#community-help
1 messages · Page 6 of 1
thanks!!
Is it possible to share all four generations at once or can I only share them one at a time
For now only one at a time :)
bummer
Question: How to generate seamless textures? That would open up a huge world of possibilities.
Seems like that would just be a matter of treating the canvas as continuous/wrapped around, instead of having edges.
Is there anybody willing to create a 16x9 wallpaper for me with payment cause I don't have photoshop
anyone know how to change the password to your openai account?
The best way to do this would be by reaching out to the support directly 
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com
You can use the method mentioned above or simply use the forgot password option to reset your password! 🙂
oh duh, good call
Oh yeah! Sorry, i read that as E-mail @tulip zodiac
i need to change my email too so both are helpful 🙂
@normal star hello
This is confusing
Is there a way I can upload a picture and generate it on a style I pick?
As of now this is not possible, maybe in the future though.
I need a pdf of this stat. Just got done reading the entire book and yeah, this is going to change the world of art, forever. We’re only scratching the surface right now.
Hi, I recall reading somewhere that DALL E 2 can be used to expand images. Can someone tell me how I can do this?
Take your image, go to #dall-e-bot and upload it with "!pad" the bot should give you an image to download
Take that image and go to DALLE and upload it and press edit, make sure to remove the watermark
Type up whatever you want around the image or just whatever the image is
Results
Can we only remove the watermark by cropping it out
You can download the image without the watermark by using inspect element on the expanded image in the results page
thank you so much! worked great!
Yep
Only condition is you don’t mislead people into believing you created the artwork yourself
I joined the waiting list for dallE 2 but now I can use dallE 1 I guess?
This is Dall-E 2 :)
i was excited thinking i could use Dall-E 2 from the email, but when i went there and logged in it seems im using Dall-E 1?
Hey how many photographs can i generate as a beta member?
As I said, this is Dall-E 2. You can access it on https://labs.openai.com

Have a look at this :)
#🌄┃general-art message
oh ok. awesome!
So 1 credit means one search or one publish?
1 credit = one prompt
Any time you press variation , do an edit, or submit a prompt.
how do i check my credits?
Profile icon top right
Sad, is there any subscription plan or this is just it?
Q: How DALL·E Credits Work
• You get 50 free credits your first month.
• 15 free credits will replenish every month after that, on the same day of the month.
• For example, if you signed up on August 3rd, your free credits will refill on September 3rd.
• If you joined on the 29th, 30th, or 31st of every month, your free credits will refill on the 28th of every month.
• Free credits don’t roll over to the next month, so they expire a month after they were granted – but you’ll get 15 new free credits.
so i havent saved anything i've made but i've closed the tab that has all my history creations and now i cant seem to get back to them without typing something in the prompt
nevermind i found the button
Can you do both shifts and transparency rectangles at the same time?
In case anyone wants to download all their saved images, you can go to the My Collection page, scroll all the way to the end (or the last image you care about) and then use Chrome's 'Save as...' menu item, and tell it you want 'web page, complete'. For me, it downloaded all the images into the directory named the same as the page name (though it didn't pull them from the browser cache, so it took a bit longer than I expected). It names the images numerically, so if you care about the prompts, you need to go through the html file, but it's better than nothing and secure.
well yes
hi there, I have a question, if I use a prompt that has someone else's IP (for example, the teenage mutant ninja turtles) as part of the promptt, but the resulting images do not really meaningfully resemble this IP, is it okay to share it?
I've seen a lot of generated art that riffs on existing IPs, but I don't want to go against the wishes of the team
here's from the new user submission channel today... is this actually okay?
#new-user-creations message
Is Nightmare on Elm Street, Freddy Krueger something DALLE would understand?
probably but I wouldnt bet money on it
Only one way to find out.
whats a good artstyle for something futuristic?
I’m a broken record but I feel like digital art can do that easily
nice
Has anybody tried making prompts using only “up goer five speak”? https://splasho.com/upgoer5/ If anybody ever does try this please ping me
I used the word "war" in reference to making a war-mech and received a warning for such, is there a more detailed reference to what is not allowed? I really don't want to get banned, I use Codex almost every day. 😦
@elder delta
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
API Terms & Policies: https://openai.com/api/policies/terms/
I've read that, I meant maybe terms that where banned
There is no word list. You just need to avoid the restricted subjects as listed in the content policy.
Is there a way to change the text to a saved image without doing a whole regen? Sometimes I want them to read as a cleaner caption not the long prompt
Should I flag this? IT seems a bit inappropriate to return when I asked for "superdad"
Lol
Like I won't lie I laughed my jaw clean off for a hot minute but it's a bit of a spicy take to be returning "DEAD" when people ask for "dad" content
It just amused me how badly DALL-E faux-pas that one
Could you point me to the FAQ or Tips on getting text to work correctly on v1. I haven't been able to find it
I'm trying to get a MVP trophy but keep getting random stuff
DE2 is not good at text. Requires a different type of AI. It may render a few very common short words, but that's about it.
oh ok. thank you
im still super thwacked out on this but maybe im unthorough: how can i like, get variations of a drawing i made in different art styles?
DE2 is not designed to do style transfers of that type.
aw i swear i saw somehting like that, maybe it was midjourney or something
btw tysm for being such a big help to the community 🐊 👍
BTW, I can confirm your report that the DE2 Dictionary does not seem to be loading images. I have reported it to the operator of that service.
thx gigachad
Dall E doesn't allow editing realistic faces. Is there a way to edit my photo only by getting consent from open AI? I really want to make my photo look cool.
hey i have a few questions. 1 if purchased credits and generated images, are they considered mine for use like on tshirt? does it have to have a water mark on it? Also is there a way to make sure its known that my prompt was a misunderstanding and/or have the prompt reconsidered as an ok prompt to ask to be generated? for example, one of my favorite animes is death note but when i ask for characters from it it says that it may violate the policy. Also is there away to see the time steps of the image being created? if not can this be an added feature? I have spent almost if not $100 in credits in less than 3 weeks so its safe to say im a big supporter. im hoping to get more user options. also is there a way we can help contribute to the training data or join the development team? this is my passion.
Why is «adhd» blacklisted? 🙄
Death is against TOS, try using the name of his other prominent work Hikaru no Go
Yes, you can use commercially, without watermark. 🙂
They have a policy against "public and personal health". I'm guessing that might be the reason.
Does Dalle do text weights like Midjourney does?
No, but you seem to be able to add weight to a concept by repeating it
Heres a reference talking about this https://twitter.com/phillip_isola/status/1532189616106881027?t=PQvlsIt8rd4GC_xq4NVQww&s=19
Language-conditional models can act a bit like decision transformers, in that you can prompt them with a desired level of "reward".
E.g., want prettier #dalle creations? "Just ask" by adding "[very]^n beautiful":
n=0: "A beautiful painting of a mountain next to a waterfall."
1266
227
Hello I asked dall-e for "a medieval village during the plague where they burn corpses" and I got a violation warning, but I'm not sure why. Does it fall under "Shocking" in terms of use?
burning corpses isn't exactly "G-Rated."
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
API Terms & Policies: https://openai.com/api/policies/terms/
so dall-e is strictly for G-rated content?
Thank you for trying our generative AI tools!
In your usage, you must adhere to our Content Policy:
Do not attempt to create, upload, or share images that are not G-rated or that could cause harm.
My rule of thumb is usually Pixar.
are realistic people blacklisted or what?
if i write a realistic portrait it just gives me art
No! You can create and share realistic faces, as long as they're not of public figures; you cannot upload and edit realistic faces, to avoid deepfakes and bullying.
ah okay so the problem is with my prompt
Oh. That's just because "realistic" is never used to tag actual photos 😅 try: "Nikkor 35mm portrait"
thank you for the explanation
a interesting
what if you want a realistic representation of somebody that doesn't exist with Nikkor 35mm?
like a roman emperor or something like that
do i have to watermark my images to use it as a pfp?
try 🙂
informally, no. but the terms still suggest you need to disclose the role of AI in producing the art
ok thanks
I want to make a portrait of myself, but when I upload the image of myself it's violated, is there anyway?
We can't suggest ways to circumvent the policy on uploading realistic faces, which exists for a reason. Regardless of whether in this case it's your face, sadly.
Can you ask Dall-e to create a 16:9 picture?
No, but you can use some simple compositing techniques to make one using two credits
Making variations uses credits right?
Can you generate variations of an image with a custom prompt alongside it?
unfortunately not at this point. It may be possible with future updates.
How many times are you allowed to get a warning? It has only happened once but the wording is a bit uneasy that it will automatically close your account
hey how can i set aspect ratio?
just wondering, is there something specific that i gotta do in order to use pictures created from DALL•E as my profile picture? or am i allowed to just save it and use it, no other steps required?
Q: I just got a content warning! What does that mean? Will I get banned?
• Please don't worry if your prompt clearly doesn't violate the content policy, but has been flagged by our system!
• We chose to err towards our classifiers/block list being aggressive to start, but have been constantly iterating on them as we learn more. Please share prompts you don't think should've been banned in suggestions so that we can improve.
• While we're not sharing the exact thresholds, a few warnings here and there won't lead to an account suspension.
The images will always be 1024x1024 px by default.
You can 'combine' or upscale them though!
Have a look into #inpaint-outpaint & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale
by @wheat zodiac

