Can I upload photos into chat gpt plus whilst using DALLE 3?

Hi there,

I would like to take a castle I have designed myself, and fuse it into the book cover design as shown (this was made in DALLE 3, bing).

So if I upgrade to chat gpt PLUS, would I be able to upload the designs and ask it to do this this? I’m not sure if there is a photo upload option when using DALLE 3 in chat GPT. If anyone knows of another way to do this, I’m all ears, I’m thinking of perhaps exploring Midjourney.


1 Like

Welcome to the forum.

DALLE (and Midjourney / Stable Diffusion) won’t be able to do what you’re describing. You’ll need to cut it out and replace yourself manually in Photoshop or similar.

Hope this helps.

1 Like

Hi @lizbass
“So if I upgrade to chat gpt PLUS, would I be able to upload the designs and ask it to do this this?”

Yes but anything you create within ChatGPT4 is not accessible via Dall-E therefore it would make more sense to explore the Dall-E route.
I assume you can crop what you’ve got so then you have a Dall-E base image - so then Dall-E will do what it is told (within reason… lol)

Ref “photo upload option when using DALLE 3 in chat GPT.”
…sure add whatever file you like, it will appear in the window…

Whist it would seem Dall-E is a go for this nonetheless it may be interesting to explore the prompt actions / implications / results via GPT4 since the create / regen times are pretty impressive.

I may be wrong but imho you can take this to Dall-E and merge it with another creation. I’m happy to prove myself wrong, if you’d like to give me two dummy images.


1 Like

Uploading: 1.png…

You crop the image in this way, you correct the faults, notably the font,… and you search the internet for book creation and you attach your different creations generated by dalle 3. you can subscribe to gtp plus which will give you access to slab 3 but when you ask for an illustration, you must not mention the word ‘book’ in your request. otherwise he will systematically create an illustration for you in a book.

1 Like

I asked slab 3 to generate an illustration consistent with your image, he gave me that. So you can do it as many times as you want, until you generate the desired illustration.

1 Like

Just a ten minute Photoshop hack job…

Hi there, thank you for your time. I looked up slab 3 but nothing came up, what is it?

Hey there, thanks! really cool to see that it might be possible in PS. Can I ask which effects/techniques you used? I’m thinking if I can find an AI that is able to match the style of the book cover to my castle, I could then photoshop it on as you have just done, to get the a closer match.

You know that story about the mechanic that billed $500 for “the right bolt”…

I started off by matting. Selecting most of the castle and inverting, then playing with gray levels to make most background transparent, then erase tools, more blending.

Then worked on the background, did some cloning, and then just did some shrinking and liquifying to make the old castle tiny.

Then needed to get in there with sizing and perspective on the new castle, making it’s base match up with the unreal drawing’s perspective, a bit more background tweaking.

Then colors, lots of colors and levels, filters, making it not too purple, not too yellow, not too ghastly. Throw some chunky grain in there, and more color selection, hues, etc.

Then another layer of edge extraction, growing the edges and colorizing that overlay to not detract from what I’d already done. But these words adding to the work… I can see I could have put more greens into the background to make it easier.

One thing you can do with AI is use the API edits endpoint. You make part of the image transparent, and then you get AI rendering that transparent area which fills that back in or outfills, with very loose prompt following. It’s about the worst of AI imagery you’ll get, but can be a technique to put a blue sky back where a deleted castle was or such.

OK, thank you for that, very kind of you to take the time to do it and also explain your process. If I cant find a way to do it with AI, at least I know something might be workable in PS! :smiley:

1 Like

Hi Liz,
Apologies, what with Christmas / New Year et al I never got back to this.

The very nice images (imho) generated by bramskiller1 are based on :
“A whimsical fairy tale landscape inspired by a fantasy kingdom, with majestic castles featuring bright red roofs and soaring white towers./”
the reminder of the prompt is truncated so I cannot see it.

No2: “A whimsical fairy-tale castle sitting atop a lush green hill, with a flowing blue river wrapping around its base. The castle is majestic, with bright…”
ditto ref the truncation.

Please forgive me, I cannot remember what your original question was?


cant you just use photoshop or even canva to do this? its a mockup you want i believe.