On this page
LlamaIndex: agent building features and setup
A framework for building agents and workflows around data and documents. It provides tools for adding context to a model and integrating information retrieval into a process.
Tool by the Otomatzia editorial team · source content checked on
Overview
LlamaIndex focuses on connecting a language model to the information an application needs. You can build document retrieval as an agent tool and assemble an event-driven workflow with additional steps around it. It is especially relevant to products where answers depend on business information outside the model's general knowledge. The framework is separate from managed services such as LlamaCloud, so you need to choose which components to run yourself and which to use through an external service.
Features
- Building agents that use information retrieval as one of their tools.
- Event-driven Workflows that combine data, tools, and agents.
- A range of readers and connectors for data sources and documents.
Business uses
- Searching project documents and answering questions with references to relevant information.
- Extracting information from several documents and passing it to a structured review process.
Getting started
- Choose a set of documents and define test questions with known answers.
- Build a small retrieval process and check which passages are selected.
- Add an agent or workflow only after the retrieved information is reliable and relevant.
What to check
- Check the framework's terms separately from the cost and data policies of managed processing services.
- Document indexing must preserve the original access boundaries, especially in a store shared by multiple customers.
Sources to check next
Vendor features and prices change. Confirm details on the linked official pages before you rely on them.