nano-banana

[gentle jasper]

Oh nice.

[limpid ravine]

He didn't change his face, he's the best nano model.

[gentle jasper]

True.

[limpid ravine]

Nano banana va original

[gentle jasper]

Looks nice.

It preserved the detail of her face and body.

[agile oxide]

@indigo forge @lyric pier check out #1397655624103493813 for more info on how to use the bot

[mental marlin]

এই ইমেজ থেকে সব টেক্সট আর লোগো রিমুভ করে দাও।

Photorealistic product ad of a sleek white wall-mounted air conditioner with soft rounded edges and a small circular temperature display, resting on a mossy stone pedestal in a breezy meadow of tall green grass and white daisies; blades of grass gently swaying, crisp natural sunlight, clean blue sky with wispy clouds, low-angle hero shot, minimal modern aesthetic, shallow depth of field, realistic shadows, ultra-detailed, brandless, no text.

[shadow igloo]

ok we get it

[waxen inlet]

Make it diy bottle doll

[hallow basin]

A lush, vibrant green jungle filled with tall trees, colorful flowers, and playful animals, Pixar 3D style, full HD, 16:9 aspect ratio. Sunlight filters through the leaves, animals are smiling and playing together.
Negative Prompt: No humans, no city, no dull colors.

how can i create video

[forest yacht]

you should go to #1397655624103493813 and read that.

[cobalt otter]

Why people think they can generate here, it is clear that it is text channels

[fossil slate]

The name is too meta lol

[cobalt otter]

maybe haha

original

[fossil slate]

Looks nice, i don't see much troubles

[cobalt otter]

hehe

[fossil slate]

And he did the other train :3 i like trains crashed (meme)

... 😏 you verified that the image of the train was the good one, not a too recent train ? ^^ haha

[cobalt otter]

Fr I didn't very care about this kind of detail

[fossil slate]

haha i only see details, too much teacher on me in the past : DO details ! THINK details YOU ARE a tool user not an artist. O_O)/ ouloulouuu 🧙

Anyway, its good work !

[cobalt otter]

Yeah sure

[robust bluff]

#nano-banana add a gray color cap on head and cap logo will be symbol of islam.

[cobalt otter]

🤣

[cobalt otter]

This is not a channel for doing generation.

[robust bluff]

Oops, I don’t know that

[cobalt otter]

You have to use it on the lmarena website.

[robust bluff]

Ok thank you

[cobalt otter]

Or gemini if you have access.

[robust bluff]

Ok

[inland bobcat]

Make a thumbnail

[pallid egret]

#nano-banana Bear throwing blue balls. i need this shot more particle dynamics of balls need lots of ball continue to fall from the basket on woman very fun and enjoy playful shot

[compact mica]

@cinder prairie 1960s mercedes benz car driving in a modern cyberpunk city

[viral mesa]

Nano Banana itself in gemini produce often also more animated or cartoon image results than realistic.

It's not so great in understanding the wish of photorealistic.

It can change perspectives of already photorralistic, uploaded images or photos. Add new things, yes. But when you try to turn an cartoon image into a photorralistic, it has trouble in photorealism.

[hazy onyx]

Hi! Please check https://discordapp.com/channels/1340554757349179412/1397655624103493813 to generate

[keen nimbus]

Do you know is there a model that could do the transformation well, Klux maybe ?

[viral mesa]

Not really. Sorry.
At least not 1:1.

LMArena Website in Battle Mode could do it sometimes.

[bronze atlas]

what is the current best ai image editor

[fair echo]

Gemini 2.5 Flash Image

[bronze atlas]

i saw on tik tok some apple image editor

and its really gooc

good

idk the name tho

[viral mesa]

To change perspectives or add new things is nano banana not bad.

To create aestetic images at all is still midjourney the best. Especially for blending: /blend two, three, four images.

But many a.i. image generators habe their own tiny profits. If you know them.