Use them however you want, as long as you stay true to the policies! :)
You can simply head over to #dall-e-bot and type !watermark (with your image attached)
and it'll give you a profile picture ready to use!
does editing cost anything in terms of credits?
Yeah it does, it's always 1 credit
Also note that you will get only 3 images instead of the original 4.
thanks 🫀
The smile smiles 😉
Indeed indeed 😌
thanks, just wanted to make sure
Did any of you have success creating a spritesheet of some sort? I have yet to create a spritesheet that can actually be used to create animation in a 2d context
i kept on describing the ai to make the person walk on top of that mirrorball but every single output were of the person walking towards it and not on it, is there a reason for that?
I'm getting a violation on this prompt: "a woman with white hair wearing a black sweater, a character portrait by Riusuke Fukahori, flickr, neo-dada, ilya kuvshinov, anime aesthetic, futuristic, 35mm f1/8 photograph", what could be the trigger?
i spent 5 credits describing the same thing in different forms of english, as i thought it didn't get me as im not an english speaker but everytime the result came out like this
It's not a language problem - it's a problem with the model. It doesn't have good data to work with for humans walking on spheres, so it resists it. Maybe inpainting is the easiest solution.
Riusuke Fukahori is getting flagged, have you tried the japanese name? :)
you mean the kanji?
It works, but it's still strange why the artist is flagged...are they controversial or political?
I'm glad that it works!
The whole content policy from Dall•E is known to be somtimes a bit too strict and sometimes too little.
It's hard to get it right though so I get it. For now just use the Kanji characters in case a
Japanese artist get flagged, I belive it's the same for Chinese. 
ah i see, thanks for the info!
Hi all, newbie question regarding the mobile site: I generated a gothic chess set, tapped the download button on each one, but it doesn't seem to download them, and it doesn't seem to save it to my collection.
How do I save generations to my collection? Thanks.
(I will try the desktop website when I'm home later today)
Simply press the "Save" button
hi
hello guys, newbie here. Is anything associated with a Military violates the content policy?
Violence is against the content policy. So most military related things would probably be in that category.
Thanks for the clarification. I'll try to avoid it.
Hej everyone, is there an option for me to create an image with a certain aspect ratio? 🙂
I’ve seen very very very cool-looking actually do something before
Not without cropping it, sorry
Hiya, yeah there is!
The images will always be 1024x1024 px by default,
you can 'combine' or upscale them though!
Have a look into #inpaint-outpaint & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale
by @wheat zodiac

Thank you both for the answers, i will look into the docs ! The thing is, combining or cropping images is not a single interpretation from the AI and then some of the "meaning" (whatever that is) gets lost. If for example i tell the AI hey, fill me a 24:65 AR canvas the results will be different than combining or cropping from 1:1 images. Dont know if that is a limitation because the training was done on 1:1 images but that would enhance creativity to higher level imho.
hello there
I was messing around with a picture of mine, it's myself sleeping in the grass
so DALLE beautifully made the place like i wanted it too be, and with regenerations of certain areas again and again i had such a beautiful ALBUM COVER
i relly want to use this image to illustrate my next album
so it would mean maybe have it on stickers to sell, things like this
how okay is it ?
(OpenAI will of course be mentionned as being part of the creation of the image
)
Just noticed this channel, so I may as well ask one last time to see if anyone has a more exact answer
Anyone know around when the Dall E 2 dataset cuts off?
@faint mantle @short salmon well thanks a lot 🙂
does dall-e knows VHS style ?
hey i cant pay credit with my debit card bec my bank need 3D secure merchant , any other ways?
probably
has not been dicussed heavily in the server tho, try a prompt, if it doesnt work post in #❌┃prompt-help
thanks a lot ❤️
Revolut has worked for some - or prepaid credit cards :)
what is revolut?
An online Bank
ok thanks fam
Hi! I've invited some members of my team and one of them can access Dall-E, but two of them can only access the API Text Section and Playground/Examples area – not the Dall E image generation – anyone know why?
Hiya! Try accesing it through https://labs.openai.com directly, and if that doesn't work give another browser a try :)
Thank you! Will try and report back!
Hi, if I’m using a image made in dall e in a render to make a sky texture must I make the water mark in the bottom left fully visible or put credit to open ai?
Hi, I've seen in the website that there are plans I can buy for more credits (1k credits per 0.03$ or so...).
I lost the website, does someone know how I can get to it?
Else then that. I want to use dell e inside a discord bot as for this:
!generate <text prompt>
and make the bot generate the image and send it to the discord channel, is there any known library that does that yet?
how do I see my oldest generations? chronology on the right stops at some point
and I can't find a button to go further in the past
after some generations, you can only go to saves generation.... if you didn't save, they are marked as temp files... and will delete after a period of time
Stay true to the policies at all times
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Removing the watermark is still a grey area in the ToS, so be mindful of that and decide for yourself.
For the time being, you should be fine as long as you clearly state that the image has been AI generated
(for example, in a description) and give clear credit to OAI/DALL•E 2.
I'm saying should because, as I already mentioned, it's still a grey area legal-wise.
Please also note that the ToS are currently getting updated AFAIK...therefore changes might happen
sometime in the future, so come back every now and then and double check 
(I'm in no way associated with OpenAI. Please make your own decision after reading through the policies)
Ah alrighty Tysm!
There's currently only one way to purchase credits, and that is through the offical webite.
115 credits for $15. https://help.openai.com/en/articles/6399305-how-dall-e-credits-work
Did you perhaps confuse the tokens from the API with the credits from Dall•E?
-> https://openai.com/api/pricing/
yes!
thats what I got confused with
what are tokens used for?
what are the difference between this and credits
Not yet, OpenAI however plans on releasing an API to
the public sometime in the future.
The only way you can generate your prompts right now is on the offical website.
https://labs.openai.com/ :)
I saw a discord bot that is generating images as I explained, I wish I knew how to do that \:
btw, is anywhere in the ToS it says if scraping from the website is allowed or not?
Credits are for Dall•E while the Tokens are being used for the API/GPT-3.
(also made by OpenAI, not related to Dall•E though)
yea, so for the bot, I'll need tokens. currect?
so there are tokens already but not yet an API?
Take a look at this, this is what I'm referring to https://beta.openai.com/playground
An API for accessing new AI models developed by OpenAI
Hello, How can I download ALL images that I generated using dalle, including images that re not saved in a collection?
It's a GPT-3, not related to Dall•E by any means.
okay
How to i download images I've developed using DALL•E 2 but i didnt save?
As of now, there's no way to do so. (correct me if I'm wrong)
You can check your browser history for urls with labs.openai.com/e/ or labs.openai.com/
though and find old images that you haven't saved to your collection via the links 
You can check your browser history for urls with labs.openai.com/e/ or labs.openai.com/
and find old images that you haven't saved to your collection via the links! :)
Alright thanks! :)
how can i generate a photo with mario without dalle thinking im referring to a random dude named mario
Is there a good way to get Dall-e to reliably generate text, often it seems to get the idea but the letters are scrambled or chaotic
Repeat what you want to have very often. Generally DALL-E is limited here though.
Dall·E struggles with the latin alphabet quite a bit.
Words with 2-4 letters sometimes work, most of the time though
it will give you gibberish (or Dallish as some call it) 
Keep on trying! Mention the company, some games he was represented in even.
The best help you'll receive though wil probably be at #❌┃prompt-help
I got a warning for not following the content policy for this prompt, but I am not sure which part of the prompt is restricted content?
"a cute photo of death incarnate hugging a stuffed bear. pastel, digital art."
"death incarnate" is the issue
Hi, can someone pls point me in the right direction - can I train Dall-e with any new data sets? ( I want to create vector branding ) specific styles. Any help towards a right documentation or paper would be highly appreciated.
ah that makes sense. is there some other wording that is allowed e.g. grim reaper etc?
Dall•E doesn't like the word death, try grim reaper and see if that works :)
ok, thanks!
Yeah, give it a try!
yep, it worked, thanks!
You can give Dall•E a "blob" of an image and make it mimic the style! :)
Take a look at this #top-tips-and-tricks message
Interesting. thanks
The best way to do this would be by reaching out to the support directly 
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com
thank you!
Do edits take up slots?
I've noticed that when I generate variations on a somewhat realistic face generated by DALLE, the variations tend to be less realistic than the original. Selecting a variation and generating variations on that tends to lead me away from photo realism and toward faces that look more like Pixar characters.
-
Is there any way to guide or fine tune how the variations are generated (e.g. how far should they deviate from the original image, and in what dimensions)?
-
Is it possible to add "negations" to a prompt, so I when generating content I could say "-Pixar" or something like that?
No to both things.
is there an application built by Open AI, which would allow me to upload an image generated by DALLE•2 (in my case, character generation https://www.twitter.com/bfg_nfts) and output a written poem, story, description, or otherwise? Alternatively, an application that let's me input a prompt for an ouput of a poem, story, description?
Good day. I just went through the onboarding process, but that initial sample image generation with special code (I assume to bind Discord to my DallE account) just cost me 2 credits?
Edits cost 1 credit, in return you will get 3 instead of the usual 4 generations. 
cool thx!
Hm that's weird.. technically it should be "free" you only need to generate
one prompt with the code, that prompt can contain whatever you want it to contain though.
Have a look at this for more info
#openai-chatter message
oh well, easy come easy go. thanks @pine void
Do the saved images in "My collections" delete on their own? Or can I feel safe that they're all there
If you save them to your collection they'll stay there :)
You can also check your browser history for urls with labs.openai.com/e/ and/or labs.openai.com/
and find old images that you haven't saved to your collection via the links!
how do I see how many credits i have left
and does one prompt take up 1 or 4 credits
By clicking your profile in the top-right corner :)
Each prompt takes up one(1) credit.
One prompt = 1 credit
This Figure!
Yuss
How's my straight arrow 😏
is a there a way to upscale images beyond 1024x1024 or change aspect ratio?
There is! :)
Take a look at this -> https://bit.ly/dd-upscale (by. @wheat zodiac)
thanks!
You can also 'combine' them!
Have a look into #inpaint-outpaint & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y (by @spring wyvern)
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
OH BETT!!! Thank you for that
@pine void anyway to download ur collection? all at once
There is, have a look at this :)
#starred-ideas message
you’re a life saver. i’m gonna attempt it rn
It's not an official feature (yet?) so the current way is a bit long
it’s better than saving them 1 by 1 that’s for sure
yupp
Neat!
Being this a beta version still.
50 credits is very low.
And paying 15$ for 112 credits is a little too much. Should be more for this price.
115* credits
and also #suggestions for this :)
is possible to add height and width in the prompt?
Are you referring to the height and width of an item or shot within your desired result or the dimensions of the output file itself?
is there any way that i can find subscriptions, how much prompts i still have and such
never mind, found it 🙂
The images will always be 1024x1024 px by default.
You can 'combine' or upscale them though!
Have a look into #inpaint-outpaint & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale
by @wheat zodiac

