1 comment

[ 5.4 ms ] story [ 17.5 ms ] thread
How does this work under the hood? Is it OCR plus semantic mapping, or something else?