It's just a question what you want do with a.i..
Editieng an image? Enhancing (upscale and add new sharper details) an image? Blending multiple images to one result? Add new Details? Change only a perspective of the same image?

So, for that i use several different a.i. image generators.
Depending what I want. (the Case).

Photoshop has also still its own profits.

PS. Googles: Gemini 2.5 Flash Image = is Nano Banana.

[potent sky]

I would like the ai can do a micro-edit the character facial expression, most of the ai (even the banana) have generic expression that are too common

Allow user to edit

like in 3D modelling

able to alter new posing live at will, not the bone pose image like stable difusion

[viral mesa]

Sometimes you have to ask an AI chat (LLM) like ChatGPT or Persplexity to formulate a prompt (for an image generator) with your preferences. This sometimes works better. Or, even shockingly, much better.

[agile oxide]

@compact whale check out #1397655624103493813 for information on how to properly use the bot.

[agile oxide]

@hot owl note you don't use this server to do image edit, check out the pinned message for how to use the model - #nano-banana message

[barren ingot]

Lmarena don't work no more

[agile oxide]

how so?

[barren ingot]

Required login now has usage limits doesnt allow

[agile oxide]

It isn't a requirement to be logged into an account.

[autumn nymph]

ok 1 sec

[agile oxide]

@thin aspen please read #1397655624103493813 for more info on how to properly use the bot.

[thin cobalt]

AI-generated grape harvest scenes in Puglia 🍇🍇🍇🍇

[gentle jasper]

I think you might be looking for #video-arena-1 .

[nocturne laurel]

I am currently working on a documentary about the Hinterkaifeck murders, a historical case in which a family was tragically killed on their farm in 1922. For the purpose of this project, I am seeking a way to generate consistent AI-generated images of the farm setting that accurately reflect the atmosphere and historical context. So kindly let me know if there’s any platform or tools that support such image generation, and whether consistency across multiple images (e.g., different angles or scenes of the same farm) can be achieved?

[mystic bolt]

I dont know what I would do. That is beyond life changing. What would you do with that much money! Honestly there is only so much we need.

I would not let anyone know I won

Hi everyone - is it just me or LM the best place to run Nano? I get the best generation HERE!! Other places are not as good. Does anyone know why?

ALso, still having issues being kicked off the chat and being asked to login or create an account to continue. Both on Chrome and Brave. Anyone know about this?

[topaz helm]

Are you italian?

(Sorry for personal question)

[thin cobalt]

Yes, no problem

[topaz helm]

Wow pure io

[mystic bolt]

so just me? LOL

[agile oxide]

Hey - @topaz helm @thin cobalt would note we ask to use English only when chatting in our Text channels.

[topaz helm]

Ok

[thin cobalt]

Ok 👍

[strange ridge]

/image

[mystic bolt]

@agile oxide thanks for your help yesterday. I am able to use nano on Brave

[strange ridge]

/

[mystic bolt]

me

TN

lol adult. whats up?

HA HA OK

GLOBAL here on LMA?

not sure I can help there

good luck!!!

[agile oxide]

Glad to hear it, I appreciate that

[fair echo]

Banana is the best for editing, so that's the way to go.

[autumn nymph]

nano banana best for coding

[clear flame]

wym no??

[short onyx]

zwd

[sacred kettle]

https://tenor.com/bobKo.gif

[forest yacht]

For everybody reading this page: Read the pinned message.
Here is the message just in case you don't know what pinned messages are or where to access them:

How to use gemini-2.5-flash-image-preview (nano-banana)
gemini-2.5-flash-image-preview (previously known as nano-banana) is accessible in Image Generation for Battle, Side by Side, and Direct https://lmarena.ai/?chat-modality=image
This channel should be shut down, it is now fully a channel where people submit prompts for a non-existent nano-banana bot. The owners of this Discord should've known better than to make a channel for a model on a Discord where many new people have just joined Discord to use nano-banana and the video arena and may have a very poor understanding of English.

