flow-general・🌊

[stark void]

Then not inactive screen or if working laptop then change tab during creation

[hexed latch]

Guys what is the best text to video ai?

[covert osprey]

do I use same browser and acc?

cant use a different acc cus I need to verify the others

[stark void]

My mean was don't change tabs during creation of image

[stark void]

I didn't understand

[covert osprey]

I dont change and still keeps showing that same thing

[crisp basalt]

screenshot this whole chat conversation, and ask chatgpt/gemini to translate in your own language

[covert osprey]

how do I fix this

[stark void]

Try to clear cache of browser

And refresh tab before prompt again

[dry frost]

Or you could just wait a few hours and try again, like the error message told you.

[covert osprey]

didnt work, waited like 4 hours

[dry frost]

Huh... well I've never got a request too quickly message before

[stark void]

Any result

[covert osprey]

nano banana 2 works but pro doesnt

and pro is better I want to use that

[green needle]

Any news about being possible to transfer projesct from whisk to flow?

[bold falcon]

General question, if I get this message with Ultra, what chance do others have? Does API get all the priority. I mean it's Sunday.

[pulsar coral]

Paid plans are being more resistant to this, however Nano Banana Pro is the one that seems VERY limited in use despite whatever plan. If you're having issues try swapping back to Nano Banana 2 and good luck with that.

And if all models still fail yeah just gotta come back in few hours.

And from the amount of people that are having probs, it does seem way more want to use Nano Banana Pro than 2.

[bold falcon]

yeah I got it on a save file for Veo 3.1 on an extend video. Honestly, it acts like a scaling problem. So hopefully tomorrow when they all get back in, they can add a little backroom magic and make it optimized more. Thanks.

[bold falcon]

Okay this is promising and I changed up a little bit of what I'm doing and it's worked on the last 3 time-lapse videos I made today. I'm not submitting a second job until the first 1 displays that it's done. Then I go to download and click download and I'm not shutting down any of the pop-up boxes with the dismissed dialog. What that means is I'm marrying them back up with DaVinci Resolve Studio as individual clips, not using the scene jobs. So this looks promising, maybe I was just sending too much to it at 1 time. So if you are sending multiple jobs, let 1 job finish maybe and then submit the other 1 once it renders. I'm doing this for 2 iterations of images for each scene and 2 videos using fast with a 10 credit burn. and letting the extension run 1 at a time with a 10 credit burn as well. Good Luck today all. Hahah, quick edit, I'm going to try scene editor after they all complete. each video is one thumb, 4 scenes. I'll let you know how that works out later. I have 5 more to go.

[icy bobcat]

Hello. One of my Google accounts doesn't get full restorage of AI credits. The other ones get 150 credits every midnight. What might be wrong? Am I supposed to reach below a certain number before the system makes up for the loss?

[bold falcon]

Not sure which plan you have but my credits don't restore until my billing cycle resets. So that would be once a month.

Oh and I solved the dismissal box issue they should just throw it up here, lol.... #1, Make it So!

[stark void]

Create a new short for YouTube using this flow
https://youtube.com/shorts/ufC-9fW3m2E?si=rGKoKV6r3Q5X2DnY

[bold falcon]

okay that did it my other 7 thumbs as well as 35 videos without a single error as long as I'm waiting until each job is done before submitting the next 1

Okay, I'll pump it in and see what I get out of it. And then I'll post a result here.

[winged mantle]

Why do I get "failed something went" wrong message for the past 1 hours on image generation

[bold falcon]

Hahahah, Here's what I'm starting with...

[rotund elm]

how do we add any clip to a collection from the lcip itself

[stark void]

Now create a new one and this time a something better
https://youtube.com/shorts/iJQUti14RD4?si=vF9I2h0yaexTWYGD

[bold falcon]

Publishes in six minutes: https://youtube.com/shorts/svl4DhhIQGs?feature=share

[bold falcon]

It's Sunday, lol.... I I have to finish my uploads for tomorrow first.

Darn it. Too late. I'm already starting it now. Grrrr. lol

[stark void]

Your video isn't showing

[bold falcon]

I know I delayed it for another 3 minutes so the 4K version loads.

[stark void]

Good lol

[bold falcon]

Oh, you're gonna love this garage one I'm doing. Last 1 though, after this I gotta get my work done.I'm trying to get a 1,000,000 views off of 5 platforms in 30 days. But I want mine to tell a story, not just a build.

What'd you think of the kitchen 1? I didn't even try very hard on that 1.

OK, All done with the garage one. Let me know what you think of this one too. Gets published in five minutes: https://youtube.com/shorts/x2nIvWPqak0?feature=share

Here's the prompt I used for my garage, lol.

[wild quartz]

A bean field, a dirt road has washed away, there is a rice field, an old woman wearing a large hijab burqa and a woman wearing a large veil, hijab and burqa, sitting picking beans, the woman has a large bamboo plate in her hand, the woman's face is covered with a niqab,

[full flicker]

Hey everyone! Just a friendly reminder to keep our discussions focused and to avoid excessive emojis, spam, or content that doesn't make sense. We've deleted the recent messages that didn't comply. Thanks for helping us keep the chat clean and engaging for all members!

[strange monolith]

finally updated the aspect ratios in flow

[blissful abyss]

what does this error mean Flow is currently experiencing high demand. Things might not be working as expected, please try again later

Is it that flow will not run well or what

[wooden prism]

can someone help w the ingredients option i understand what im supposed to do i just cant figure out how to do it n all the vids on it from older versions

[pulsar coral]

If you payed for a plan in that account the credits in it don't restore daily, it only gets what it got as soon as you payed, like the $5 plan comes with 200 credits. The free accounts do restore like 60 credits daily, at least that's from my location.

Idk what's logic behind that, but it is what it is.

[blissful abyss]

50 credit and 100 for new users

[wooden prism]

do i @ my 3 imgs i wanna use for the ingredients n its good or ?

[blissful abyss]

just add your images

but for frames first and last

[wooden prism]

i dont wanna do the frames

[pulsar coral]

150 for new ones from mine.

[wooden prism]

so this

[blissful abyss]

then use ingredients then prompt

[pulsar coral]

Seems to be regional.

[blissful abyss]

then u can the maybe the @manimg standing in a @boatimg

[wooden prism]

its just 3 diff angles of the same hoodie

[blissful abyss]

Nope i got 150 too but I mean only 50 will be renewed daily

[blissful abyss]

then just prompt

[wooden prism]

appreciate you

[blissful abyss]

Has it worked

[wooden prism]

generating

[blissful abyss]

U joined this server a day before one year you created your discord account

thats nice

[wind rapids]

does anyone know how we can use the syle feature on flow or any way to get a similar result?

[pulsar coral]

That's my attempt at Flow, it might not look bad, tho if you've already seen how Whisk's Nano 1 does this stuff, you probably can't unsee now.

Probabaly only the middle one looks ok enough so one would have to refine more from a simple prompt like that.

[blissful abyss]

you can also use the @ feature

[pulsar coral]

Which would explain why is Flow been straining so much, you need some good 10 attempts to really get what you want from Nano 2, while Whisk's Nano 1 would get it at the third one or so

[pulsar coral]

Yeah or maybe the prompt needs to be less strict and say something like

"Keep it true to the first uploaded reference image. Base the image of the cat on the cartoony art style of the second uploaded image."

[blissful abyss]

just use the @ featyre

[pulsar coral]

"Base" instead of "Make"

[blissful abyss]

U can just use "keeping image 1 the same transfer the style in image 2 to image 1"

[pulsar coral]

Yeah or "Plaster" has worked better for me, forgot that one

[wind rapids]

what is this feature you guys are talking about?

[blissful abyss]

transfering a style from one image to another

[pulsar coral]

or can also be done more precisely like

You just just gotta explain what you want and it will get it, I'm sure my prompt there can be rearranged better in a single line, this was just an idea test

Literally telling it to "Do not add effects" removes all the superhero effect from syle image. Nano 2 does give more complete control, but it's a little harder to keep it like it.