Did you ask a question? I must've missed it
If you do not have an answer. There is no need to reply
What's your question, I'm confused?
Why the price. If you did not understood.
We are beta testers.
Asking for this amount seems a little bit against the idea of evolving the tool
Running the graphics cards is expensive
You need data to make it better. And only by using you can improve.
I mean, you literally just stated the cost of Dall•E in your original question.
Since it sounded more like a suggestion I directed you towards the channel #suggestions
I'm not working for, nor am I related to the team of OpenAI by any means. So there you go :)
Running the graphics is not that expensive @short salmon
okay 🙂
And as I said. A so high price for so less credits beats the purpose of being a beta tester.
And that's why I directed to you towards the channel #suggestions :)
I already left it there.
You are getting incredible pieces of art for almost every credit. At around 14 cents a prompt
I want to know... Why the price?
The peaces of art as you call. Are square and 1024x1024
yep
So
We're not part of the Team of OpenAI, nor are the community volunteers. Why/How should we know?
Please dont @ the entirety of the community volunteers
unless its absolutely necessary
You didn't ask in the first place :)
mb, forgot it did that
is it ok to use a dall-e image as an album cover
Yep
Ok. No answer.
Have a good day.
And keep up the good work. 👍
Nope
lol
awesome, i saw that it said not for a commercial use but i likely wont really profit
Have a wonderful day
The price exists partly as a barrier to protect OpenAI's computing resources, and partly to help them fund future research and pay their team. If the product isn't worth the money for you, don't pay it!
As I said you have full commercial rights to your images
Who talked about not paying it @past geyser
Don't be so black and white.
It is just too expensive for so little credits.
Maybe you should think before answer @past geyser
Usualy gets better results 👍
#suggestions is a great place to write up your extended thoughts on pricing.
Please treat others the way you would like to be treated, and assume best intentions. Don’t harass or attack others, and don’t engage in hateful or generally malicious behavior (e.g. sexism, racism, homophobia, etc.) Community volunteers will issue warnings and then mute, kick, or ban you from the server — depending on the severity.
If you read the conversation before
You should see that I already did it.
Who is attacking who?
😂
👍
This is what gpt runs on
If DALL E is comparable, it is very expensive to run and cool that amount of equipment
hey there! quick question. How does one remove the DALLE watermark? I dont really mind it but the website says its for commercial use.... is it just for commercial use with the watermark? TIA
You can remove the watermark with a third part application or by cropping it out
Ah I gotchya, thank you. I can easily remove it, I just wanted to make sure it was alright to do so and if there was a way to download it without one on the website. Thank you!
As long as you dont mislead anyone to believing you created the artwork yourself, you're fine
greetings - its unclear how to move images from my history to my collection - there is no "save" button. Sorry - I'm coming over from MidJourney and the website seems to lack any sort of save function for images.
to be clear - I can download them to my computer
Save button is in the top right as shown in the image
You can also remove it via 'inspect-element' in your browser!
#❌┃tips-n-tricks message (1:20)
take a look at this video :)
You're currently editing the original, that's why the "save" button won't show up.
If you go over to the prompt where that picture has been originally generated you'll be able so save it :)
weird
click an image
You'll be able to save any image but not the very left one
ah @pine void is correct, I didnt see you were doing variations
Do credits accumulate over different months? Or do they reset?
I see - yes I see the save button now. sorry about the noobie questions
#openai-chatter message
Take a look at this :)
Q: How DALL·E Credits Work
• You get 50 free credits your first month.
• 15 free credits will replenish every month after that, on the same day of the month.
• For example, if you signed up on August 3rd, your free credits will refill on September 3rd.
• If you joined on the 29th, 30th, or 31st of every month, your free credits will refill on the 28th of every month.
• Free credits don’t roll over to the next month, so they expire a month after they were granted – but you’ll get 15 new free credits.
No worries, just gotta learn :)
Thank you
Hello I am trying to generate characters of "furry" types (animal human hybrids in the sytle of Carmelita fox in the game
video "sly racoon") and I'm having a hard time getting the generator to understand that I want a human-animal face with hair. In addition, the more I am told that the longer it is, the better it works, but in the end I only have lines that are more and more vague, almost abstract. Help please that made mad i had already spend all my credits... T-T
Have you tried using the word 'anthromorphic ____' to describe the animal?
Also, quick question for you all. How worried do I need to be when I accidentally submit a prompt with a banned word? I feel like it's pretty generous so far, but I can't help but get a bit anxious anytime I see the warning pop up.
err... i feel dumb now i had tryed everything but that...
Q: I just got a content warning! What does that mean? Will I get banned?
• Please don't worry if your prompt clearly doesn't violate the content policy, but has been flagged by our system!
• We chose to err towards our classifiers/block list being aggressive to start, but have been constantly iterating on them as we learn more. Please share prompts you don't think should've been banned in suggestions so that we can improve.
• While we're not sharing the exact thresholds, a few warnings here and there won't lead to an account suspension.
@elfin hill
Oh, shoot. My apologies, I'm pretty sure I had read that already too.
update when?
Are the images i generate free to use for ex. my websites?
From what I understand, you own the commercial rights
Yep. You have full commercial rights. The only restriction is you can’t mislead people into believing you created them yourself.
All right, cool! Thank you for the answers:)
Is it legal to remove the watermark?
Yes, you can remove the watermark. See Joanne's message here #community-help message
Can i use this for a school project?
Hahahahah ok ok👍
it won't let me press generate when i try to outpaint an image
You need to erase the transparent part, like not erase all just drag your finger/mouse at the transparent part
i am, but it's not doing anything
Can i see the image?
it's just my profile picture trying to be outpainted
Idk how to explain it properly but, try drag it only the edge
dragging it along the edge doesn't fix the button graying out
Ohhh, try erase any part of your image not the transparent part
erasing a corner of that doesn't do anything either to the button
I meant like erasing the part of your profile picture
Not all part ofc
generate button still gray no matter how i do it
i'll do that later
@chrome elm You need to write a prompt for it to know how to fill in the empty space. Then it'll generate.
oh, gotcha- i thought it would automatically try and do it based on the source
hi! you need to edit the prompt as well, to complete the image. You might get best results with using a similar prompt to the one you used to generate the first image
thank you
Does anyone know a word that specifically refers to old-timey diving suits like this, that doesn't also refer to modern diving suits? (Not sure if this is a prompt question, the prompt I have is fine I just don't know the name of these things)
you can also use an external editor to combine images, and DALLE will take the clue
" diving dress" might work
gotcha, thank you
I've tried Standard Diving Dress, Heavy Diving Gear, Old-Timey Diving Suit, Diver Dan, and Copperhat suit with no luck 😦
I'll check with the prompt room, may just need some clever prompt engineering
"vintage diving suit" maybe ?
a good trick for that is looking at Google images
search for an image that is close to what you want and do the similar image seach, and will get you some words matching
Vintage Diving Suit got me close. Problem is the rest of the prompt is futuristic so it might be confusing "vintage" with "modern" haha.
Yep, that's what I did, but every phrase related to those suits I tried got me modern scuba gear
Vintage got me this though so I think I can finagle the rest, thanks 😄
very bioshock!
neato
Hahaha, I GIS'd Vintage Diving Suit and got this and now I am determined to find a way to recreate it with dall-e https://4.bp.blogspot.com/-q2UjHBg50ew/XuuPQIbN0bI/AAAAAAAAYXw/EYCX9mJqW-Y_HMgvQLJbOHDdo8LXaIPzgCLcBGAsYHQ/s1600/diving-suits%2B%25281%2529.jpg
"Standard diving dress, also known as hard-hat or copper hat equipment, or heavy gear, is a type of diving suit "
ive found this illustration ... is there a way to upload this image to use this style to create images of other things?
Just below the generate prompt there is the option to upload a picture to edit
ive seen that ... but it just gave me variations of more basketballs on fire ... trying to create a drawing of Dove in this same style
omg, does the draws are deleted forever?? the first i made arent there anymore!!
From my understanding you have to save them if you want to keep them. After a while they're removed from the server
I have a question
Does anyone know how to do this?
I have some album covers that I want to expand
it uses outpainting and pure editing
Aw damn
What apps?
after effects and ofc dall-e 2
ahem in MacBook?
sorry idk, i didnt and have a macbook
but i think it works on any editing apps, its just you need to know how
ayo i can make you one just send me the image
Fr? Let me send you it
bet alr
DM’s or here?
is there a way to generate writing? like for it to right our exact words?
Style transfer isn't possible at this stage. The model is also not trained well enough to replicate text in anything but the most common visual contexts (e.g. a STOP sign)
Check your browser history.
appreciate it!
Kinda worked!! Thanks 😭
Remember to save the files locally for anything you can't afford to lose.
Yes, thanks!
DE2 is not designed to do style transfers of the type you are requesting.
My browser history is the darkest
Not interested, cheers.
If one were to sell artwork generated by DALL-E 2 - whether 'as is' or edited, modified, or otherwise transformed - how would you reasonably disclose that the image was originally AI generated?
I've reviewed the Terms of Use, Content Policy, and Sharing & Publication Policy and am still not quite sure how to do so in a fair way that also doesn't devalue the work put into prompt crafting, editing, or altering the AI art
While I'm aware that any Joe Schmo could generate some art and try to sell it as their own, that seems disingenuous and if I were to sell the art, I'd like to, y'know, follow the rules.
Is there a way to see “recent history” on mobile?
Not without switching to a desktop view unfortunately.
That's largely why the watermark exists.
I understand that, and I think it makes a lot of sense for images that are solely created by the AI!
But I think it becomes trickier once the image is edited or altered significantly; despite being co-created, I think the watermark would give the impression that the AI generated it in its entirety.
Although, admittedly, I'm hard-pressed to think of a better solution.
I believe there may be a ToS update in the works on this matter. For now, read it and make your decisions based on your best judgement and local legislation.
Is there a way to alter the style of an existing image? I was wondering if Dall-E can be used to make all or a part of a digital painting more photorealistic, but so far editing an image seems to try to keep the original style of the image
Currently there is not, maybe in the future :)
OK! Here's hoping - if it worked, it'd be a huge timesaver 😄
@past geyser Hey, do you think prompts like "Fireman holding a Jaws of Life hydraulic rescue tool, digital art" is potentially risky in terms of account status if there are likely images in the data set that show violent imagery?
Did it get flagged? I wouldn't worry about it but change the subject just in case if it's triggering the filter.
No, but I'm heading into first responder territory, so wanted some insight, I'll just discontinue if I get a flag. For example an image of a tank, is that G rated?
Or report it if it goes through and creates sensitive imagery.
Grey area. Use your best judgement and the policy as a guide.
Okay, I'll use the white background modifier where possible to focus prompt.
On the topic of the ToS, I've been trying to decipher the content policy, sharing and publication policy and terms of use but none of them seem to state who owns copyright, what I can/can't do with the generated images. Unfortunately "generally permissible" is a bit vague for my taste 🙂
Edit: found this page https://labs.openai.com/policies/terms that is isn't actually linked from the menu on the other policy-pages
The ambiguity probably comes from nebulous law, particularly when international contexts are considered. If you need hard answers, you may need to do some research.
I would say the questions are unanswered and in the future they will be fought out in court and politics.
At moment there is no absolute certainty regarding anything about AI.
Was wondering if I could use cigarettes in my creations
it's usually fine, but it feels like the filter has a problem with it if you re-upload the image for edits
ooh ok so I can use it in text prompts right?
and btw can I use people that are no longer alive in prompts?
As long as the person is not a symbol of hate, yes.
Great, thanks
Does anyone know if you can ask Dall e to place exact text on something? If so, how? I can only get it to place an idea of what I wrote, but its usually incoherent
You can, but generally it won't work that well.
ah I see, thank you. I suppose it'll just be trying until I get something good. Also, where did you get that screenshot from? I was looking for something which showed examples with explanations, for that, but couldn't find it.
I was trying to get it to say a long sentence. I'll try playing around with single words
thank you very much, I'll have a good read of that
People who got accepted in july are actually using DallE 2 right?
The invite you received from OpenAI is for DALL·E 2. The first version of DALL·E, version 1, was never released to the public.
Read more here: https://openai.com/dall-e-2/
Great 🙂
hello! joined because someone's been advertising their dall e 2 account for sale in my twitter replies. wasn't sure if that was allowed, and it left a bad taste in my mouth, so i was wondering if i should/could report it?
https://twitter.com/ProSNAP2
this is the account btw, check the replies tab to see all their spam
i just blocked them as a temporary situation but thought i should report it somehow
Cheers, noted!
hey thanks! may i ask what will be done about it, if anything?
There may not be enough identifying information for OpenAI to suspend the account but it's against terms of service so they'll possibly look into it. I'll make sure it's elevated.
thank you! i hope it gets easily resolved
Is it ok if I use lisa frank style prompt concerning copyright?
I saw it in DALLE2 prompt book.
hey guys, how can I use a character from a specific franchise? For example Solid Snake? The AI just shows a real snake
Sometimes you can easily, sometimes you can only with very careful phrasing, other times it's just not in the data and you'll never recreate it.
Yep all good, at least from OpenAI's perspective.
So should I write "Character X from the Y-franchise"?
Hi, in welcome message was written - to connect to the server I need to enter a code after any prompt and it won't "waste my credits" - after I did all of that - I'am here, but in the credits I see 49... I did something wrong?
I'm trying to edit a picture of my gf but it says it violates the terms
To prevent deepfakes, we are currently prohibiting uploads of all realistic faces, even when the face belongs to you or if you have consent.
oh never mind
Thank you
It can help. Try "Solid Snake from Metal Gear Solid" with nothing else and you'll soon see what the model knows.
You don't waste a credit because you can append the code to the next prompt you'd do otherwise.
Hi, I want to do a research project and publish it. DALL-E 2 really got me on text-to-image concept. Is there anyone to help me on this?
🤚 ^^
Hey Jeremy, just a suggestions since this was brought up, would it be possible to highlight the fact that a desired prompt can be appended so that the users are aware beforehand? Thanks!
I'm pretty sure that's clearly articulated in the verification bot already.
oh wait it is!?
Oh yeah it is mentioned in Joanne's message... Could have been added after I joined...
I have a question. I've been looking into using DALL-E 2 to generate textures. I'm not that great of an artist and actually It worked pretty well with some of my test prompts. With some editing I could use them, I'm just wondering about the legality. I'd have to remove the DALL E watermark. I don't wanna use it for profit. Any game or 3D render made/art With the help of DALL-E would be free and credited. I was just thinking about making a quick demo of a game made with AI generated textures. So that's why I'm asking
Where can I read which things I am not allowed to do?
Read up on the content policy and terms
#old-server-rules message
Yeah the system was changed at one point for this reason.
Totally fine. Acknowledge the role of AI in the credits.
ahhh gotcha.
I'll put it as a disclaimer when you launch the game and in files to be safe. Thanks for the answer
Is there a tutorial anywhere on how to put in modifiers in an image I upload? Do I write in a description and then upload? Ex. I want generate a specific cat photo I upload and have it leading a marching band
Hello, I wonder which AI is the most technically demanding (in complexity and dimension) between DALL-e and GPT-3? (And is this an ok question for this channel ?)
Still no luck sadly! Any other advice? One of the people on my team can access Dall•E but the rest can't 😵💫
is DALL-E something I can post a video about on youtube to share some of the designs created or is that not allowed?
I know we can share individual images, but is there a way to share my entire collection? I want to just get a link for my whole collection so my friends can look through them all, instead of my downloading/sharing them 1 at a time
I've just been tossing mine into a dropbox folder
Yeah, that works. But I'd like to share the prompt with the image
Hm... That's really weird. You can give another device a try,
so switching from a PC to mobile and vice versa e.g
For some it also worked turning on a VPN.
Give these a try and come back to me to let me know! 
In case none of those work the best thing to do is to contact the support directly.
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com
There currently is no way to do so, maybe in the future though :)
If you have a suggestion post it in #suggestions so that Dall•E can improve in the future!
I'd say the best place for this would be #ai-discussions...
I like the question though, ask it in there! I'm interested in seeing the convo :)
Hi I'm new to Dall-e, is there a way to generate images other than with 1:1 aspect ratio?
Take a look at this, it's gives you a super introduction to Dall•E :)
and I'm sure that you'll understand the way that it works waaaaay better after!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern (page 55)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8 by @wheat zodiac
Awesome, thank you 🙂
The images will always be 1024x1024 px by default.
You can 'combine' or upscale them though!
Have a look into composites & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale
by @wheat zodiac