Once again, this is NOT a channel for putting in prompts, there is NO nano-banana bot here to generate images for you.
You have to go to the website at https://lmarena.ai/?chat-modality=image&mode=direct and generate your images with nano-banana (selected as default, referred to as gemini-2.5-flash-image-preview (nano-banana)) there yourself.
@agile oxide if you are reading this (and you should, that's why I pinged you) this channel should be shut down ASAP as most people are now putting their creations with this model in #ai-creations, the channel has no purpose anymore. It is still just endless prompt spamming.

[potent sky]

Any tips on how should I approach Ai for a question like this? Because What I meant was that sometime a model or the cameraman would've a sudden spontaneous idea and then a new wonderful picture being created, that's like a masterful photograhy

[calm blaze]

Hlo

[grand wagon]

what purpose does this channel serve? nano bannana has been unmasked

it should be renamed to "video gen" or something

and share-prompts to share-prompts (text)

the server is now just walls of images and videos everywhere...

[little fulcrum]

The only purpose is to payback The Nani Banani because they are burning money, paying many people in this world to glorify it. In fact, it's just another type of language model similar to ChatGPT with gazillion of restrictions..

[hazy onyx]

@celest geyser Hi! Please check https://discordapp.com/channels/1340554757349179412/1397655624103493813 to generate content and #nano-banana message to generate using nano banana

Bluedog
[LMai]
— 7:42 AM
@als Hi! Please check ⁠how-to to generate content and ⁠nano-banana⁠ to generate using nano banana

[agile oxide]

Thank you for sharing. I'll be chatting with the team today about potential changes.

[fringe mauve]

is it slow for anyone else

[agile oxide]

appears to be working pretty normal for me, was able to generate in less than ~5 seconds.

is it just nano-banana that's slow for you or are you seeing the same for other models?

[fringe mauve]

other models too

qwen and flux

[agile oxide]

Would you mind putting together a post in #1343291835845578853 and tag me once posted? Include as much info as you think is relevant

[fringe mauve]

alright i will do it once i am free in an hour thank you

[agile oxide]

@wheat geode if you're looking for the Video Arena bot be sure to check out #1397655624103493813 for more info

[graceful magnet]

sorry, we cant post unreleased models?

even if anonymous still?

[agile oxide]

I was removing more-so because this channel is for discussion related to nano-banana, but you posting in #general (like you did) is all good 👍

[graceful magnet]

tyty

makes sense

[agile oxide]

@arctic mason Be sure to check the pinned message - #nano-banana message

[arctic mason]

yes and? what if I do it on gemini I can't post here?

[graceful magnet]

i think he was showing the prompt result

[agile oxide]

Ah okay, that's my bad.

[arctic mason]

it was an image made with nano banana

[rocky grove]

@agile oxide

do you have the new openai image model?

on lmarena

[agile oxide]

You can check what models are available in the dropdown menu in Direct & Side by Side.

[little fulcrum]

Hopefully he will finish running tomorrow, not next week..

[gentle jasper]

https://tenor.com/view/joe-biden-presidential-debate-huh-confused-gif-9508832355999336631

[midnight meteor]

nanobanana starting to give random results no matter what reference image i upload……

[gentle jasper]

Hm.

[midnight meteor]

Also the “something went wrong” up significantly

[gentle jasper]

Interesting.

[little fulcrum]

Why is it banana, by the way?

[agile oxide]

Do you have an example of this you can share?

[rocky grove]

random name

[agile oxide]

That looks stuck, I'd start a new chat.

[little fulcrum]

Hmm... Nano Cucumber.. 😏

[midnight meteor]

probably found the reason. in one chat window, lets say i uploaded [image 1] along with a prompt. later i uploaded [image 2]which is entirely unrelated to image 1, yet result gives me something related to image 1

is that a bug or cache problem within a same window?

[agile oxide]

Oh interesting.

Okay yeah I'm going to look into and will report.