[wind rapids]

the problem is that we need to arrange things almost perfectly meanwhile whisk does the task with simple instructions

so imagine for someone who is generating 100+ images, it's pure torture

[pulsar coral]

Ah yeah I'm still in love with Whisk's easiness too, specially the default compositions it picks. Gotta save some pics of it to try to tell Flow's models to reproduce it.

Anyway point is with Nano 2 one's at least gotta try.

[blissful abyss]

But the realism on whisk is bad

[pulsar coral]

idk I'm only trying anime and games, who even needs reality.

[blissful abyss]

I need

[wind rapids]

i need the cartoon style and flow is not very good at that it seems

[blissful abyss]

flow is just better with reality look at this

[pulsar coral]

Is that Nano 2

Imagen 4 can also get that type of image surprisingly well, but I don't think it'd get much better than that

[blissful abyss]

yeah

[pulsar coral]

Now I added "keep the tonalities true to the image of the cat" that got rid of the random palette.

Just gotta talk to it

Eventually you can save a functional layer prompt

[wind rapids]

whisk vs flow. second image looks like something chatgpt could do in the beginning, which is not very good yeah

[pulsar coral]

Now about the cat that came out as "Cubism" that can be fixed by also adding "Keep the image's structure true to reference image"

I think that should keep it there

[pulsar coral]

I think it's the colors mostly, ending prompts with something like

[Image Quality: All professional but completely true to reference image including vivid vibrant tonalities also accurate to reference image. Remastered 4K HD ultra high resolution, sharpest definition, denoised, unblurry, clean lines, High contrast, fix any jpg artifacts.]

It's mostly the vibrant vivid colors that seem to be missing there.

But yeah it does start piling up

I must have 3K token prompts somewhere, Whisk's nano wouldn't need to be explained everything

There's other more complicated tricks about artstyle and compositions, that Nano 2 can barely get but still seem to work better in NB1

[wind rapids]

is there any other tool similar to whisk?

[pulsar coral]

Wherever you can find original Nano Banana (Gemini 2.5 flash image) should work around the same, tho only Whisk seems to have the little trick about styles and scene. Google AI studio still has NB1 as introductory model but only gives few free tries daily.

It can't be "Pro" or "2" usually sites only call it plain simply "Nano Banana"

[wind rapids]

i will miss the little trick about styles so much...

[pulsar coral]

Yeah I think the style and scene features are the second most requested feature right after aspect ratios, they need to take it into consideration ASAP.

Only they can read whatever prompt layer Whisk has for it and everyone's so in love with it for whatever reason.

[wind rapids]

do you think they will?

[pulsar coral]

They did add aspect ratios so there's a chance. Gotta make a thread at "Labs-bugs and feedback" too.

That's worth of a "Today's special"

[frosty niche]

I had so many 🪲 reports last night it was nuts.

[pulsar coral]

Also if you knew who that cat is, it's perfectly capable of that.

[frosty niche]

Hey those are only on weekdays guy!

[pulsar coral]

Yeah that's why we be posting cats

[frosty niche]

⚫

Those old dolls don't even need a seatbelt i guess.. 🙄

[cyan kestrel]

Hey, not sure if this is the right place to ask this, was the scene builder removed from flow, I haven't used it in a while but when I try adding to scene i get an error .. a 500 error from api/trpc/project.createScene

[dry oar]

i've been getting 500 errors too, they usually clear after waiting a bit and refreshing

scenebuilder is still there

[cyan kestrel]

thanks.. i thought i was going crazy.... it seems like some projects work ok others do not... also when looking at the request it looks like it was sending a request for a landscape sceene builder even though i was trying to send a portrait vid ... oh well.. crossing my fingers.. 🙂

now I'm copying videos from one project to another to make it work 🙂

[pulsar coral]

It's probably Sunday not in working hours for Google atm, they should be getting on it in a few.

[tidal torrent]

how many free video gen can you make per day
how many free daily credit you get and is it enough to make 1 video

[sage otter]

Devs have done a good job making flow's ui dogshit. Just wanted to ask, can we push this further and make it even worse? Thanks

[tidal torrent]

this youtube that is pretty well rated said that it's 150 per day, is that accurate?
if its not how could this video get so many views and kinda highly rated ?
https://www.youtube.com/watch?v=QlzLbWp92YE

[pulsar coral]

He probably means 150 credits not 150 videos. And more often at free accounts people just get 100 credits for their newly made account, and then 50 free credits daily, so that's why he probably saw 150 credits in one of his accounts. I also coulda sworn my new account got 150 credits for day one too, but here we are.

Each video uses 20 credits.

But if you pay for a Google plan, then you get a certain amount of credits and no free daily credits anymore until the end of month, like if you pay for the $5 plan you only get 200 credits for the whole month and no more free daily ones

Why'd it work like that beats me, sounds like really bad planning but yeah I have an account like that and can confirm.

[pulsar coral]

idk I'd say the UI is survivable and even nice once you get used to it. The lack of original Nano Banana is the actual deal breaker here.

Also Nano Banana 2 has an awesome auto aspect ratio feature that can be controlled by prompts like "Aspect ratio is 2:3" and so on most any you'd like, Gemini has it but Flow doesn't, which is a complete waste of the model's capabilities. And it's something original Nano can't do unless you already made the image in the aspect ratio you wanted and the site has Auto aspect ratio enabled too.

[pulsar coral]

Man Grok's charging a full $30 only and no cheaper plan? That's nuts. I know it's just like $1 daily, but you're not gonna be using all those vids in one day, I don't think every human sits doing videos every day for hours to not waste those $30 at the end of month.

Or DO they...

Flow does seem to have a better plan there, if you at least pay for Pro when there's an offer

Not to mention your Flow Google account is the same one that connects to Gemini and other services, so you actually get double the value there. Including more use of Nano 2 there.

[sharp trail]

heyy guys is anyone online ?

idk why but im getting watermark on the videos im generating in flow with ultra account

and its happening only today

until yesterday everything was fine , so can someone help

[crisp basalt]

if thats what u want sure. weirdo

gettin this pressed over ai video tool is crazy lmao

[sharp trail]

hey bro you know smt about this watermark stuff ??

[crisp basalt]

sure. what country are u from

[sharp trail]

india

???

[crisp basalt]

im not a google staff lol calm down let me finish my thought

I can't confirm exactly how many countries are on their list for the watermark even for Ultra users but...

[sharp trail]

is that really what they are doing bruh

[crisp basalt]

this is the only thing Google staff have said about it, and it's been an ongoing discussion since earlier this month

yes , no solution for now

sorry

thankfully i didnt get the watermark thingy even tho im on ultra. and im from malaysia

[sharp trail]

wth bro i js bought this subscription for client work , and cant use clips with watermark for professional work , this is nuts

[sharp trail]

what do you think using vpn would work ?

[crisp basalt]

uhhhh, i cant confirm if that would work becuz

My theory is that Google checks your account's region based on the country tied to your payment profile

[sharp trail]

thanks bro

now what will i do

i have a project i need to complete today , this is impertinent from google

[crisp basalt]

idk tbh , . It sucks because the 'no watermark' feature is exactly why most of us paid for Ultra

Google Labs Team/Staff really dropped the ball by not warning us about this. People paid for Ultra purely for the 'no watermark' benefit, so this is completely unfair

[crisp basalt]

Give it a try tho

[pulsar coral]

We're all off to your house then. Or a residential Malaysian VPN.

[craggy belfry]

My images have been trapped behind a Collection wall. My images suddenly went into a collection. I did not want that. Now my images are gone and instead there is a square box that says "Untitled Collection" It then spawned another box that says "Untitled Collection". CLicking on one box just changes the image displayed in the other box. How do I get my images back and prevent collections form happening?

The instructions at google about collections oes not relate to what is actually on the screen, I can find no way to remove images or access images in a collection

After numerous tries and leaving the project and coming back to it I was finally able to access the images in the ever growing number of collections. Now I would like to know if there is a way to prevent collections from happening?

