Today we’re excited to introduce Llama Packs 🦙📦— a community-driven hub of prepackaged modules that you can use to kickstart your LLM application. Import them for a wide variety of use cases, from building a Streamlit app to building advanced retrieval over Weaviate to a resume parser that does structured…

Today we introduce RAGs, a Streamlit app that allows you to create and customize your own RAG pipeline and then use it over your own data — all with natural language! This means you can now setup a “ChatGPT over your data” without needing to code. Setup and query a…

(co-authored by Haotian Zhang, Laurie Voss, and Jerry Liu @ LlamaIndex) Overview In this blog we’re excited to present a fundamentally new paradigm: multi-modal Retrieval-Augmented Generation (RAG). We present new abstractions in LlamaIndex that now enable the following: Multi-modal LLMs and Embeddings Multi-modal Indexing and Retrieval (integrates with vector dbs) Multi-Modal RAG One…

Authors: Avthar Sewrathan, Matvey Arye, Jerry Liu, Yi Ding Introducing the Timescale Vector integration for LlamaIndex. Timescale Vector enables LlamaIndex developers to build better AI applications with PostgreSQL as their vector database: with faster vector similarity search, efficient time-based search filtering, and the operational simplicity of a single, easy-to-use cloud…

(co-authored by Ofer Mendelevitch, head of Developer Relations at Vectara, and Logan Markewich, founding engineer at LlamaIndex) Introduction Vectara is a trusted GenAI platform. Exposing a set of easy to use APIs, Vectara’s platform reduces the complexity involved in developing Grounded Generation (aka retrieval-augmented-generation) applications, and managing the LLM infrastructure that’s…

We’ve added capabilities in LlamaIndex allowing you to fine-tune a linear adapter on top of embeddings produced from any model (sentence_transformers, OpenAI, and more). This allows you to transform your embedding representations into a new latent space that’s optimized for retrieval over your specific data and queries. …

Authored by Joe Reuter, Software Engineer at Airbyte (cross-posted from the Airbyte blog; check it out here!) Content It’s now possible to utilize the Airbyte sources for Gong, Hubspot, Salesforce, Shopify, Stripe, Typeform and Zendesk Support directly within your LlamaIndex-based application, implemented as data loaders. For example, to load the Stripe…

UPDATE 9/10/2023: We’ve included embedding finetuning abstractions into the LlamaIndex repo, so this repo is technically outdated! Please check out our embedding fine-tuning guides in the core documentation. We’ve created a comprehensive, end-to-end guide showing you how to fine-tune an embedding model to improve performance of Retrieval Augmented Generation (RAG)…

(co-authored by Jerry Liu, CEO of LlamaIndex, Jeffrey Wang, co-founder at Metaphor, and Adam Hoffman, Software Engineer at Hypotenuse Labs) We’re incredibly excited to launch an integration between LlamaIndex and Metaphor: combine the capabilities of LlamaIndex data agents with Metaphor as a native LLM search tool to enable knowledge workers…