven’t yet, join our newsletter via our website to get all the newest LlamaIndex news directly in your inbox.

🤩 The highlights:

1. Corrective RAG LlamaPack: We have launched LlamaPack with CRAG insights, refining information retrieval for enhanced accuracy and relevance. Tweet, LlamaPack.
2. SELF-DISCOVER LlamaPack: We have launched SELF-DISCOVER paper implementation as LlamaPack, leveraging meta-reasoning in LLMs for adaptive, complex problem-solving. Tweet, LlamaPack.
3. RAG Production Guide: A comprehensive guide to production-ready RAG, featuring insights and strategies from Sisil Mehta at JasperAI.

✨ Feature Releases and Enhancements:

• We have integrated insights from the Corrective Retrieval Augmented Generation (CRAG) paper as LlamaPack, enhancing our system’s ability to evaluate and refine retrieved information for more accurate and relevant responses. Tweet, LlamaPack.
• We have integrated SELF-DISCOVER a novel approach as LlamaPack to empower LLMs with meta-reasoning, allowing them to self-discover and adapt the most suitable reasoning modules from a selection, enabling more versatile and complex problem-solving capabilities. Tweet, LlamaPack.

🎥 Demos:

• DanswerAI, an out-of-the-box ChatGPT integration for enterprise knowledge, enhances efficiency across sales, IT, engineering, and customer support teams by connecting to common workplace tools like GDrive, Slack, and Jira, powered by Llama Index.
• GenAI for ADU Planning: A comprehensive app by Rujun Gao that navigates local ADU regulations, analyzes buildable space via satellite imagery, offers floor plan suggestions, and connects users to local contractors, showcasing the power of multi-modal AI automation in enhancing productivity

🗺️ Guides:

• Guide to building production-ready RAG covering practical tips and tricks inspired by Sisil Mehta from JasperAI.
• Guide to Nomic-embed-text-v1.5 with LlamaIndex: Get embeddings of any dimension from 64 to 768, inspired by Matryoshka Representation Learning.
• Guide to creating a RAG-powered research agent, notebook on building a workflow for scientific investigation, leveraging ArXiv, Wikipedia, textbooks, and more, with capabilities for fetching abstracts, generating ideas, and comprehensive information lookup, powered by Llama Index and LionAGI.

✍️ Tutorials:

• Jerry Liu tutorial on Introduction to LlamaIndex v0.10.
• Jerry Li tutorial on Building Agents from scratch using Query Pipelines.
• Ravi Theja tutorial video on Building Multi-Modal applications with Ollama and LlamaIndex.
• Brett Young tutorial on Building a RAG-Based Digital Restaurant Menu with LlamaIndex and W&B Weave.
• Raghav tutorial on MultiModal RAG for Advanced Video Processing with LlamaIndex & LanceDB.
• Tech With Tim tutorial on querying unstructured data, analyzing tabular data with Pandas, and actioning results in a concise, step-by-step approach.
• Florian June’s tutorial on using RAGAs and LlamaIndex for RAG evaluation.
• Davide Gallitelli’s tutorial on Deploying a HuggingFace embedding model to Amazon SageMaker and consuming it with Llama-Index.

🎥 Webinar:

• Webinar with Henry Heng, co-founder of Flowise on building advanced no-code RAG apps over your data.