[pulsar coral]

Just be careful where you click, also you can just reupload everything anyway. I did just ask again for a button to unsort collections at feedback... cause tbh me and all could also use it.

[craggy belfry]

OK. Thanks. I hope they put it in. I am using Flow for the first time and I miss Whisks ability to have images as subject, style and background. I guess I could write it all out in a prompt?

[vocal bridge]

Who’s sharing the Flow website? It’s in high demand right now. Image generation is currently free, so please don’t leak it. If usage gets too high, they’ll likely remove the free access.

https://giphy.com/gifs/crying-matthew-mcconaughey-interstellar-l3vR4CdLInXOhr3rO

[crisp basalt]

The extension injection user

[vocal bridge]

whats that

[crisp basalt]

People who purchased an ultra plan at a low price from a third-party seller used a very unsafe method. This method involved injecting the account into the Google Flow website using an extension.

[vocal bridge]

oh

i hope they patch this method

and ban any third-party seller or buyer

so usage gets lower lol

[crisp basalt]

They’ll likely remove the free benefit if people abuse the system. Few months ago, Google removed the 0 credit video generation for ultra users temporarily due to high demand and abuse of the Google Workspace trial method to access the Ultra plan for free or at a low price.

[pulsar coral]

It really doesn't matter, with exposure or not things eventually stop being completely free, because there was a good reason for it to be free for starters, go check the youtube vid above about Grok for reference. Once the beta phase of things is done, you as free user are done too.

Best you can do is move and learn as much as you can while you can.

And if you got really good at the tool, it'll be worth paying.

[vocal bridge]

damn

but im poor asf

hope the free access stays atleast for more 2 years

@owner

@googleowner

bruh

[pulsar coral]

Someone please explain this dude how to do that @ Whisk trick about the prompts, I don't seem to be good at it, thought my prompts were fine so idk. @glad yacht

[pulsar coral]

I think at worse they'd keep the $5 tier, it's the one I use and lasts long enough.

In fact I just woke up thinking about that and started practicing again.

[wide moon]

How cuz they’re getting rid of imageFX and whisk

[vocal bridge]

bruhhh

we're cooked..

[vocal bridge]

i dont even have 1usd in my bank dude

[wide moon]

Of what

[vocal bridge]

free access

[pulsar coral]

Who am I to judge anyway. Everyone has hobbies.

[wide moon]

If they did they wouldn’t consolidate whisk and image FX into flow

[pulsar coral]

Well admittedly Google has kept the beta test phase of everything for very long, maybe they'll still put up a Nano 3 or something later and they'll still need people testing it and sending feedback

Training models cost an F ton, so that's likely what we're doing here.

[wide moon]

With the much higher stuff like nano banana pro etc that’s stuff you’ll have to pay etc

[wide moon]

I’m seeing it for pro users

[pulsar coral]

idk much about image model costs, but training the language models can indeed cost millions, we're probably more cost effective than we think as long as we send feedback

[wide moon]

Pro users gets first dibs on beta stuff before testing it out for public use

[glad yacht]

The collection in projects is a folder 📂 for you to organize photo/video in group. You can create multiple of them in the same project. You can drag and drop to move files in and out from collection to project on the desktop environment.

[gloomy verge]

Hello. Can I ask question? Fast video is so in bad quality? Because I tried create new video (8 sec.) and when I written prompt (Girls turn off to pool and jump to water), girls´s head rotate so strange, when they must rotate to pool. Thank you.

[frosty niche]

Pat, did you ever finish that western piece man?

[glad yacht]

Try Start and End frame to create the video 🤔?

[gloomy verge]

No, because, I wanted create some western video with this desperado, but I haven´t any ideas for story for movie, so I deleted it. 🙁

[gloomy verge]

Will it create a better quality video?

[frosty niche]

You had some decent material there, too, man. Well, a clean slate is a good start too.

[gloomy verge]

I want to create some short movie too, but I haven´t any ideas, what to create. And second problem is, that it is in my opinion useless, when I have this problem, bad rotated person, bad rotated head, automatically enlarged pool after jumping to it, etc. It will be bad movie with this mistakes. I don't know how to not have these mistakes.

If you can halp me with this problems...

[gloomy verge]

What can I set as End frame? Some image, or set it empty?

[glad yacht]

Character at point a, Start 🖼️, character at point b, end 🖼️. Describe what happens between start and end 🖼️. The ai will fill in all the blanks

[gloomy verge]

Example, I place image of woman on the edge of pool as start and in pool as End and I write prompt: Female jumps to pool. This you mean?

[glad yacht]

That works

[gloomy verge]

And how can I create, if I want same person on the image on edge of pool and in pool too?

Because, I don´t want two different characters on both images.

[glad yacht]

Ask Gemini create 2 image

Crop each to uploaded as start and end frame

[gloomy verge]

Pool on first image and second image is same?

[glad yacht]

You have described on both frames with the same detail of the pool(if you want the consistency of the pool)

Example ^

[silver viper]

I have google ultra subscription. My flow doesnt have prompt to video option. It has only frames and ingredient. What should I do to fix it? Is there any setting that I need to enable? Would appreciate your reply and help

[storm kayak]

yo why i cant use flow? i got the google ai pro
It looks like you don't have access to Flow. Check availability here and contact support if you have any issues.

[silver wigeon]

Compartilhem meu filme

https://caisroom.com/movie/janaina-the-poor-sirena

[quiet kindle]

Friends, is Scene Builder currently unavailable?

[rapid shell]

Since Saturday night I haven't been able to generate videos through the flow, can someone help me.

[rain rapids]

does everyone have these issues with video generation?

it takes so long time

[unique abyss]

Just type in a prompt, and don't add ingredients, and you'll get a video.

This confused me at first too, because I was so used to the old way, but honestly, this is less cluttered makes it easy to change your mind if you later decide to add (or remove) images.

[crisp basalt]

show us the prom,pt u wrote

[bold falcon]

Just take your clips put them into DaVinci Resolve and go up into the Transform top right and zoom it to about 1.2 and it will disappear. It's your content. or cover it up with your own logo.

I was having issues for a little bit yesterday then doing the images and then making scenes with the video. I'm using Nano Banana Pro 2 with Ultra and I'm creating 2 iterations of images waiting until those images completely render then submitting the next task. I found by doing that I'm getting a lot less errors and I created about 45 last night without any errors once I let the current job finish. And if you think about that, if you let it finish you're gonna be saving time versus having to redo things because of failed generations. So give it a chance wait until each job finishes then submit the next 1.

[flint vine]

Does Flow usually not work at around this time?

[stiff rapids]

help me i have problem plz

[broken horizon]

i got the same issue i can also provide the prompt that got used and its definetly not agains any policies

[pulsar coral]

Then it's possible the prompts, the image itself a face or whatever element is being recognized by the models as something copyrighted, maybe if you upload it to Google AI studio and try it in original Nano Banana (Gemini 2.5 Flash image) you could change some elements so then Nano Banana 2 accepts it, you could also do it at Whisk but yeah... can't get too used to that anymore for obvious reasons.

Very popular recent stuff alike from movies or famous people can be rejected easily specially by Nano 2.

[broken horizon]

It's nothing like that I just wanted it do draw out the Lewis formula for some molecules

[pulsar coral]

Funny enough there's a famous "Lewis Hamilton" guy in formula 1 racing... I don't think it'd be beyond Nano 2 to be recognizing that as "Hamilton formula". Yes is the type of nonsensical recognition Google-search type Nano 2 is doing atm.

Maybe try removing the name Lewis from the images if they have it, I've seen those diagrams come with that

[rain rapids]

@VEO TEAM

Let us know what happend to Scene builder!! We cant use

@main saddle here are so much community messages, we event dont get an answer

[full flicker]

Hi there, this channel is for discussing flow with the community. The appropriate channel for any bugs or feedback is #1238621511401734254 . The staff monitors those threads and it won't get lost here in the chat.

