[Feature Request] The Agency Needs You: Support image source
Opened this issue · 1 comments
Hi, we are a certain Secret Service agency and are in need to process large amounts of sensitive data. Our main source of intel are photographs our spies take during missions. This has proven to be far more efficient than having to copy all the text to their little notebooks.
However, it would be great if you little library accepts digital images as well so our spies can feed their gathered intel into this lib to determine whether some numbers are odd or even. Our spies have heard rumors about so called multi modal LLMs, if it helps we can steal the details and pass it to you so you can implement this feature?
This could also have more uses than just for checking numbers. Being able to input any picture (not just of text) and asking "is this odd?" is probably a very useful feature for spies. Odd pictures require deeper investigation.