When you parse a document with the API, the extracted data is returned as Markdown following the format below.

# Name of Element 
Description of element. 
comment that describes the location: <!-- title, from page 0 (l=0.074,t=0.112,r=0.444,b=0.147), with ID 83765a10-0d93-4b1d-9473-83c90ad10798 -->
Corresponding JSON ObjectDescriptionExample
chunk_typeThe element typetitle
groundingThe page number (starting at 0) and relative coordinates of the element on that pagefrom page 0 (l=0.074,t=0.112,r=0.444,b=0.147)
chunk_idThe unique ID assigned to the chunkwith ID 83765a10-0d93-4b1d-9473-83c90ad10798