[crisp basalt]

https://tenor.com/view/never-heard-of-her-mac-kahey-macdoesit-didnt-heard-about-her-before-i-dont-know-about-her-gif-20290616

[pulsar coral]

Flow is is just going through heavy usage atm, they're probably on it since it's monday already.... yeah Sunday's not very under watch.

Anyway, is this how people are doing the "Style transfer" from Whisk to flow? I don't see the results as accurate as a plaster prompt, maybe I'm missing something here

Whoever's so in love with them might wanna start saving them, the randomizer prompt from this must be headache

[rain rapids]

maybe they prepare release veo 4

[pulsar coral]

Two models in a year? Why would... no they would.

[pulsar coral]

Ah bingo I know what's missing here

These style prompts from Whisk are so incredibly intricated... no wonder people don't even wanna try, a normal human wouldn't be able to prompt it this way to Flow manually.

...never had much use for this so never noticed

Also "Style transfer" to Flow my carrots, this was simply done opening the Style Whisk image and copy pasting the prompt.

It's surely a really fast way to learn prompts tho

.... yyyyeah the Style randomizer dice icon from Whisk with whatever prompt is using most likely would be nice to bring to Flow if it could still do this.

Well if you didn't before now you know, got like 8 days to save them.

[sinful hatch]

Hi, i am the only one that when importing a vertical image to flow, it automatically flips it horizontally?

[full flicker]

Hello! Just a friendly reminder to keep all conversations in the supported language for this server. This helps our entire community engage effectively, and we've gone ahead and deleted the recent posts that didn't comply. Thanks for your cooperation!

[turbid lintel]

bytedance already release global seedance 2.0

when veo 4 😭

[craggy belfry]

There are commands? It would be great to know them. Is there a site with them somewhere? I have been searching for tutorials but there is not very much.

[amber silo]

Does anyone know why my interface always looks like this? When I return to the project and re-enter it, the images will be displayed here. This indicates that the images were actually successfully generated for the first time, but they cannot be displayed properly.

[arctic kelp]

2027

[blissful abyss]

I don't even know how to use whisk

[lucid bloom]

When copyright rulling become less strict

Seedance 2 is the prime example for the benefit of 'Scrap everything and Train on everything'

[unreal viper]

@bitter sage Hey! Just a friendly reminder to keep your personal information, such as medical data or contact details, private. To keep our community safe, we've gone ahead and deleted the recent posts that shared this type of information. Thanks for your cooperation!

[bitter sage]

why

[gloomy verge]

Hello. Can ask, what does it mean this? I write prompt only "car" and generate 4 images and I get this information. From some reason, befor some minutes ago I normally created images of peoples. Thank you for help. I tried change internet browser, but same.

[summer oyster]

Have you tried reading what the message says?

[gloomy verge]

Yes, but whenI am trying create only one image, it write it too. no lot of images. And before I normally generate 4 images, now no. It is strange

[summer oyster]

It simply means you are trying to generate too much in a short period of time.
The amount you generate is what counts.

[gloomy verge]

I tried create again after a 30 minutesd, but still I can´t create images. How many moments I must wait?

[summer oyster]

That i'm not too sure.

[cobalt canyon]

Why, if I have Ultra, does it put a watermark on my videos?

[crisp basalt]

So I wasn’t actually crazy when I find out that Gemini can do video edit for generated video , I wonder if Flow will get this capabilities too??

I do know the first and second bullet points are basically Video insertion/removal and Camera Reshoot in Flow but modifying the scene?? That’s like “remix” actually omg

[summer oyster]

so now you will spam channels because you need help?

[crisp basalt]

I mean I would too tbh but did you make sure you are logged in on right account

Also have u checked in “my library”

FLOW WE NEED THIS PLSSS

[summer oyster]

charge your phone

[young lantern]

Yes it's the right account

[young lantern]

Yes, When I open it, it’s now showing, “There doesn’t seem to be a project here…,” but this project is in my library and both the image and ID are there. What exactly is going on? I’m really about to lose it

[summer oyster]

i'm bald

[young lantern]

this project is in my library and both the image and ID are there. but when I open it , it’s now showing, “There doesn’t seem to be a project here…,”

[crisp basalt]

https://labs.google/fx/library?type=videofx

Try here?

Wait we are talking about Flow right ? @young lantern

[young lantern]

Yes

[summer oyster]

@young lantern You said you could not find support, yet you lied and you actually posted in the support section.

[crisp basalt]

Have u checked the link I sent you firstly?

[young lantern]

I've more than 10 projects in my library，earch project has >1000 imgs + videos

[crisp basalt]

@young lantern

Need to check that first

[young lantern]

Checked, I opened it, but it’s not grouped by project

[crisp basalt]

Yeah I Know

But I just want to make sure the images /vid are actually there

Meaning it’s not completely gone then. Uhhh now do you remember the timestamp of the project you created?

[young lantern]

I have tens of thousands of videos. At this refresh rate, it’ll probably take me until next week to scroll to it 😭 😭

[crisp basalt]

And how many projects you have right now so far in your account?

Project you’ve created btw

[young lantern]

Let me check

40 + projects

in my library

[crisp basalt]

If you created many projects, you might need to check each one individually to see if videos and images are included in any of them.

[young lantern]

The issue is that I clearly know what each Project contains, so I’m not failing to find the image because I forgot where it is

[crisp basalt]

what

its not about you know which project contains omg

🤦‍♂️

[young lantern]

The problem is that when I open one of the Projects, after a long wait it just says, “There doesn’t seem to be a project here...”

[crisp basalt]

ok there u go, see i just wanted that details

[crisp basalt]

ok can u provide screenshot of the project thumbnail? like this screenshot i sent,

[young lantern]

For example, when you open this project, it now says, “There doesn’t seem to be a project here...”

[crisp basalt]

yes i am aware of that @young lantern but right now i need to see if your project has any thunbnail images or not

[young lantern]

This one

Yes ofc

[crisp basalt]

ok uhh let me try something on my end

i wonder if its becuz of ,, non english text on your project that causes that

eh its not .

[crisp basalt]

is it that only one particular project that doesnt load? the others are fine right?

[young lantern]

Most of my projects have Chinese titles, but this issue never happened with the older projects

[young lantern]

Yes it's the only one

[crisp basalt]

yeah its not really the cause of the issue (the chinesetitles)

have u tried all of these steps like : open flow in incognito tab / clear cache/ log out and login , use other browser , use flow in mobile device

and also this. opening the project in new tab

[young lantern]

Thx, I'll try

I tried all of that, but nothing worked, and it’s the same on mobile （ios chrome）

1. The project exists and has a thumbnail image.
2. When I open it, it stays on “Loading” for a very long time.
3. After a long wait, it says, “There doesn’t seem to be a project here...”

[summer oyster]

Seems like a server side issue then. Nothing you can do, other than waiting for it to be resolved. Perhaps try again tomorrow.

[young lantern]

Yes, because most of the other projects usually finish loading within about 3 seconds

This issue started the day before yesterday, and I’ve already been waiting for two days before asking for help

[crisp basalt]

can u check the network log?

right click -> inspect ekement -> network -> refresh page

[young lantern]

I even tried it on multiple computers

[crisp basalt]

and see if theres any log that shows "red colored" meaning error

[young lantern]

Thx I'll try

[crisp basalt]

reason i asked this becuz its helpful for labs team/google staff here to try reproduce or identiy the cause

[young lantern]

Should I send it here or in DM? Might contain personal info

[crisp basalt]

I think u should try send it to one of the staff

@tough yacht is one of the people who worked on Flow btw

but id recommend u to post it in #1238621511401734254

easier for the staff to keep track of ongoing bug/issues, posting em other than that channel would be a problem cuz it will get buried by other messages

[young lantern]

I’ll forward our conversation there as well, since that might make the situation a bit clearer.

[summer oyster]

i told you from the beginning, but you ignored it.

