Vision for gemini models through api?

Gemini models support vision. but when I try to send an image:

is this configuration related?

it’s probably been 3 months or so that people are asking this question everywhere.

why not?
and why only claude for agentic? gemini also supports that.

is the pipeline different? why don’t you support this??

more importantly

WTF you guys don’t answer?

1 Like

This happens to me too… and Ive tried all all the Gem models…

Hey, we don’t internally have support for image capabilities on the Gemini models yet, but I’ll log this down to see if we can add it!

this should be relatively easy to implement? everything that I do with gemini just supports images if I attach them as base64 and works out of the box.

is there some hard-coded structural issue that makes cursor devs focus solely on claude (for agent flows too)?

is Anthropic paying you :smiley: (joking)

I know that cursor is not open source. but it’s obvious that it’s growing and maybe you could allow people like me from the community to chip in, or at least give us a way to extend / modify cursor’s capabilities with some kind of scripting / api etc?

Hey, our team is still small (but growing) and as such, we have to prioritize heavily on features that benefit the majority of our users.

While I appreciate the implementation may seem trivial, it’s more about what would be delayed elsewhere in the editor that makes it hard to fix or improve on features like this, especially when other models already have image capabilities built in!

Regardless, I will speak to the team about seeing if this can be enabled.

1 Like