That doesn't sound like expected behavior

[little fulcrum]

I don't think it's specifically related to lmarena. Happened to me when I run the model on Google's AI Studio. The Nano Cucumber ignored my prompt and gave me a monochrome photo of a cowboy bringing his children to a ride on a horse .

And I've just realize that in these last 3-4 hours, the image results from this Banani are getting less photographic, even when negative prompts to avoid CG or 3D inserted. Seems like they have more bigger tension than lmarena, after their avant-garde bombardment to all the news and social networks in the universe to glorify their existence..

[frozen tendon]

Bro waited 20 minutes

[frozen tendon]

Yeah the lmarena's banana struggles with two images

Did you try Google's ai studio

[midnight meteor]

sounds like a reasoning degradation which often been seen in chatGPT

happens during overload status

[midnight meteor]

I totally gave up on Ai studio, performed worse than arena since the beginning for banana. Always referred to irrelevant uploads in a chat box.

[little fulcrum]

I personally hope that lmarena includes image-to-image from Seedream and Imagen 4. Both are really good on Dreamina and Whisk..

[midnight meteor]

Also heard Seedream 4.0 on live since Monday

[little fulcrum]

Never tried this one. I believe what is used on Dreamina is Seedream 3.0

[midnight meteor]

I saw a bunch of posts from my China social media groups for 4.0, quite amazing if true, a comparable model to banana especially it outputs 2k.But my seedream app hasn’t offered 4.0 yet lol

[gaunt chasm]

Good day everyone! Just wanted to ask how many image to video are we allowed to create here 🙂 Wanted to make sure first before I prompt unnecessary things 😄

[little fulcrum]

8 per day

[gaunt chasm]

Thank you my friend. Appreciate you! Do you happen to know what time it resets?

[little fulcrum]

They have serious steps about it. Try to compare face recogniton between Imagen 4 on Whisk, and Seedream 3 on Dreamina. The Chinese one is much more has accurate reading..

[fair echo]

#1413627255988686950

[little fulcrum]

I don't know, Bro. Haven't use video generating so much..

[wary spindle]

Is there any way to turn off multi turn image editing?

[gaunt chasm]

Still thank you bro! Lastly, do you guys have any suggestion of similar AI with LMmarina that generates image to video for free?

[wary spindle]

I really liked that non bannana was not multi turn

Any way to restore that?

[little fulcrum]

Try "aistudio.google.com/prompts/new_video" They give 10 free Veo 2 videos. For Veo 3, try Whisk or Flow on "labs.google/fx/" each are giving 5 free Veo 3 videos.

[gaunt chasm]

Thank you! You're a rockstar bro!

[wary spindle]

You know how to turn off multi turn image editing?

[little fulcrum]

I don't even know what is it..

[fair echo]

#announcements

[wary spindle]

It’s when your chat history for image editing doesn’t reset

Nano banana gets worse with longer chat history best edits are fresh

Don’t want to start a new chat every time for image edits please

[little fulcrum]

Dang.. That explains why I've got so many delay lately. You should ask one of the moderators about it..

[dapper snow]

nano nanana

[fair echo]

Nani Banani

[sick oxide]

Someone else is failing you lm arena

[little fulcrum]

When Nani is Japanese, and Banani is the name of my Arabic friend..

[fair echo]

https://tenor.com/view/nani-huh-what-mustache-gif-24250081

[sick oxide]

The page crashed

It fell or saturated

[little fulcrum]

🤧

[agile oxide]

The site is down for you?

[sick oxide]

It doesn't work for me, it doesn't let me edit even in kontext flux

Not if it's because of the comet hahaha

[little fulcrum]

The fate kicked me out to images with Gemini's watermark. Again..

