Overview
When Cardinal extracts figures and images, the Markdown output now contains numbered placeholders like:image_metadata
field of the API response.

How to Enable
To enable this feature, setimageMetadataDetect: true
in your API request. By default, this parameter is set to false
.
How It Works
- Markdown output: contains
[Figure N]
placeholders (e.g.,[Figure 1]
,[Figure 2]
) - API response: contains an
image_metadata
array with entries that match those placeholders by ID
image_metadata
entry includes:
figure_id
— numeric ID that matches the Markdown placeholdercaption
— short description or label for the figurecropped_image_url
— direct link to the cropped figure imagemetadata
— additional context (format, description, brand info, etc.)bounding_box
— both original and pixel coordinates of the figuresubfigure_count
— number of subfigures (if any)processing_times
— timing breakdown of how long each step took
Note: By default,cropped_image_url
contains temporary URLs to the extracted images. To prevent URL generation and keep images ephemeral, set theephemeral: true
flag in your API request.