Skip to main content
Version: 2.0

Vectara and LlamaIndex

LlamaIndex is an LLM orchestration framework that focuses more on the data aspect of building LLM-based applications. LlamaIndex provides a different set of abstractions and tools than LangChain, focusing more on data quality.

Vectara is integrated into LlamaIndex as a ManagedIndex, providing a natural interface for an end-to-end RAG-as-a-service platform. For LlamaIndex developers, using Vectara provides a powerful semantic retrieval engine and end-to-end RAG capabilities without having to integrate additional components like a vector database, an embedding model or even an LLM.

Integration benefits​

Similar to LangChain, integrating with Vectara from a LlamaIndex application provides the following benefits:

  • Powers semantic search or a full RAG pipeline, and removes the need to integrate with and maintain additional components - all is handled in a scalable manner through Vectara’s platform.
  • Offers tight integration with the LlamaIndex ecosystem for low latency and reduced cost.
  • Enables enterprise grade scalability, security, and privacy features right out of the box.

This example notebook includes code examples showing a few ways to use Vectara in a LlamaIndex application, including RAG, semantic search, as well as AutoRetrieval.