Skip to main content
A representation is an alternative asset for a file stored in Box. These assets can be PDFs, thumbnails, or text extractions. Representations allow you to access content in different formats without manually converting files, such as by extracting text from documents to send to language models for processing. Representations are automatically generated for the supported file types, either when uploading to Box or when requesting the asset. These representations are exposed through the GET /files/:id endpoint by using the fields=representations query parameter and the x-rep-hints header. The workflow involves:
  1. Getting the list of available representations for a file.
  2. Checking a specific representation’s status and retrieving its url_template.
  3. Downloading the representation if available.
The Box Python SDK provides the file.get_representation_info(rep_hints) method to access these capabilities.