1 comment

[ 3.2 ms ] story [ 15.9 ms ] thread
How does this work under the hood? Is it OCR plus semantic mapping, or something else?