[crisp basalt]

,

MOVING ON !

Scene builder is back now yay

after 84 years!!

[cursive plume]

Indeed, but it's a very basic version; it's impossible to create a prompt from scenes. If you look at YouTube tutorials and see what you can do, there's a big difference.

[frigid furnace]

.

[winged plinth]

..

[terse dawn]

https://youtu.be/ZHqhI15sllQ

[odd turtle]

You are from Türkiye

[terse dawn]

Yess

[native star]

can i prompt it to use one of my color palletes if i have them on a project?

[blissful abyss]

when did this feature come

[night ivy]

Salom

Video yasa

[sage otter]

how do we search our porjects in flow

[cyan dragon]

I saw on YT people speak about Daily Credits of 150 but i get only 50 ?
can someone confirm is it 150 or 50 ?

[blissful abyss]

there is no 150 daily reset

It is 50new users only get 150 and 50 reset every day

[weak wagon]

Hi evry one any help please. I have account gimini ultra but the watermark still whene i genrate a video ... Any feedback

[pulsar coral]

I just did a couple posts above your question. Basically you go to Whisk, press the dice icon generate a style, then click on the image and you'll see the prompt it uses, copy paste onto a notepad or whatever to save it forever. There's still like a week to save some.

Apparently there are like 10 different styles anyway, not much different one from other, so it can't be hard.

Let's see what this does

This is much easier than people make it seem to be, so no wonder

...Sadly it'll be gone tho, but now I know what it does will try later to explain later at feedback what people want with it

There's also the sticker one, that's easier

Plushie

Anyway these can be pasted to Flow to get similar stuff, idk if better

Let's try then, this is like 80's anime, auto generated by Whisk's random Style dice icon and it's original Nano Banana. Now I paste it to Flow's Nano 2.

This is Flow's Nano 2. same prompt.

It def has a different composition but yeah that's probably as good as a "Style transfer" gets

[crisp basalt]

https://tenor.com/view/eating-minding-my-business-no-gif-15059806

[pulsar coral]

Also it kind of complicates it a bit... pretty sure one can just prompt "80's anime style" and it would already do most of that without a complicated prompt.

Just gotta have some confidence when prompting, people are to used to that little button from Whisk. At least with that type of "Style" images it can't be so hard, you just gotta find and know the terms.

[pulsar coral]

Some countries are getting the watermark due to their laws and regulations about Ai, can't be helped. In your region at least Google's terms should have already removed the no-Veo watermark policy for Ultra plan from the buying page.

USA and seems Malaysia are still not having watermark for the Ultra plan. We still don't have a full list of countries I think.

[crisp basalt]

Don’t jinx it

[pulsar coral]

Those regions pretty much count as USA anyway, up above every Filipino speaks English too. War era colonization stuff business probably.

Also I just noticed I do my encasing at Flow like this

[Artstyle: 80's anime]

But Whisk's does it like this

Color Palette:

I mean like this

Maybe it does get things better like that at Flow too idk. But gotta be clear about the encasing for sure.

The model itself is talking like this so it can understand itself

[crisp basalt]

If you want the whisk auto prompting thing,

I have a solution

[pulsar coral]

Go ahead then

Tho I don't really know what are people doing with that completely random stuff, seems to be going a little beyond than "just for the fun of it"

You'd imagine when someone's sitting to generate something, they'd already have something in mind so randomness would beat the purpose of it

Maybe they're indeed just learning from it

Also in those 80's anime reporter images I put it can also bee seen how Flow's Nano Banana 2 prefers to give the characters huge heads and small bodies, while Whisk's original Nano Banana image kept the head to body ratio natural.

If that's how's it gonna stay can only imagine Fortnite style art benefitting from using Nano Banana 2

[pulsar coral]

Considering the prompt literally aims for "Vintage" Whisk's original Nano is doing it much better than Flow's Nano 2. Flow's looks like something cel-shaded yesterday, not cellulose acetate/nitrate film like Whisk's Nano does.

Literally unplayable

Well month's almost over, by then it should be enough to decide what to do.

[crisp basalt]

Dm me

[pulsar coral]

Awesomesauce

Your solution is great, tho I guess they did somehow autogenerate this at Gemini somehow.

It hurts to think about it tho

[crisp basalt]

everything within flow / whisk runs using gemini

[crisp basalt]

in term of processing user input

speaking of gemini

Flow team , when can we expect this in Flow 👉 👈

modifying the scene

[pulsar coral]

I don't think I'm gonna sit down to try to debunk that, even tho it can probably be auto generated as one of those how do they call them

[pulsar coral]

Gemini simply calls it "Create canvas" that could probably pull one out

[crisp basalt]

@pulsar coral

[pulsar coral]

... this thing thought about famous painters just like I would, didn't have to tell about it

This really isn't what Whisk is using tho, it'd need much more refinement

Could probably be deconstructed from a series of prompts from Whisk, then applied at Flow.

But yyyeah after those other tests, it's not worth the time without original Nano Banana.

But yeah more likely when Whisk was auto generated, they included it to have a randomizer dice button that can do this type of stuff, but they didn't bother with it at Flow. it's that simple.

[frosty niche]

Yeah I made 1 myself for a project last year for an AI room designer app, they probably voted it out, idk.

[pulsar coral]

Well if really caring about it, this'd be a solution to give to people.

[frosty niche]

The specific dice for a different prompt is chill i was cooking on this project.

Been awhile since i thought about this one. lol

#MadeBeforeStitchlol

[opaque wagon]

Create Collection doesn't work now

[wet flame]

Batch download also doesn't work

[wet flame]

And now I am generating too quickly

[crisp basalt]

Yes

[opaque wagon]

Only Ultra, poor pro.

[wet flame]

NB2 now works but pro doesn't

[fresh scroll]

I am using the scene build but I cant see the JUMP TO feature?

[crisp basalt]

its no longer part of the scene builder

[fresh scroll]

Well anyone there that can help me generate a face of a person consistently? Despite of changing her outfit?

[crisp basalt]

use ingredients to video

[hot crown]

yes I'm having problems with the download feature too, download project gives me only a handful, and select-all/download doesn't work at all.

[proud vapor]

happening to me too

[hollow jungle]

hey guys can anyone guide me that if im using google flow for video generation, i am only able to give 3 ingredients and none after that, should i use the frame method after i generated the ingredient video? how do i stitch them? i was using elevenlabs flows but it has a lot of limitations too. i want to create an ad/tutorial video for my app

[neon raven]

am i the only one getting constant errors

[blissful abyss]

I get it too but I guess the team said they're working on it

[neon raven]

do u get it constantly or not

[frosty niche]

I was getting them alot to earlier.

[hot crown]

all the failures, none of the downloads, it's going to be that sort of day 🙂

[real mirage]

pain

[green needle]

Any news about being possible to transfer projesct from whisk to flow?

[loud gull]

I have been getting this error as well. I dunno if it's a glitch or something else

[astral fern]

it doesn't even attempt, it just fails 🙁

[loud gull]

Yeah I'm really confused about this right now

[wide ether]

I have a very large project for which I'm wanting to download all my favorites. When using the shift key to multi-select the videos, not all of the selected videos make it into the downloaded zip file. Do I seriously have to manually download each and every one of my favorites??? I have like a thousand!!!

this function used to work, it looks broken now

wondering if it has to do with file naming ... if you've used the same prompt over and over, would the file names be the same on the download? Meaning you'd only get one file per unique file name?

[limpid ore]

VEO4 IS COMING

[wide ether]

Yeah I think it's the file naming ... I'll report that bug

where do we report bugs?

[lyric dust]

The generated resolution has decreased

[astral fern]

Where’s your source love

[limpid ore]

I'm excited about the new update that Google is going to release; who knows, maybe Veo4 will be announced on May 19-20 at Google I/O.

[wide ether]

I created a bug report for the multi-select download issue ... can we get that fixed ASAP! It should be a pretty easy fix.

