Trying to submit images without a vision-enabled model selected

Discussion
1

This one drives me crazy.

Trying to submit images without a vision-enabled model selected

I will frequently drop from “Agent” to “Ask” to get an analysis from Deepseek R1 (a remarkably cool tactic if you haven’t tried it), then I’ll shift back into an Agent model to apply suggestions that I like. Deepseek is professorial… really good at spotting holes and potential enhancements but terrible at implementing them in a realistic way. But I digress…

Hey Cursor… that image was like 3-5 chats ago REALLY? That’s the restriction against doing this chat?

Give me a way to context dump, or just skip submitting images to the non-image compatible model and proceed with a warning about context thinning.

But for the love of all that is binary… just stop with the hard rejection of teh prompt.

2

Hey, since all the images added earlier are in the context window, the solution is either to delete these images from your previous requests or start a new chat.