[sick oxide]

It's good to saturate the page

[rocky jewel]

From my tests Seedream 4.0 was better than Nano on most tasks

though not all of them

[midnight meteor]

@agile oxide is opening a new chat the only way to aviod multi-turn editing? let's say i don't want context window information has a single effect on my latest prompt

[rocky jewel]

Seedream is better at keeping the image consistent with the original, here's an example

[hot lion]

Suddenly Gemini 2.5 image generation is not working!
It is processing for a long time. 4/5 minutes it still processing.
Why is it happening in a sudden?

[rocky jewel]

"Exact same pose but camera looking from the side"

Nano Banana / Seedream 4

not perfect but seed is much closer to the original pose

[agile oxide]

I believe so, but will double check on this.

[hot lion]

Still processing
Usually it used to take maximum 10 seconds to generate
But suddenly it is taking so long and not even generating anything!!!

[midnight meteor]

Very clear example👍🏻nano understood it as if character is looking aside

[agile oxide]

There are some issues that we're looking into.

[rocky jewel]

probably

[hot lion]

How long will it take to solve the issues?

[agile oxide]

I don't have an ETA for you sorry to say

[agile oxide]

yeah confirmed

[hot lion]

ETA means!

[midnight meteor]

also i think nano has a serious shortcoming when it comes to text other than english

[agile oxide]

Sorry that stands for "Estimated Time of Arrival", meaning I don't know.

[midnight meteor]

even gpt 1 does better on mutilingual text generation in image

[hot lion]

Get it

[rocky jewel]

tried "show how this would look from behind, same pose, especially the arms behind the back" instead

Nano Banana / Seedream 4

[little fulcrum]

From what I saw on Gog's AI Studio, each chat have what is called "token," and it's limited. When a chat reached the limit, you should create a new one. I don't know is it the same in lmarena..

[rocky jewel]

(nano looked a little sus so I covered her to avoid getting banned here)

[little fulcrum]

Nani always sus on everything..

[wary spindle]

Is there a way to turn off multi turn image editing for nano?

[rocky jewel]

"Make an illustration of this character, in this artstyle, with a dynamic pose and exaggerated fish eye lenses foreshortening perspective"

Nano Banana / Seedream 4.0

both are first results

hopefully Seedream 4.0 is added to lmarena in the future

[obtuse tangle]

Where we can try seedream 4.0

[rocky jewel]

only chinese websites like space.coze.cn for now

a bit of a pain to get access to since you need a phone number

[midnight meteor]

It originally announced went on public since Monday, but seems like a rumor

[obtuse tangle]

Chinese number or any number works?

[rocky jewel]

I'm brazilian and it worked for me, but I had to try multiple times since it kept giving "Too many attempts" errors

[obtuse tangle]

Fake site ?

[rocky jewel]

yeah

never trust websites that have the tool's name

[midnight salmon]

How can I make him wear this it never work

[midnight salmon]

Ok

[main flicker]

nano banana is failing ?

[hazy knoll]

Keep Bruno Fernandes by my side

[fringe mauve]

Taking this much time to generate 1 simple prompt.

ok it was faster on the app

so issue is with arena ig

[midnight meteor]

Don’t know if it’s google’s problem or the multi-turn edit, image quality and accuracy is horrible now

Also nanobanana now gives exactly the same output from last result

[fringe mauve]

i have the same issues too ^^

keeps spitting out the image from the previous prompt

[wary spindle]

It’s because of multi turn image editing seriously benefits nobody

[fringe mauve]

who asked for it

Also generation is slow today

sometimes fails completely

Even simple prompts

[wary spindle]

Was snappy this morning before multi turn image gen 😭

[midnight meteor]

Honestly it’s a bad idea to incorporate multi-turn edit cuz I believe it just brings a lot more messy situations than actually benefits…

[obtuse tangle]

And how about censorship

Is it more censored than nanobanana or less

[rocky jewel]

less, but it's not like it will do anything either

[agile oxide]

There is not; however, if you start a new chat that'll effectively "turn-off" multi-turn.

[agile oxide]

We were seeing a higher error rate earlier; however, this since has been resolved, are you still seeing errors?

[agile oxide]

Can you help me get a better understanding of these messy situations so I can bring to the team?