[cedar shell]

Because Flow doesn't let me animate images with people, or create images of people talking on the phone in a circle.

Do you know of any other Google tools that can help me create explainer videos? Like motion graphics.

[raven siren]

Yes, hardly any images finish...

[frozen totem]

flow get something went wrong on 90% generate image and video?

[austere zodiac]

Is there any official release date ?

[elder flint]

I have credit pro but i want now download 2k images and didn’t work why

[wet flame]

Flow rn be like:

https://giphy.com/gifs/trump-consequences-NTur7XlVDUdqM

[wide ether]

Are you trying to download them using the multi-select and download all as a zip? If so, that's broken ... see my messages on that earlier today

[wide ether]

@elder flint ⬆️

[wet flame]

I felt like there should be an option to disable the UI animation

so if it failed, I don't have to wait for it to slide to delete it

[mild swallow]

why???

[wet flame]

Which model are you using?

For me NB Pro doesn't work

[mild swallow]

I'm losing it.

[wet flame]

Yeah, like I said

I had the same issue this afternoon

and rn

[mild swallow]

How can I handle this? It’s starting to affect my work performance.

😭

[wet flame]

Return to use Whisk xD

only if the result doesn't need to be ultra realistic

[main saddle]

hey all! we're investigating this, thanks for continuing to report!

[wet flame]

Thank you

[mild swallow]

Thank you

[brittle hearth]

I can't generate any images please fix this

[arctic kelp]

hey all! we're investigating this, thanks for continuing to report!

[dull rapids]

Looking forward to Veo 4

[bright heron]

Very much so

Need something to compete with seedance and kling

[bronze owl]

Hi everyone, I’ve been using Google Flow (Veo) for generating short scenes and noticed a couple of issues: 1. An 8-second scene is consuming around 20 credits, which feels quite high. It would be much better if credits were charged proportionally based on duration (e.g., per minute). I think this would make the tool easier to use and encourage more users. 2. In 8-second scenes, the character’s position keeps changing repeatedly. I tried using identity lock, but the issue still persists and the character does not stay stable. Is anyone else facing the same problems? Any solutions or tips would be really helpful. Thanks!

[blissful abyss]

U can kindly report this in #1238621511401734254 following the message guideline which is pinned in the channel

[muted scaffold]

Hi everyone—could someone tell me why this issue is occurring, or who I can reach out to here (or where I can submit a support ticket)?

Error:

You are sending generation requests too quickly. Please try again in a few seconds.

[rain rapids]

@main saddle

[rain rapids]

sora died, will veo 3 also be cancelled?

i hope not 😭

[bright heron]

I don't think so. OpenAI has been slacking behind on image and video generation for awhile and couldn't compete with others. I'd hope google has the funding and motive to continue veo 4 development....

[wintry solar]

is flow free

[pulsar coral]

Speaking about Open AI, ChatGPT just added something called "Styles" for the image generation

Ain't that funny, wonder what else has/had that feature and people keep requesting it

Flow really needs to get their disposables together before others do.

[placid sky]

Hey friends iam new here, does anyone know how to get the SEED ID from a generated Video ? Cant find it anywhere in Google Flow ? Thx a lot

[wet flame]

For the faces, I usually put stable anatomy, consistent proportions, consistent identity at the end of the video prompt to keep the faces correct from the starting image, I'd say it work quite nicely in Veo

[glossy island]

delete my library data not working

[wet flame]

It is free to use, just around 100 image limitation daily, and 3 8 second video generation

[placid sky]

is someone active here who knows how to pull the json metadata from a generated picture or video ? I need the seed id

Chat is dead ?

[wet flame]

About the cost, tbh it's quite cheap compared to other video gen, also depends on what type of content you're generating.
I've tried Kling 2.6 and Veo, both are good for realistic video, Veo's 8 seconds video with audio cost about 0.25 USD, but Kling does 10 seconds video with sound costs about 50 cents

[wet flame]

no idea

and timezone exist, it's 10pm in NY, 2~3am in UK and Europe

[wintry solar]

ye i tried costed me 100 credit to make one video. was wondering if any more options available

[wet flame]

Are you using Veo 3.1 Quality?

[wintry solar]

ye lol was not supposed to?

[bronze owl]

Hi everyone, I’ve been using Google Flow (Veo) for generating short scenes and noticed a couple of issues: 1. An 8-second scene is consuming around 20 credits, which feels quite high**. It would be much better if credits were charged proportionally based on duration (e.g., per minute). I think this would make the tool easier to use and encourage more users**. 2. In 8-second scenes, the character’s position keeps changing repeatedly. I tried using identity lock, but the issue still persists and the character does not stay stable. Is anyone else facing the same problems? Any solutions or tips would be really helpful. Thanks!

[wet flame]

Well fast costs 20

[wintry solar]

oh wow

[wet flame]

bro just copy and pasted his message ._.

[wintry solar]

i provided a image to make a video out of it somehow it didn't used the image atol

[bronze owl]

Here to explore AI tools and figure stuff out if anyone knows a fix or workaround for this (doesn’t matter where copy nd paste it’s from), please let me know 🙏

[wet flame]

@bronze owl read here

[wet flame]

Did you use the image as starting frame or ingrediants?

[bronze owl]

yes

[wintry solar]

seem like ye

[bronze owl]

ys i use the image as starting frame or ingrediants and lock their pose but still issue

[wet flame]

seems like starting frame or ingredients?

[bronze owl]

In every video scene, the character’s pose keeps changing. It doesn’t stay consistent across frames.

[wet flame]

You want your character to stay like a statue?

[bronze owl]

Yes yes

[wintry solar]

idk just drop the image and gave a prompt, ig by default it did something itself

[bronze owl]

If 8 sec costs 20 credits, then 1 min costing ~150 credits is fine ...character should stay consistent and not change pose every frame,its easy to use

[wet flame]

What's the prompt?

[bronze owl]

I’ve already tried giving detailed prompts and using pose/identity lock, but after one 8-second clip, the next clip still changes the character’s position.

[bronze owl]

I'll send

[wintry solar]

ye seem like it will take time to figure it out

[wet flame]

It should look like this, and apparently Quality doesn't support model references

[bronze owl]

LOCK Video Creation Instructions:
keep lock the same posittion or background using the provided reference image
lock a hyper-realistic cinematic video using the provided reference image
Do NOT add or remove anything from the woman
Keep the same woman exactly as shown in the reference image
No visual enhancements, no stylization, no filters
Facial Movement & Performance

The woman speaks naturally, with:
Smooth, continuous lip movement
Realistic facial expressions
Natural blinking
Speech flows smoothly as if recorded in one single take
No awkward pauses, no sudden cuts
Eye Direction (IMPORTANT)
She must look directly into the camera lens
Do NOT look side-to-side or away
Maintain steady eye contact throughout
Identity & Pose Lock (CRITICAL)
The woman must remain EXACTLY the same as in the reference image

Do NOT change:
Face
Facial structure
Age
Skin texture or details
Body proportions
Keep the same pose, posture, stance, and orientation
Preserve her original expression style and overall presence
This must look like the same real person, not a modified version
Body Language & Gestures

Subtle, natural movements only:
Slight head movement
Gentle hand gestures while speaking
Relaxed posture
No exaggerated actions
No sudden or unnatural motion
Visual &Technical Quality
Lighting is soft, natural, and realistic
Skin texture, fine lines, and facial details appear ultra-realistic
Background is clean and silent
No overlays, no graphic elements, no visual effects
Final Quality Requirement
The video must look like real camera footage
It should NOT appear AI-generated
Cinematic realism is the top priority

[wintry solar]

@wet flame what more video generation you use

[bronze owl]

i have use

[wintry solar]

i used frames lol

[wet flame]

Veo 3.1 Fast and Kling 2.6 nowadays

[wet flame]

well it definitely will be AI-generated, so you just want her lips to move or..?

