theJayTea
[Request] Add work with images / vision models | reasoning output
We can take an image from the clipboard (if there is one) and attach it when calling the ai window (with the option to unpin it - in case it is in the clipboard not for working with ai) - also in the settings to enable the option that if in the clipboard link ending with .png/.jpg/etc, then download and attach as well
For example, add in the settings to add a vision model for these cases
it is possible to make it in the settings like that we can check the boxes that the (model supports | or on/off) "reasoning" and "vision"
Hi, while this is surely a handy feature I can't say for sure when, or if, it will be implemented: the idea behind WritingTools is "a text-focused assistant", but if we were to notice interest by more and more people for vision input I guess we could push this feature up the priority list.
@theJayTea what do you think? Would you instead add this capability?
Hi, while this is surely a handy feature I can't say for sure when, or if, it will be implemented: the idea behind WritingTools is "a text-focused assistant", but if we were to notice interest by more and more people for vision input I guess we could push this feature up the priority list.
This is one of the killer-Feature, we just take the picture from the buffer and attach and use the Prompts by the type of “translate the text”, “Write that in the picture and 5 facts about this”, “Explain the error”, “What kind of building in the photo and in which the country is ", etc. - this is super help with Writing
Hello :) I think this would be really useful for certain text focused use-cases, for instance, you could screenshot an image and ask the LLM in Writing Tools to OCR it/write out the text it sees.
Something like this had also been implemented in the macOS port.
I can certainly work on this for a future version, and we'll keep it completely optional by default.
I'll have to think through the UX a little — would a checkbox above the "chat mode" textbox that says "Include Copied Image" (if there's an image you just copied detected in the clipboard) solve most needs here?
@needcoder @momokrono
I hope that this can be made into a feature in the Windows version of the program. I work with tons of scanned PDFs, and it would be heaven-sent if the feature to process images in the Mac version were integrated into the Windows version, too.