got it, thanks!
:-) @fair night in #ai-discussions explained that Dall-E is bigger because it's apparently layers added to GPT-3.
Is there an equivalent OpenAI Discord to this one for GPT-3?
I would be very interested in confronting tests and ideas about GPT-3, and learning more about it. With textual interaction.
@pine void ty for taking your time and helping people out with their questions 🌹
Is anyone else not from the US having trouble purchasing credits?
It doesn’t let me buy them despite me putting in my correct details and everything
Nope. 4 images per token
Okay, thank you
What country are you from if I may ask? :)
If you don't want to answer/keep it privat you can always message the support:
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com
I’m from the UK
Is there any way to change the email address associated with my account?
I work as a researcher, and I am moving to a new institution. I will lose access to my university email address.
Please email support@openai.com for account related concerns.
Thank you
How do i change my password?
Hello, do I own all the rights for the art? Can I print the art on canvas and sell it if it doesn't infringe other copyrights or are there special commercial terms regarding the use of DALL-E 2?
somewhere I saw it mentioned where a dalle account can be used with a team.... how do we do this? I want to let my staff use this to make images for some of our web tools. are accounts sharable or something?
how do I do this?
Does dalle ignore small typos? For example I misspelled "Peaceful" as "Peacful" earlier, would my results have been different had I spelled it correctly? 🤔
I don't have an actual answer for this, but in my experience it doews seem to ignore/correct typos
so before i did my first prompt, i watched a video of someone doing it just to see how it worked. they got 10 results, and I only got 4. did something change with dalle? the video was posted at the end of april this year
do all photos download as 1024x1024 px with 72 res or can that be changed? I want to print some of these at large scale and high quality for my office
Try to not input any spaces whatsoever in the card prompt's postcode entry.
A few people from the UK had a similar problem, so let me know if that does it! :)
The images will always be 1024x1024 px by default.
You can 'combine' or upscale them though!
Have a look into composites & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale <- For upscaling
by @oko