You kinda over prompted the prompt

[bronze owl]

I want,,In the image, the woman’s position should remain the same, but her head and hands should move naturally like a real human

[wet flame]

You mind sharing what you want the woman to say?

I still have some credits, I can try to do your prompt

[bronze owl]

I’m new in this field, so I’m facing quite a lot of issues. I think it would be good to discuss with people who have more experience maybe they can find a solution. I’ve tried different things, but the issue is still the same, so I thought of discussing it on Discord

[wintry solar]

oh ye i wanna see this too, provide the image @bronze owl

[bronze owl]

Sure sure; I’ll come back. I’ll share it with you after some time,

[wet flame]

Yeah, usually I will start with a simple prompt, then correct to what I want. It's AI afterall so it takes multiple tries to make it right... unless you hit jackpot

[wintry solar]

@wet flame can i give you image

[wintry solar]

ok we have a image

[bronze owl]

If more than one is made, then it becomes different.

[bronze owl]

Here you can see the difference

[wet flame]

Ok, what's the dialogue?

[wintry solar]

make her speak how yll doing

[wintry solar]

Kling 2.6 is asking $

[wet flame]

of course

[wintry solar]

which one is free

[wet flame]

none of them

Video generation takes resources, yk

[wintry solar]

so just flow is with me

[wet flame]

@bronze owl I tried it in my way, see if it met your conditions

[wintry solar]

crazy which one is this

[wet flame]

Veo 3.1 Fast

[wintry solar]

cool

[wet flame]

The prompt is
A confident young woman with deep brown skin and natural hair in a high puff, wearing a colorful patterned wrap and hoop earrings, stands in a sunlit field of tall grass, looking directly into the camera with calm presence as she speaks naturally in a soft, grounded tone, subtle facial expressions and eye contact, gentle breeze moving the grass and fabric, shallow depth of field, warm golden-hour lighting, realistic skin texture, handheld camera feel with slight natural motion, human-like micro-expressions and lip sync, cinematic realism. Her dialogue is: “Not everything needs to be loud to be powerful.”

[wintry solar]

this was frame or

[wet flame]

frame

[wintry solar]

you created another pose to have frame ye?

[wet flame]

No, you can just use the given image

[wintry solar]

can you show the settings you applied

[wintry solar]

Sora got closed already?

[wet flame]

No idea, I've never used it

[astral fern]

Maybe it’s cause I’m tired, but I don’t see a suggestions channel.

Flow really needs an app, when I use it in my iPad, my iPad overheats on the site

[pulsar coral]

Use the labs-bugs-feedback area, and check the pinned guidelines there.

[north wigeon]

was the extender prompt feature removed?

[glacial slate]

what do you guys all use flow for primarily? youtube?

[wet flame]

Oh, so the dialogue isn't in English?

[bronze owl]

Look at this, this is how it moves when changing position

[wet flame]

I think it's fine...?

[bronze owl]

No bro ,2 clips but position changed

[bronze owl]

i want the position in every scene to remain the same as the first clip. That’s the only issue. If you can provide a solution for this, we can discuss it . Otherwise, everything else is working fine.

[wet flame]

I think you're using ingredients instead

[bronze owl]

Yes

[wet flame]

If you're using the same image as (starting) frame, it shouldn't do that

[bronze owl]

Then I should use a frame; this issue won’t happen with that?

[bronze owl]

Okay, I’ll check and let you know.

[wet flame]

yes, the frame also means "starting frame", an AI video is basically generated frame by frame into a video. So if you start with that specific frame, it will always start with that frame.

[bronze owl]

So when the next clip is generated, the position will also remain the same ok ok got it

[wet flame]

Yes, but this is something I don't suggest you to do

since you want to continue the flow, you can use the Extend feature instead, it basically take the last frame to generate your next video/dialogue

[bronze owl]

Okay got it , Thanks a lot

[pulsar coral]

Try using

[Camera/perspective/composition: Always locked to reference image without any creativity freedom, no angle panning, no transitions.]

Even if you're not using reference image, that will lock it to the first image from the video

[wet flame]

<@&1366831844502274180>

It's hammer time

[bronze owl]

Yes It's working Thanks @wet flame @pulsar coral

[pulsar coral]

There's a lot of people who need that same composition, probably cause it works nicely for reels

[wintry solar]

using ingredients doesn't made seem flawless atol

[crisp basalt]

Skill issues

https://tenor.com/view/cat-macska-cica-vomit-oklendez-gif-14738230240638170539

[wintry solar]

what generation tool you use apart flow

[warm kite]

make the man in the image look straight with a smile

[crisp basalt]

look what now

[remote vortex]

veo 3 ultra is not generating videos at all! Is this happening to everyone? 😕😕😕😕😕

[crisp basalt]

Refresh the page and try generate again

[lime quiver]

Welcome to Google—nothing works properly here

[dark bolt]

clear your cookies and change your ip

[crisp basalt]

Nvm it’s the em dashes for me

🤖

[lyric dust]

Flow's image generation has started failing in large batches again...

[trim lintel]

It's failing a lot here.

[placid wing]

same here

[brittle pelican]

Imagine you pay more than $100 for a service, and it doesn’t work properly for 2 days in a row. Can you imagine that?

It's good that Google cares about its users

[brittle hearth]

Fr I have to do the work and i can't

[gloomy verge]

Hello. Can anyone help me with something? I want create video about two peoples, example reporter and female. But I want use in one 8 second video different angles on both charaters, white they are interviewing. How can I create it? It is possible create it? Thank you for help.

[crisp basalt]

like this?

"A reporter interviews a woman on a city sidewalk in daylight as pedestrians blur in the background. Start on a medium two-shot as the reporter holds a microphone between them, then cut to an over-the-shoulder close-up on the woman answering. Cut to a reverse over-the-shoulder on the reporter nodding and asking the next question, with clean natural lighting and realistic handheld camera feel."

[gloomy verge]

Yes. But, can I create more than 3 camera angles in 8 second video?

[crisp basalt]

second?

yes you can, that one i shhared was 8 sec

[gloomy verge]

And can you write me, how I create this angles?

[crisp basalt]

by asking google gemini

[gloomy verge]

Maximum seconds of video is only 8?

[crisp basalt]

yes

[gloomy verge]

But I seen here videos from other creators and it have example 1:10 etc. Longer.

[crisp basalt]

cuz they stitch the video together using scenebuilder/video editor

maximum is 8 second. no more than 8 when u generating a video using veo3'

[gloomy verge]

In other than VEO3 is more?

[crisp basalt]

@main saddle @tough yacht i am so sorry was just trying out if it maintain consistency 😂

[main saddle]

my true form

[crisp basalt]

well its up to u which video generation service u want to use,

[gloomy verge]

Is strange that on the begining of video hold microphone otter and after it looks as man with kid. No?

[crisp basalt]

im so sorry i dont understand what are u trying to say here-]

[crisp basalt]

https://tenor.com/view/well-yes-tokyo-toni-zeus-tokyo-toni-red-carpet-interview-gif-9160985937715866183

[gloomy verge]

Here hold microphone otter. And second image it looks as man hold microphone. Microphone is in different angle.

[crisp basalt]

yes

[gloomy verge]

It is some mistake of video generator?

[bright heron]

I'm getting a 100% fail on all generations right now. Anyone else?

[crisp basalt]

idk how to help u any further cuz you are asking all over the place

[crisp basalt]

yes

nothing is perfect my friend 🌠

[dry frost]

I'd give you the best hug, you deserve more though

[crisp basalt]

you just have to refine your prompt , til you get the result u want

[night tide]

Same here. Cleared cookies/cache. Tried different browser and incognito, used VPN and still the same issue.

[bright heron]

Are they just failing or telling you too many generations and to wait? I am getting the ladder even though I've been really conservative with how fast I'm generating

its been 30 mins and still telling me to wait a moment

[gloomy verge]

I tried create video about reporter and female with only 2 sentences: one from reporter and one from female. And I get this. Why?

[night tide]

Im getting the error "Failed. Oops, something went wrong!"