[midnight meteor]

Repeating previous result is one big issue. Plus, it seems like responses weigh too much on previous result, which cause new outputs ignore latest prompts and reference images. If only solution is opening new chat window, then it will be tons of windows to manage for each project or random art thoughts

Technically speaking, if there is no editing tool directly done on the outputs(like REGION tool in MJ or sora’s edit), word based edit or remix is quite inefficient for nanobanana at least for now.

[wary spindle]

Can you please add an option to turn off multi turn? It was so much more convenient to use nano banana on LM arena due to no multi turn, many people will find this feature very valuable if possible

Now I feel bad for not taking advantage of nano on LM arena sooner I assumed it would always be this way

Let me explain, the context of each edit gets muddled by the next, for true multi turn editing you can just screenshot the previous photo very easy on Apple devices, and immediately feed it in for the next edit

You get higher quality edits every time

[wary spindle]

Let me know if you find any solution to this

[little fulcrum]

Ahh.. Nano Cucumber..

[lilac storm]

Why not work the website to make AI photo?

Using the nano-banana model, turn this image into a realistic photo of a collectible figure holding a banana, in realistic style and environment.
Place the figure on a computer desk. Next to the computer screen; place a packing box with rounded corner design and a transparent front window, so that the figure inside is clearly visible. The box should also be printed with the figure's original artwork and "PI " logo.
On the computer screen, display the Blender modeling process of the figure.

Now can’t work the code? Why?

[gentle jasper]

Well, for certain things, it just isn't really good at them.

You could try using Qwen's image editing model for that, though.

It's sort of the same quality as nano-banana.

I've tried it so I know this.

[lilac storm]

I see, thank you

[gentle jasper]

My pleasure.

Even if I'm not a mod, I still like to assist people with their needs.

[main flicker]

stuck on the generations

[agile oxide]

Sorry to see it's stuck for you, I've already flagged this to be looked into.

[agile oxide]

It's possible, the feedback will be shared

[wary spindle]

I’m not sure if your aware but in the animated show king Julien, a spin off of animals of Madagascar, there is a pantheon of gods let by their leader a pineapple lol, so I plead that the pineapple shall hear my requests

[fringe mauve]

@agile oxide generation is stuck for me too

the flux generation came in 9 seconds

[agile oxide]

Lol I did not know this but good to know

[long hornet]

how does lmarena nano banana have no visible watermark?

do they not have that watermark on the api?

[gentle jasper]

They do, but it's a toggleable thing.

And I suppose LM Arena just has it disabled by default.

[long hornet]

ok so if i pay for the api i can toggle it off

[gentle jasper]

I believe so.

[main flicker]

i think that i found a bug. basically is like to if old generations stuck in the same chat, mix with the new generation try

[gentle jasper]

So basically you're saying that if old chat logs somehow get stuck within the current chat room that you're in, that it'll attempt to mix with the new chat logs?

[little fulcrum]

Proven good. But it's too good for the details, so you need to smooth up the prompt to avoid "over-detailing" when editing photos.

[lapis gazelle]

Hi, all

[gentle jasper]

Greetings, Stephen.

[gentle jasper]

#video-arena-1

[short lion]

A strong, confident man standing in a dramatic pose, looking to the side with intensity. He is wearing a fitted black shirt, slightly unbuttoned at the top, with rolled-up sleeves. His style is sharp and bold, paired with dark trousers and a belt featuring a round metallic buckle. The background is blurred with dark, moody grey tones, creating a cinematic and powerful atmosphere.

[tidal harbor]

hello

[midnight meteor]

can we really fix the multi edit thing? now keeps giving same outputs from previous chats or rerolls

[fringe mauve]

whattt

previous chats too?

[midnight meteor]

yes, not only previous chat in the same window, but cross window as well

[cobalt wraith]

Is nano banana still having problems with errors and extremely long wait times for everyone? I can still generate in one of my chats with 2 images but attempting 3 images in my other chats results in failure.

