Can combine 2 images in browser UI but not with API

API
,
1

Hi. I’m trying to use gpt-4 in order to combine some images, think of it as placing a logo onto an existing image. The idea es it that given the image and the logo gpt would give me a suggestion about the logo placement and then return an image with the logo placed.

I know this can be done, as when I tried it on the browser UI I was able to get the combined image back.

But when I tried to do by calling the API and using gpt-4 it didn’t work, and got responses like this:

  • “Since I can’t physically combine the images for you, following these steps in a photo editing software will help you achieve your desired result.”
  • “Sure, here is the combined image: (fake url)”
  • “Sure, here is the combined image: (real url with random picture)”

I also looked into the docs and it seems only dall-e 2 is able to edit images, but as far as I could tell it was not possible to give it two images in order to combine them, as the expected input is an image, a mask and a prompt.

Any help would be greatly appreciated.

1 Like
2

Welcome to the forum!

This is not doable via ChatGPT or the API.

If you want to do style-transfer, check out Deep Dream Generator.

3

But I actually had it done via the web UI. Here is the conversation:

image
image1240×412 52.4 KB

and the provided link actually downloaded the combined picture

4

Here is the image I got back (had to do another post because I wasn’t allowed to post 2 pictures on one post)

camera_with_logo
camera_with_logo640×640 329 KB

5

Ah, it might be using Code Interpreter to just paste the logo onto the image randomly. You can’t really control it, though.

6

Oh, that makes sense, thanks for the quick reply!

1 Like