PageIndex

Meet PageIndex

PageIndex helps people and applications retrieve information from long documents. It builds a tree index that an AI model can navigate to find relevant passages, then connects answers to source references. A chat application supports working with uploaded files and comparing information across documents, while a developer SDK exposes the retrieval workflow to other products.

The local open-source option works with text-based PDFs and a user's model credentials. Managed cloud indexing adds support for scanned files and images, hosted storage, and more precise block references; local citations stop at the page. Teams can choose the deployment that fits their document types and review the linked evidence when checking an answer.