DALLE3 Inpainting: Editing your images with DALL·E

Editing your images with DALL·E

You can now edit images you create with DALL·E
Updated this week

The DALL·E editor interface enables you to edit images by selecting an area of the image to edit and describing your changes in chat. You can also provide a prompt with your desired edit in the conversation panel, without using the selection tool.

You can access the DALL·E editor interface by clicking on an image generated by DALL·E.




No freaking way! That is amazing. Out & in painting was what imo made Dall-E epic.

Regardless, thanks for sharing! Awesome news. I have to try this out. I wonder if it’s charged as a completely new image :thinking:

EDIT: oh my god it’s only via ChatGPT fml


I can’t find the option on ChatGPT4. Is this available in France?

Whenever OpenAI releases anything they make it seem like it’s instantly available to everyone.

The truth is that they always roll things out. Give it some time and you’ll have it


It reminds me of Photoshop’s generative fill. It looks great, I can’t wait to get it rolled out to me.

1 Like

I happened to notice this feature by chance today, very cool. I notice that if you don’t select a region and instead just describe the edit, it seems to generate a different but mostly similar image with the edits, as opposed to the exact same image with only certain parts changed. It basically uses the same prompt as before but might change a small part of it.

So if it’s a very specific part of the image that needs to be changed, using the brush to select a region is better.

1 Like

Pretty cool, at least from a time/workflow standpoint, to not have to copy and paste previous gen_id and prompt details repeatedly for image iteration as we did before.

However, that’s all this update is actually doing if you check the data and the iterative prompts. Great for a broader audience, but I’d really like to see something new from a release that we here in the forum didn’t sort out on our own months ago……c’est la vie.

Great news here! One caveat though is that this function is not yet available on Custom GPTs though, namely, there is no way to impaint a generated image in the custom GPT (see example screenshot where the impainting function is missing when enlarging the picture). Does anyone know whether this is on the roadmap?

I’ve noticed that the ability to edit images is not working or appearing when attempting to edit images created by a GPT, only in a standard chat. Is this an oversight?

While I don’t know for certain, they’re likely getting all the wrinkles ironed out before moving it over to custom GPT ecosystem? We shall see. I’ll report back if I hear anything.

mine stopped working, and chat CPT is dodging the questions regarding it :slightly_smiling_face:

Well Said in this Thread Team you are Great!

Well Said in this Thread Community you are Great & We Are Very Glad!