The Optical character recognition (OCR) models we use work best if the language
they are detecting is known. Type in the
language box below to narrow down the range of available languages and click
the most appropriate option. If you do not select
a language, English will be used by default.
To reduce noise, the output of the OCR process itself is filtered to
remove some extraneous characters, and
for the following languages a spellchecker is also run on the output to improve it.