can I change my display name or go anonymous on the dall-e 2 website?
Back then it was indeed 10, this number has now changed to 4 though :)
If you have signed in with your Google account you might be able to do it through that.
Otherwise, the best way would be by contacting the support directly :)
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com
I signed up with google, I dont know why it took my real name and not my nickname or displayname
that makes no sense
maybe when it asked for my real name
I was dumb enough to write it in and not fake something
Oh yeah, that's most likely it. The best way to change it would be by simply reaching out to the support
ugh
Have a read through the policies :)
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Ok
they say I have to give my real given name?
also "
Our services aren't available right now
We're working to restore all services as soon as possible. Please check back soon."
oh nvm
missed the part where you replied to him
okay reached out on chat
6 hours ~~ reply time
Is there any way to see images created that are no longer in "recents". I JUST realized you need to Save them for them to show in your gallery 😦
will there be a point in the future where you have more free credits with dalle? and what's the reason for its decrease from a few months ago?
wait so is this dall-e 2 or just dall-e?
Why is it that every prompt i do with kanye west in it i violate the rules?
Still nothing sadly, keeps either declining or saying something went wrong
You can check your browser history for urls with labs.openai.com/e/ or labs.openai.com/
and find old images that you haven't saved to your collection via the links! :)
This is Dall•E 2 
I have used a prompt to get verified on Discord, using my personal code but I still had that credit being used, rather than not being used as it said when I was trying to verify my account.
Is there any way I could get that credit back?
I can't give you an answer to either question, I don't believe any of the offical volunteers can either(?)
It's up to OpenAI, the best thing to do is simply to wait and observe :)
For now we're recieving multiple amazing pieces of art with almost every prompt at a cost of 14 cents, that's pretty fair imo.
Have a read through the policies
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Current public figures are not allowed but historical figures are typically acceptable
That's indeed weird... I'd say your best bet would then be to message
the support, if everything else has been submitted correctly that is.
It's "free" you only need to generate one prompt with the code.
That prompt can contain whatever you want it to contain though
and the code doesn't really mess with the prompt itself. 
It wasn't though, I had 34 credits left, I used the prompt to get verified and now I have 33 left.
As I said, it's "free". The credit will still be used, you can however write anything for that prompt.
Here an example "a low poly render by Scott Listfield, rendered in maya,
the code"
The prompt would still give you the wanted result as the code itself should't really mess with it.
@summer siren
I am so confused, it is what it is, is just one credit but I thought it won't cost a credit, as it said when I started to verify my account on discord (The bot will ask you to incorporate a numeric code into your next DALL·E prompt, using a DALL·E credit -- and the code won't affect the results (so you wouldn't be "wasting" a credit). Thank you anyway!
For sure! :)
If you want to you can name me a prompt and I'll "give" you that credit back by generating it for you.
It's fine, thank you.
Hello OpenAI team, is there any roadmap for when an API will be launched for developers to use DALL•E?
Just to let you know, as long as you don't see a yellow name you're not talking to an offical team member of OpenAI. :)
To answer your question; There currently is no roadmap, the most we know about a public API is this:
#suggestions message
Q: Is there a publicly accessible API for DALL·E?
OpenAI: We do not at this time, but do have a longer-term plan to offer this. Please also do not attempt programmatic access of DALL·E, which is against our Terms of Service.
I can't reset my password. The email just doesn't arrive. I want to detach my account from google auth. Is this possible?
Did you sign in with Google?
yeah originally
If so the best thing would be to contact support via, support@openai.com.
Ah thanks, will do
Question: Can I invite my artist friends to try Dall.E if I have been already invited on the platform?
Negative, DE2 does not support user based invites. However I suggest you get them to apply for the waitlist as the number of admissions are really high at the moment.
Yes, the wait time on the waitlist is long. That’s why me and my collector were planning for give an invite to an artist in our next twitter space.
Is there a way we can this?
No. Currently the only way to get an invite to the beta is to join the waitlist
Understood, thanks for your prompt response team 🙂 🙏🏼😁
Will Open AI accept payment for credits via paypal eventually?
I just got an email saying I now have access to Dall E, but I can't seem to find out where to use Dall E 2. I only seem to have access to Dall E. Does anyone know why this is?
This is Dall•E 2, there is no Dall•E 1 :)
Maybe, for now it's credit card only.
Revolut has worked for some too - or prepaid credit cards 
Dear OpenAI Team;
Unfortunately I’m currently living in Turkey and I need to buy credits in Dalle but I can’t pay from Turkey but select as billing address from Turkey, the system rejects. I really need to buy credits but I’m in a really sole remedy right now. I don’t know if it happened because of taxes, political sanctions or any other reasons but I DO REALLY NEED URGENT SUPPORT TO PURCHASE CREDITS; I really love and appreciate your impressing and amazing content and services as a usual developer and content creator… Could you please tell me if there is any way to purchase credits from Turkey🙏……Excuse me for disturbing and thanks as a fan of your stuff and a tech enthusiastic:)
Do you know where is the exemple gallery. Can't find it
Anyone tried the arc face restoration tool recently? Not working for me, uploading an image results in a error. Are there any other tools like this?
How many images are in DALL-E 2's dataset? If that's the proper way to phrase the question
Try it some days ago and have no problem, result is pretty impressive.
May be the frame ?
Don't know about the DALL-E dataset, but to get an idea LION-5B who is used by another text image generator is 5,85 billion "Clip filtered images pair"
I thought rather too close or too far
Is it possible for someone to view my DALL-E saved images or do I need to post them (on the discord or elsewhere?)
Why can’t I buy credits with PayPal?
Images not shared via a link or file are only visible to yourself (though OpenAI can also access them for content policy reasons)
Can't confirm or deny anything, but my account has gotten a lot of warnings since I rely on image prompts (Once put a picture of a dog and it got flagged) but I think since I have a large volume of inputs and pay for a lot of credits, it seems to be okay.
I think the warnings are more about the proportions (like 10 okay prompts per 1 warning prompt) than total amount (i.e. if 10 warnings got your account frozen).
No exact numbers, and I have no confirmation from Open AI, but if you are a large-volume user like me this might comforting to hear
I mean I think I read that in the server. "A warning here and there won't get you banned" could be right, but they prob won't confirm
Guys would psychedelic prompt be an issue?
If it's simply about the word 'psychedelic', no, you're all good the content filter doesn't block that :)
Hi all, I am new here. I had a question about my history of generations. How do I see all of them? I can only see "recent" generations. Thank you
Apologies if this is already explained in an FAQ
I have the same issue. I didn't realise, that new generations override the older ones in "recent". Is there a way to get back to them?
I really hope so! I will be certain to save my favorites to my collection right away now
I've found this @stone cove
Ah great find! This will do for now I think
Yeah, there currently is no other way sadly. Best thing for now, save them individually 
I was able to sort my history by location and shift+click to select them all and backup those old Dall E links from my previous generations. I will go through them one by one eventually haha
Thanks!
!pad50
How do you download generated images on Android?
There should be a download button when you click on an image you've generated
I have tried clicking it however after it loads no image is downloaded
Its hard to help without seeing whats happening
Is it maybe in the photos area of your phone?
Why is my goose still one-headed? Am I using the photo editor wrong or is DALL-E's fill feature not meant for major additions like this?
The little original photo next to the prompt shows what I erased from the original. There are some slight variations in the tail feathers, so DALL-E is registering the erased area.
DALL E struggles with things like numbers of heads
Numbers of anything really
How much energy does one generation use?
Hm. Not sure actually
DALL-E won't lengthen my geese either! What's the point of this technological marvel if I can't create eldritch goose abominations? 😤
I've seen people do some pretty impressive things with inpainting and uncropping, but it mostly seems to ignore my inputs. Seems like its dedication to realism is stronger than to my new input.
I've noticed that recently
Currently there's an official "Emoji Suggestions" thread, and DALL-E-generated emoji that are good enough might get added as server emoji. But... wouldn't this contravene OpenAI policy on clearly marking all outputs as AI-generated?
I mean, OpenAI is always free to disregard its own user policies, but I wanted to check on this because there's some dissonance there and I might be misunderstanding something.
(unless there's some plan to somehow mark them?)
And to be clear, it seems like a harmless and fun thing! But so are a lot of the things people might do with unmarked images. 🤷
hey strange question maybe but where does dall-e get its dataset? :0
i've noticed it can do pretty good with some fictional animated characters but with others it totally flops (even from the same series, i actually generated a picture of a side character which came out better than the main character)
also, are there any plans to expand the dataset? or is it gonna be kept as-is?
I had some lucky with it, just dont tell it that its a goose :)
Does DE2 actually read image text when generating variations? It seems to have been successfully bamboozled in that first set. 🤔
Third set with a closer original image
Hello?
Hiya, what's your question? :)
Just saying hello
Greetings everyone :) I go by Suki, and I'm from America
I enjoy talking with people and exchanging ideas ans whatnot
I'm 32 and in my free time I enjoy gardening, playing video games, volunteering, traveling, watching silly things on YouTube!, and seeing where the wind takes me.
I also have insomnia ans have trouble sleeping, so I'll probably talk at odd ours where other Americans aren't awake.
That's all I've got for now. Thanks 😊
is this still used?
Hey, I had no idea that dalle2 erases your history after a certain number of creations. Is there anyway I can get those back? I didn't save them.
Ok so I have got this reference image. I want a Chinese dragon in a realistic style that has the same color scheme but dalle always fucks it up and I am always out of credits
I am only looking for the colors aka a "Black chinese dragon with dark purple hair and a grey belly in a beautiful Forest"
How do I fix this
Hey,
I`m new here and i've got a question... I tried to turn a picture from spring to autumn with this prompt "Landscape in autumn, 9mm tilt-shift photography."
but got this alert ⤴ What am I doing wrong?
Edit: It works only with this prompt "Landscape in autumn" - But why?
You can check your browser history for urls with labs.openai.com/e/ or labs.openai.com/
and find old images that you haven't saved to your collection via the links! :)
Hiya 👋
Dall•E associates '9mm' with weapons/guns...
Instead of '9mm' go for '10mm' or '8mm' 
Yes it is! Feel free to submit :)
thank you smile!
Bruh... didn't expected that xD
Thank youu
You're welcome, happy to help! :)
I used the online chat thing in the openai site for a question, usual reply time is "under 6 hours" it's been more than a week
thank you, thaat worked out!
Is there a platform to look at openaidalle pictures other than say instagram, that is more accessible to the elderly?
There is no publically accessible gallery, if that's what you mean
Good question... I'm not sure to be honest.
If you simply want some example pictures you can always take a look at this:
https://dalledictionary.notion.site/dalledictionary/DALL-E-DICTIONARY-2ec3dd6dc2184a45a9b6b4b8a021f0a6
&
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y
While these are not a publically accessible gallery they still give you a lot of example pictures and lots of different styles.
The channels on this discord server such as
#photography #art-and-illustrations #daily-theme #inpaint-outpaint and many more are a nice addition too.
I hope this helps you at least somewhat 
as i have already faced some results that didn't match the prompt at all, wouldn't it be worth trying my prompts on dall-e mini before, to see if the ai understands what i meant?
my grandad just enjoyed some of the pictures uploaded by some of the users created by dall-e 2, and I was wondering if there was any easy way for him to look at a couple of pictures daily. I might just email it to him instead
Yeah, I think making a little collection yourself and sending them to him is going to be the best way :)
I appreciate the links:)
Fs! If you need help with anything else just shout, there are a lot of helpful people on here! 
Are there any plans for DALL-E to replenish Credits over time for free users?
For example, 10 free credits per month or anything similar?
Thanks 💜
Q: How DALL·E Credits Work
• You get 50 free credits your first month.
• 15 free credits will replenish every month after that, on the same day of the month.
• For example, if you signed up on August 3rd, your free credits will refill on September 3rd.
• If you joined on the 29th, 30th, or 31st of every month, your free credits will refill on the 28th of every month.
• Free credits don’t roll over to the next month, so they expire a month after they were granted – but you’ll get 15 new free credits.
Thank you so much ! 💜
"realistic photo of elon musk standing on mars while wearing a sombrero"
does this violate the terms in any way? I have tried a variation of this prompt twice now with the same message saying my account could be automatically suspended because it thinks this violates the terms of use some how. I feel like this is a genuine error.
Yep. You cannot generate images of real people.
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
API Terms & Policies: https://openai.com/api/policies/terms/
oh I missed that near the bottom lol, thank you
so I got some pretty good results from Dall E and I would like to get the bikes into frame but I don´t know the promt https://labs.openai.com/s/icE63D6UPGGdByRVZC9LTLx5
@silk echo if you like this one your best bet without wasting credits is to use outpainting to get more of the image
alright but what do i do if i want to create a new image?
That’s outpainting! Check out #inpaint-outpaint and #dall-e-bot
ah thank you
How to get the dalle dictionary in my own notion app (on iPhone)
To copy it in the app
Where can I get copyright for an image I generate?
Need it for my company
Or, where can I contact an openai employee to ask this
I don't see any contact info in the DALLE app
I am currently working on a AI generated song, can I use Dalle art work I created so I have 💯 AI record and cover? ( all credit for artwork would be given to Dall-e )
Yep
You have all commercial rights
@short salmon that's great thank you 👍
~ dall-e just down for me?
getting "something went wrong" on all prompts ~
never mind 😉
Hi, was wondering if there’s a mod who could help with a question. My friend and I signed up at the same time, got our invites in a similar time frame. My friends invite was sent to her spam. When she found it, the invite had expired. Is there a mod who could help with this, advise?? Thanks:)
The best thing to do for your friend would be to message the support.
There's no guarantee that it'll work, it's worth a try though! :)
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com
Okay, thank you very much, appreciated:)
is any horror allowed or no
Yes horror and dark things are allowed, but be aware there are many banned keywords relating to gore, so try not to overdo it :)
bump
Hmm this may be on a future roadmap, I have not seen this question before, definitely a good one for support@openai.com
Are there any plans to loosen the content restrictions? I keep getting "warnings" for things like Francisco Goya's the black paintings or "The Coming Storm," the famous painting of John Brown in the lead up to the American Civil War. Also, is my account really in danger of being banned for this (classic art references or remixes) or for other PG-rated image requests? I get the need to be very careful with the technology, but in the long run this seems like it will prevent a lot of interesting art that falls into the category of acceptable for public consumption.
Can anyone tell me why this was violating? I am guessing Dr, but want to get clarity.
it may have used words that other people used to request pornographic images. that's why "fursona" is banned, even in PG requests
Real living figures may also cause violations
interesting, hulk hogan does not
The parameters are very narrow and fussy for what's acceptable. The A.I. is powerful and fun to play with, but beyond creating anachronistic images or weird juxtapositions (Dinosaurs driving spaceships! Dogs doing brain surgery!) it can't really illustrate anything interesting. You can't even have a detective eating a sandwich at a crime scene or something if you're trying to illustrate a story.
You can have Hulk Hogan dressed as a cow eating spaghetti, it was the first prompt I ever wrote LOL
LOL well that's something. I mean, it's an amazing tool, I'm going to keep using it and I love it, I'm just surprised at how narrow the scope of what you can make is.
Just depends on the story you want to illustrate :)
wait is there a dalle app?
Nope! just the website
Hi, I generate an image with Dall E but i dont know where can I see it.
Hi is there a way to remove the rainbow icon on the button right?
Right below your prompt :)
I cant, google closed automatically after generate it
So, I dont know what can I do
There is, however please note that this is the watermark of Dall•E so have a read through the policies:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
How to remove it:
#❌┃tips-n-tricks message (1:20)
Or with an image editing software like Photoshop, Gimp, etc.