[mystic halo]

yeah, it's still too slow here

[cyan mulch]

hello guys i have question its possible to have a private romm to generate a videos ?

[sullen nova]

Can someone help for a prompt?

Let's say I want to apply a hairstyle from a reference image to my image. Suppose those images are named as 'ref.jpg'(reference image) and 'me.jpg'(my image).

Can you please give me a prompt which works accurately everytime? (I faced some errors while trying my own way)

[fringe mauve]

what if i use incognito

also btw mods nanobanana seems to be slowing down everyday right around this time

while others work well

i am facing same issues i expressed last night

it was working fine few hours back

[round birch]

@round birch

Pic

[night tapir]

/text-to-image

[fierce ibex]

Hello @pastel whale and @night tapir please check this link to learn how to generate content on this server: https://discord.com/channels/1340554757349179412/1397655624103493813

Nano-banana is only available in Battle mode. More info can be found in the pinned message ⁠#nano-banana message

[proper bronze]

nah its also available in direct chat mode

[fierce ibex]

Great, thanks for the correction.

[dull thistle]

can u specify image dimensions? or is it all squares?

[agile oxide]

Adding aspect ratio is something we'd like to add, but isn't available right now sorry to say.

[agile oxide]

@dim pollen @sleek crag looks like you're looking for Video Arena, be sure to read #1397655624103493813 for more information on how to use the bot. Let me know if you have any questions!

[regal hazel]

https://tenor.com/view/cat-cat-hit-cat-hit-and-boom-cat-boom-cat-hit-and-boom-kat-gif-15865521055253202019

[hushed quail]

pretty decent meme honestly

[next jetty]

replace design on her hoddie

[viral mesa]

The more precise you go into the details, the more precisely the AI ​​will know what you want. If you leave something out, the AI ​​might misinterpret it. Then try to correct it and explain what you want more precisely.
AI is only as intelligent as it can bind itself to information. Otherwise, it will freely make its own decisions that might not correspond to your wishes. So be precise. Guide the AI ​​where she should go. That is, how she should serve you.

[viral mesa]

But don't expect always a wonder. Ai is only as such good as many informations in a databank it has and how well he was trained. Ai is still far away from perfect. Some Ai's are still very "stupid".

[little orbit]

Prompt: "Create a high-quality, minimalistic design showcasing a product with a clean, modern aesthetic. The image should feature a white Air Max sneaker viewed from a top-down and front-facing angle. Include annotated lines pointing to specific parts of the shoe: 'Solado' (Sole) and 'Contraforte' (Heel), each with detailed close-ups of the sole and heel sections. The background should be sleek and white, with sharp contrast between the shoe and the minimal lines. The text 'AIR MAX' should be prominently displayed in a bold, contemporary font at the top. The design should be clean, professional, and tech-inspired, making the sneaker look like a product from the future."

prompt: make a professional design for the purple airmax in a cartoonish style like this

[fringe mauve]

Puny god

i win

[gentle jasper]

Amateurs...

[fair echo]

You don't generate Images here.

[proud cave]

student learning physics in out of the box philosophy

[hazy onyx]

@dim pollen @left steeple Hello! 🙂 Please check https://discordapp.com/channels/1340554757349179412/1397655624103493813 to generate content and check #nano-banana message

[distant kelp]

Question, is it possible to somehow fix the nano banana glitch? When you upload two images and ask to transfer elements from the second image to the first, it takes the same cropping from the second. That is, if there was a rectangle there, and the image you want to transfer to is square, it will make it rectangular

[hazy onyx]

@quiet pike Hi 🙂 if you´re tring to generate content check https://discordapp.com/channels/1340554757349179412/1397655624103493813 nad here espcifically for nano banana #nano-banana message

@hollow bough ⬆️

[agile oxide]

Would you mind creating a post about this in the #1343291835845578853 forum and provide details (& screenshots if you have them)?