Have you tried switching your browser yet? :)
Well on phone you can add it to your Home Screen on when you visit the website
Yes, and Im right here now:
And the moment you type your prompt in it just closes itself?
Hulk Hogan is a stage name for Terry Gene Bollea
Yes, but i saw a image that I loved. So Im trying to recuperate It
I just Wikipedia-ed it.
And you tried multiple different browsers such as Chrome, Opera etc. as well yeah?
Chrome and Safari (IOS) only
does anyone know if there are other ways of paying for credits? i dont have a credit card sadly
Give Opera a try and if that still doesn't work I would try it on another device such as a laptop :)
or another phone
Revolut has worked for some too - or prepaid credit cards 
Ok, but can I retrieve the image I saw before?
ty, ill try that
Sorry to interrupt but does anyone know why my dalle wont let me upload a picture? I’m able to select the picture and all that. And then I get to the page where it tells me to crop the photo to a square, and it won’t let me move on from there
Never mind sorry
Is there any way to restyle an image but keep the general shape? I've got a map of my town and I'd like to make it look like a medieval fantasy map of my town.
Hi, is there a limit to the "recent" section? I made a lot of images previously that I would like to download but I can't see them anymore 🤔
The sidebar keeps the last 50 prompts. Beyond that you'll need to have saved the /e/ address or have them in your browser history. OAI may be looking to expand the history features but we don't know much for sure yet.
thanks
Understood, thank you! I will look into my history then
(web browser history)
Thank u so much!!!
How do I know if DALL-E knows an artist?
Ok, where i have to go to see the recent section?
It's this icon
Im in iPhone and I dont have these button
Flip the screen to horizontal view
Or enable the desktop site setting
Ok, thanks
Anyone know where to find a catalog of styles and specific details to try as prompts?
I feel like I’ve just been stuck saying “digital art” 😂
Sweet thanks!!
The prompt book mentions 50 daily variations, is this old info or are edits inside Dall-E 2 separate from prompts?
If I remove my face from an image, like totally, is it then okay to use that image as an edit because there is not a face anymore? (or a face being deleted out of an image in general, really)
is there a limit to how many pictures you can save to your collection?
if you buy credits (say you have 115), would you still get free credits the next month (so you'd have 130)?
Correct! But free credits do not stack with credits from the previous month. So you cant just leave your account and accumulate them over time :)
The current limit is 10,000
thanks toror!! :D
All it does is prefill the prompt -- with one of the prompts you've already seen as a "tip" on that page. It's not actually very surprising!
I don't think it actually submits the prompt, though.
(I only tried it once I ran out of credits, haha.)
I just see recent tutorial and demo on DALL-E on youtube and I saw that they get 6 images, but I only have 4, is that normal?
The 6 images was from early in the beta, it is 4 now
ok thank
will there ever be different or more kinds of payment methods for credits?
The 6 images was from early in the beta, it is 4 now
Thanks again haha
How do I do outpainting with dall e
Basically, you upload an image with transparent areas -- download the image, shrink it, give it a larger canvas with a wide transparent border, upload that, and prompt it to paint in.
(I think you also have to "erase" part of the transparent area to get the UI to let you proceed.)
Oh, no.
You have to give it a prompt for what you want it to paint in the transparent (erased) areas.
Take a look at this visual guide :)
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y (starting page 55)
My english isn't the best, does this affect my dall-e generations in any way?
how often (if ever) is dalle 2's database updated?
hello is there a way i can upload a base photo for my prompts?
are there ever plans to let people use the Dall-e tech with their own training data? The style translation powers are so cool and I think they could be used to make 3d animation look hand drawn given the right data. Basically will we ever get access to more than just the prompt feature?
There is an upload button - if you want it to create its own variations. You can also erase part of it. Check out https://discord.com/channels/974519864045756446/998384442785660958
is there a way to get rid of this
Is this something that would be against the content policy to share on social media, or is it too close to drugs?
The prompt was "Uhh, I think that was a macrodose bro" but it doesn't show any actual distinguishable drugs in it. Just bottles and uhhh... clumps of moss flying around?
Should be fine, just don’t use it to promote any drugs or claim that’s what the AI is meant for :)
Where can I read the rules and things that aren’t allowed to generate in AI
There is, however please note that this is the watermark of Dall•E so have a read through the policies first:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
How to remove it:
#❌┃tips-n-tricks message (1:20)
Or with an image editing software like Photoshop, Gimp, etc.

Yeah, to some extent, it will affect your results.
Take a look at these resources; they will help you understand
how prompts work, what's important, and, most importantly, how to achieve the results that you would want :)
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
https://bit.ly/3SHBBi8 by @wheat zodiac

how do i disconnect my credit card with my account?
Hi so I was trying to generate some fun comic book art of "superman converting to Islam" and I got a warning that my prompt violated the content policy? I've been able to generate Islamic art and comic artwork in the past so I'm wondering why the crossover of these two things is throwing up alarms.
Not too sure if that's possible without help from the support,
best thing to do would be messaging them:
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com
ok thank u!!
Im kinda new here but how do you use your padded image from #dall-e-bot ? I want it to outpaint the blank part of the image but for me it only makes the blank part black and the middle different
Oh i got it nevermind
do i have to save images as i generate them? i've just blown through a load of credits and seems i can only see the last 50 or so?
You can always check your browser history for urls with labs.openai.com/e/ or labs.openai.com/
and find old images that you haven't saved to your collection via the links! :)
ooh let me try 🙂
Why is this prompt against the content policy? It doesn't violate any of the points.
And if you want to save them you can simply do that by
adding them to your collection or download them :)
^
Celebrities/Politicians etc. should be avoided from being generated 
Please have a read through the policies:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Does anyone know how to contact DALL-E i've looked on the website unable to find contact form.
E-Mail: support@openai.com
Live Chat (available from Monday - Friday, at least I think so): beta.openai.com

Thank you for that.
Hmm, it sounded like the point on politicians was for living politicians and elections, not historical people
can i outpaint an album cover? or is that against tos
That's mostly the case, however,
hitorical figures such as Stalin e.g are supposed to be avoided :)
How can I use dall-e 2?
Can you be more specific?
You go to labs.openai.com and type in a prompt in the text box
Oh ok! So simple
Guys can I change aspect ratio?
You can’t change the aspect ratio but you can use inpainting to extend it
The images will always be 1024x1024 px by default.
You can 'combine' or upscale them though!
Have a look into composites & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale <- For upscaling
by @wheat zodiac

Can i pay for more Credits with PayPal, Crypto etc.?
At this time, credits can only be purchased with a credit or debit card. We are exploring supporting additional payment methods in the future.
how to see all historical images generated on the web?
If you are talking about the history of generated prompts
you can always check your browser history for urls with labs.openai.com/e/ or labs.openai.com/
You also have a history of up to +50 generations on the right side :)
Very new and learning but wondering if i could get pointed in the right direction. I am looking to redesign an old logo. I have tried "Crisp edges, high detail, professional logo redesign, "Red Beard Collectibles" text, "Est 2011" Text" but am getting nowhere. I just joined today and have about 2 weeks of ai art exp, thank you for any suggestions!
when dall-e becomes free for people? will the credit limit be the same or unlimited?
Is it possible to define resolution/aspect ratio?
The images will always be 1024x1024 px by default.
You can 'combine' or upscale them though!
Have a look into composites & also at these resources!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale <- For upscaling
by @wheat zodiac

Has anyone found a tag for the images to resemble Dungeons and Dragons art?
I have made many images but only some show. Many are missing and are not showing in the sidebar. They also don't appear in my collection. Do they expire in 1 day?
The sidebar keeps the last 50 prompts, beyond that you'll need to have saved the labs.openai.com/e/ addresses (they'll be in your browser history if you have it enabled).
OpenAI may be looking into expanding the history functionality sometime in the future.
To send an image to your collection, click the "save" button.
(Images aren't being deleted from OpenAI's servers at the moment)
Wow that is really surprising. This is a serious issue. This should be stated very clearly.
Since images aren't being hard deleted and are still attached to your account, I (personally) expect "lost" generations to become available to you again in the future, but that depends on when/if OpenAI implements that functionality. I agree that it should be made more clear though, lots of people don't realize at first.
I did a search before I posted and saw a bunch of people are discovering this the hard way. If you are paying for credits to generate. And it takes many attempts, you can move through that recent 50 very quickly. Thank you for the tip on the link history. I am going through yesterdays history and that seems to work. In addition it is time consuming to have to both download and manually save to your collection.
Thanks very much for your quick reply and information as well as your history tip! I really appreciate it!

Is there a section where I can look up the data base of dalle? You know because there are some times that dalle give as a result something too neutral/stock and it is not what I expect.
Hi all - Are we not allowed to reproduce famous works of art and edit them? I got a content warning for this prompt: "reproduction of Edward Hopper's Nighthawks, with Kermit the Frog added"
are you looking for a way to make your prompts more fine tuned to gain your expected output?
If you want something like a public API the answer would be no.
They do plan to add that however in the future :)
If you simply want to fine tune your results etc. take a look at these two links.
They are great for getting started with Dall-E and also learning about prompts and much more! 
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y @spring wyvern
https://bit.ly/dalledictionary by @wheat zodiac
For help with prompts head on over to #❌┃prompt-help
In your prompt the word "reproduction" is the problem, simply get rid of that and it should work :)
Also have a read through the policies:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Does anyone know if we can use DALL-E photos without any copyright issues? I want to use a generated (and slightly edited) photo for a music album cover
Yes! Per the terms and conditions, as long as your generations follow the content policy, then you will retain exclusive rights.
Do note that the terms may change at any time as DALL-E 2 is still in beta.
Have a read through the policies here:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms

Awesome! I appreciate the links, I'll take a look.
Can someone confirm this?
I had read in the FAQ and content policy that it's not allowed, however I've been seeing many articles pop up like this — just wondering!
Feel free to generate images of realistic people that do not exist by using general prompts (eg, a police officer, a teacher).
Generating images of celebrities and public figures is prohibited, unless they are historical figures / no longer alive.

Is DALL-E planned to have variable image sizes? Like, 1:2 or similar?
Thanks for the info!
There is currently no way to generate images in any ratio other than 1:1; however,
you can use compositing to stitch together multiple generations in different ratios.
Take a look at these!
https://pitch.com/v/DALL-E-prompt-book-v1-tmd33y by @spring wyvern
(the section starting page 55 is what I'd recommend you to look at. Page 63 is the topic combining specifically)
This whole notion-site in general is super helpful!
https://bit.ly/3SHBBi8
https://bit.ly/dd-upscale <- For upscaling
by @wheat zodiac

Thanks!
Another thing you can do is generate an image, pick the one you like, open it in image editing software, expand the canvas so it's the size and aspect ratio you want it to be, then expand that so that it fits within a square, with all this new space transparent, save it out as a PNG, and then upload that with the same or a similar prompt to expand it with outpainting, and then crop it back to where you want it
A related question I just came in to ask:
If I upload an image to edit, does it work better if my prompt describes just the missing areas, or is it best for my prompt to describe the entire image?
For instance, let's say I try to generate Conan the Barbarian and Gandalf confronting a huge ancient red dragon, you know, the most D&D of all D&D prompts... but it only gives me one figure, the barbarian warrior for instance, and the dragon is incoherent.
My thought would be that I can first erase an area in close to the shape of a wizard I want in there, and then have that fill in, and then take that result, in turn, and upload edit it to remove the dragon to try and make it more coherent.
In each case, would it be better to keep with my original prompt (or a similar one, perhaps adding detail but removing art style etc because it can get that form the imagery that remains), or would it be better to first describe just hre wizard, and then just the dragon?
Well, if you start with something that looks zoomed in, when you do the outpainting to expand it it won't be anymore
You should describe the entire image, that usually works better :)
I was hoping otherwise, because 400 characters is limiting
it's like it feels like the characters are wasted on describing somehting that's already there anyway
Try both and play round, I myself have gotten the best results that way too though.
You can also try and create different parts of the character first and stich it together afterwards, that requires a lot of work though
(depending on your experience with image editing software)
yeah i just don't have an endless bank acocunt and I burn through these 13c prompts quick
Yeah, that's very understandable
For now I'd say try to describe the character as a whole and then add bit by bit, so like this
1: "a character wearing a white cloak"
(edits picture a bit)
2: "A character wearing a white cloak and a red cap" and so on
You can also add/edit multiple things at once in a prompt, just not too many :)
If you head over to #edit-inpainting & #characters you'll get superb help! 
any tricks on how to keep things zoomed out for full-height portraits, since from what I hear you can't say anything about the type of shot because you can't say "shot" because (grumble grumble didn'tthinkthatthroughdidthey)
how to make dall e 2 show 6 variations instead of 4?
Run it twice and discard two. It only does things in fours.
It used to do 6, but I guess it only does 4 now?
Yes, If you're using Dall-E, you work for OpenAI
Only Dall-E 2 is open beta.
You probably have two accounts on 2
How do we get DALL-E to zoom out?
got to the dall e bot tab and then input your image and say !pad
its dall e 2
Thanks. What I'm trying to do is get DALL-E to show the whole subject, head to toe. Right now I seem to be getting the middle, like head cropped off or feet cropped off.
can you send a picture of what your trying to edit
im just not sure what your asking
https://labs.openai.com/s/ErGK0rs0zyZKXaEvPR1uUVHb
How do I get something similar but with head and feet in the shot?
i would do command !pad40 then put it back into dall e
OOOOOoooooohhhh. 🙂
lol
go to #dall-e-bot and insert that image then only type !pad40
and then save the image
put it back into dall e and just erase the watermark in the bottom right corner and just tap once in the transparent area
then run what ever you want the FULL image to be
I have a question, is it allowed to use dall e images to make designs to sell
hey does anyone know how to change the password on your dall e account?
Hello all! I was wondering if I should report results that seem completely unrelated to the prompt? Or if that feedback is desired anywhere
here is an example, the prompt was "Sasquatch in a tracksuit, squatting and arms crossed, wearing sunglasses, in front of two expensive cars, fisheye lens photograph"
I submitted quite a few of my generations via the Google Form, but only recently learned about the "Share/Publish" feature, which seems to allow your pictures to be seen by people other than you. Are the generations I submitted, all of which I did via links, not show up because they were private?
Can we make our creations NFTs
Yes you can, just make sure to give credit.
Did you use the traditional sign in method or a sign in partner (google, Microsoft etc.)?
Yes, you have commercial rights to the images, just be sure not to deceive anyone by saying it was not created by DE2.
It is DE2. Dalle was never released to the public.
Are you referring to your excel sheet used to report to OpenAi?
The system reject me this text for policies issues but I can’t find the problem “Maine coon cat drawn by Yoji Shinkawa in style of metal gear solid” I already generated a similar prompt with another cat race without problem
Unfortunately the word after Maine is restricted due to it being a slur. Others found success with prompting it as just Maine cat.
Thank you
Hmm, why is it that users from the UK have to pay a $3 tax and other countries don’t?
ur paying for shipping to the us
moneys gotta get here somehow
🤣 yeah literally that was my thought! I’m sure they’ll say it’s for something like that
@trail orchid Is this after adding your payment method? Or upon clicking "Buy Credits"?
This is after adding payment method, but "Buy Credits" page would click through to blank page before adding payment method..
Oh, this should not happen! The "Buy Credits" button should show this upon clicking before adding a payment method.
Kindly contact support@openai.com for your concern as we volunteers do not have backend access here in Discord.
Ok thanks. I sent email to support.
@verbal crag I addressed this issue buy using another computer. Older Macbook I was using has outdated browsers that can not be upgraded any further..
There seems to be a trend with outdated browsers & Safari regarding the functionality of the website. Thanks for letting us know, glad that worked out for you.
Is it allowed to make a bot which can connect you your own account and download the saved images? Or is any bot usage against the TOS?
and whatabout the watermarks
Stay true to the policies at all times
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Removing the watermark is still a grey area in the ToS, so be mindful of that and decide for yourself.
Please also note that the ToS are currently getting updated AFAIK...therefore changes might happen
sometime in the future, so come back every now and then and double check 
In case you simply want to download your entire collection take a look at this
#starred-ideas message
:)
Can I flip the image before adding the watermark?
are you referring to the watermark added by the bot?
Yes
yes feel free to 🙂
Thanks
How to uncrop images in dalle? Any simple solution?
Would like to know this too ^^ I’ve tried adding landscape and largescale which seems to help a little but some things still get cropped out
There's a bot in #dall-e-bot that helps you resize your image easily.
I think what they mean by "uncrop" is basically zooming out or a longer shot, if that's the case @spring jacinth @tacit coral refer this page for more info: https://dalledictionary.notion.site/Inpaint-Outpaint-3b3a6ec7964f44cfa257f047f45ed400
Yeah but that dalle bot can only expand only profile picture but I want to use other images
Upload your preferred image in #dall-e-bot then type in !pad[number here]
Thanks, I didn’t know there was a dictionary. I’ll take a look
The bot won't take your profile picture to "uncrop" it, you can
upload any image file that you'd like to pad and the bot will do its thing :)
np 🙂
Actually if you dont upload an image, it will automatically pad the pfp.
Hence why you should upload an image :)
I have this image, generated with Dall-E, and I need to add a snake head over the shoulder of the woman, so that it matched the coiled snake body. I tried uploading this image, removing a part of it above the woman's shoulder and adding to the original prompt something like "snake head over her shoulder" and similar modifications, but nothing works, it just fills in the removed part with the background, no snake head.
Any ideas how I can add the snake head?
Try asking in #❌┃prompt-help and the friends over their might have an answer for you 🙂
ok thanks 🙂
Thank you so much 😇
You're welcome :)
I think "guiena pig" seems to set off the contents filter for some reason?
"guineapig" seems fine
How many times can you set off the content filter before getting suspended?
#openai-chatter message
Have a look at this :)
Cool, Ty
I don't think so, I interact with it thru a Google Form, this one https://docs.google.com/forms/d/e/1FAIpQLScDtxdZQ0o_XpbYuJ7c7TaFVc7UEcsDuB5bvEPKOBB7fir22A/viewform
If its not published, im afraid you wont be able to see it, whether its visible to open ai is a question a volunteer or a staff member will have to answer!
why does Dall-e mini have significantly less content filters?
DALL-E mini is not made by OpenAI.
okay thank you
Can you make NFTs with Dall-E (not like i’m gonna do it but still)
Yes, but before doing so, do read the terms & policies found below:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
API Terms & Policies: https://openai.com/api/policies/terms/
Thank you
Is that Dall-e 2? Or first version (just Dall-e) ?
Can't understand which version I'm currently using
ah okay, found in chat that there are beta only for version 2, so it's DALL-E 2, thanks
does editing a description waste a credit?
Yes, editing a description to generate new pictures will cost you 1 credit :)
ty!
Is it allowed to generate a face? (Not an irl one, but a genery one (like, eye colour,hair, etc')
I tried typing "Boba Fett shooting a large scorpion in front of a house" and it triggered a policy violation & possible suspension warning. I looked over the policy and was curious if you think it's because of the word "shooting" or if it's because I asked for a character owned by Disney?
Yes, shooting is the problem here :)
Gotcha. Thank you
Always keep in mind that you should never generate images depicting any of the following:
hateful symbols, harassing content, violent acts, self harm, sexual content, deceptive content,
content related to politics (ballot boxes, protests, etc), or images related to the treatment,
prevention, diagnosis, or transmission of diseases, or images of people experiencing health ailments.

I don't really consider what I asked for to fall into that.
'shooting a large scorpion' is not an act of violence to you?
Not particularly, no. If I asked for "Pest control technician eliminating a scorpion infestation" is that violent?
Like, "a photo of a human face with blue eyes and blond hair"
Technically it is yeah. :)
Either way, everyone has their own opinion so simply
keep in mind that the things I have mentioned above
will get flagged by Dall·E, shooting something is also part of that.
If you're unsure about wheter something will get flagged by Dall·E in the future again, have another read through the polcies here:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Everyone definitely has their own opinion. Shooting in itself is not a violent act though but that's a whole other debate. Regardless, I appreciate your responses.
Feel free to generate images of realistic people that do not exist by using general prompts (eg, a portrait of a person smiling). Generating images of celebrities and public figures is prohibited, unless they are historical figures / no longer alive. :)
If you have a look at #photography you will find a lot of portrait generations! 
Are we allowed to use images we create as album art/cover art for music?
I just got access yesterday and I don't want to lose access by breaking a rule haha
Welcome! Yes you are :)
Please have a read through the policies first though:
Content Policy: https://labs.openai.com/policies/content-policy
Terms of Service: https://labs.openai.com/policies/terms
Hii :) maybe I’m just blind, but I read through the policies twice and I still don’t have an answer to my question..:
if I take an image that dalle generates and change it in ps/procreate etc.. do I have to take away the
or am I not allowed to take it away? Because if I leave it people could think it’s all dalle, but if I take it away people could think it was all me.. I so far always left it because imo it’s still so much more dalle than me.. but not sure if I’m actually allowed to. I know that by adding a description I can clarify but people can not always see one immediately for example if u have it as a pfp or if it’s posted in one of the channels in here
Thanks!
Thanks for that info
im pretty sure you can take it away but you cant say that its your art
There are no official terms for the watermark yet :)
#community-help message
Joanne has shared that they are updating the language for the terms regarding crediting/attributing DALL∙E though,
so we simply have to wait until then.
The best thing to do for now, if you're unsure about the policies etc., is to not remove the watermark
and to clearly state that the image has been generated by an AI.
(I'm in no way associated with OpenAI. Please make your own decision after reading through the policies)