For almost every generation

[gloomy verge]

Where can be problem?

[halcyon drum]

Guys need help in veo. How to keep the dialogues locked in. The dialogues are coming out by the wrong person mouth.
Here is my prompt:
Strawberry Nurse says warmly: "Good morning! How are we feeling today?"
Cucumber Patient grins and says, voice smooth: "With a nurse as pretty as you around, I'll be healed by tomorrow."

but the dialogues are coming out mixed up.

[dapper narwhal]

Am I just completely devoid of any clue here? Is this a complex request for Flow? Whisk handled stuff like this, no problem.

[bright heron]

I cleared my cache and cookies and it seems to be working now

[crisp basalt]

its not complex, its flow website having high demand issues

[pulsar coral]

It's a little harder with reference images yes. But that "cut" thing you do seems to be working

But more often seems to work much easier and perfectly without a reference image. I think the VEO model has a little problem to recognize images when there's two subjects

[bright heron]

I wonder if traffic is high because higgsfield is trying to break a guinness world record for most video generations at once, at 11am pst

[pulsar coral]

Like this one's without ref and works okeish

[crisp basalt]

try this A strawberry nurse stands beside a hospital bed and speaks first, her mouth clearly synced as she says warmly: "Good morning! How are we feeling today?" The cucumber patient then grins and answers second, his mouth clearly synced with a smooth voice: "With a nurse as pretty as you around, I'll be healed by tomorrow." Medium two-shot in a bright, clean room with soft morning light, slow push-in as the second line begins.

[lilac hearth]

I'm having a question hoping that someone can help. I have created collections for pictures and videos in flow. But when I download the collection I only get a small number of these items. Just tried it again and got a collection with 64 videos and got a zip file with only six of them after downloading. Is there anything I do wrong? Any help would be highly appreciated

[crisp basalt]

did u use this

[halcyon drum]

It worked. what is the concept here? speaks first? So I understand and write my prompts accordingly

[edgy holly]

No. OpenAI needs to focus to compete. Google has all the money it needs.

[crisp basalt]

ok honestly i- didnt expect that it works well omg

i guess u can try follow the prompt structured that i wrote see if it works in other kinds of story

[edgy holly]

No. AI video is expensive. You do get some free gens, though.

[crisp basalt]

and it refreshes daily too, very generous

[lilac hearth]

no I have clicked the three dots on the collection and choose to download there.

[edgy holly]

Yeah, they ran this server into the ground. Used to be great.

[edgy holly]

Not consistently, but if you burn a bunch of generations you can get that.

[pulsar coral]

Yeah that fruit video problem sounds like the problem about the Kiwi and Banana I couldn't find a total solution for the other day... and did try something like that over and over. Maybe they updated the model or something.. Its mostly about the model being able to recognize a "Cucumber" and a "Strawberry" in a reference image, maybe it can do it now

[edgy holly]

Fails come in waves, during peak hours.

[edgy holly]

Underprovisioning.

[lilac hearth]

I use this button with the three dots.

[crisp basalt]

try use the one on top right

@lilac hearth besides the "gear " icon

[lilac hearth]

I have used this button now and there it is even worse getting only three items out of 64😂 I am using Edge maybe I will try it in Chrome as well

OK it is the same here. Of course I can download every single Clip on my own but the download collection button would really Help.

[gloomy verge]

I can´t now create videos?

[edgy holly]

I don't understand. How would I know if you can or can't right now?

[gloomy verge]

I tried it 4-5x, and still I can´t create videos. If it is problem only now with something, or longer problem.

[edgy holly]

"Underprovisioning."

AKA: They don't have enough compute during peak hours, and fails spike.

[gloomy verge]

I am not english speaking, so I don´t understand. From this reason I am asking. Thank you for help.

[edgy holly]

I didn't even know that word myself until last month, when a Google employee told me.

During the week, there are often a few hours where they don't have enough servers to run all the generations, so they fail out. Wait an hour or two, and it'll usually work. If this happens a lot, take notes on when you see fails happen, and try to schedule your generating work around those peak use hours.

[gloomy verge]

Okay, thank you, I try wait some hours.

[worldly coral]

Help! I didn't finish creating the video, but I've already had my credits deducted.

[frosty niche]

Anyone elses flow libraries are not showing up for them?

[austere heart]

I had this happen a week or 2 ago, nothing fixed it, but they did eventually come back after awhile.

[frosty niche]

Interesting.

[frosty niche]

I fixed it, nevermind. Just swap browsers if it happens to you again bud.

[hidden compass]

hello i want to know the credits of free video maker on flow restarts everyday?

[long hawk]

Hello!

[hidden compass]

can you help me?

[autumn sluice]

I'm new here. Does the #flow-videos・📹 channel allow still pictures made on that platform as well?

[main saddle]

hi, welcome! the #flow-videos・📹 channel is meant to be a space for folks to share stuff they're making with Flow!

[main saddle]

yes we actually just shared another #FindyourFlow on X & Instagram which covers this idea -- you can upload a color palette as an ingredient and prompt from there! https://x.com/FlowbyGoogle/status/2037271961651282233?s=20

Flow by Google (@FlowbyGoogle)

💡 Pro Tip: Try uploading a color palette into Flow as an Ingredient, then add the prompt: “Based on the provided color palette…”

The result? Visual consistency for all your images and videos locked into your signature aesthetic. #FindYourFlow

X · Mar 26, 2026, 8:54 PM

[elder flint]

I have pro account but when i do download 2k didn’t work why

[rancid silo]

I keep getting “you’re requesting generation too quickly ” for NB pro even after waiting like 30 minutes I’m still getting this same message.

NB 2 works fine though

[main saddle]

I wasn't able to reproduce this, are you still having this with refresh/browser swap?

[crisp basalt]

I’m gonna assume you were trying to upscale your uploaded image ? it won’t work that way

Nvm I’m dumb

[mighty bison]

It’s gonna take a while to get used to Flow

[opaque wagon]

BN Pro has high quality but low stability.
BN 2 has high stability but low quality.
Mixing them is the best way.

I resized the UI; the default is too big.

[mighty bison]

Why do I keep getting this? I’ve waited for a good 20 minutes

[wet flame]

well well well

[opaque wagon]

Dragging the PNG into Illustrator failed, jpg is fine.

[pulsar coral]

Bottom line, everyone is pretty much saying the same thing yeah. Literally only one person in the entire Discord server has ever said to actually like NB²—and even then, he mainly uses it to generate Fortnite character images with oversized heads or zoomed-in headshots with small bodies. That’s perfectly fine, but it doesn’t work well for most other types of art people create at Whisk.

Most people also say they prefer Whisk’s image quality over Flow’s NB² and Pro. However, there’s a lot of confusion around this. People think Flow lacks Whisk’s so-called “precise mode,” which is what makes the difference for many users. In reality, that mode from Whisk is just the original Nano Banana (Gemini 2.5 Flash Image) model, which Flow also had until about three weeks ago, when it was suddenly disabled without explanation or logic and visibly against overwhelming ongoing user feedback.

There have been daily complaints about this, yet some people still believe NB² is superior. In some aspects, that’s true—but not when it comes to accurate, adherent to reference image, yet creatively flexible art composition, which NB¹ still handles much better.

I asked whether they would ever enable it again, and only one developer responded. His answer was: “I don’t know"

[pulsar coral]

Flow just isn’t displaying the correct messages in some cases. It’s also possible that you’ve run out of image generations for the day, and it’s just not telling you with the correct message.

Your account should be fine again in 12-24 hours from the last time you successfully generated an image... but not after 5 mins of it yeah.

[opaque wagon]

Gemini uses NB² by default now.

[pulsar coral]

Yeah and I don't think free accounts get NB¹ anymore after they run out of NB² image generation tokens there. I don't think I've lurked enough about it, but probably going through the hassle of paying for an API key at Google ai studio is one of the only ways to truly work in peace atm without people disabling your essential tools

Flow at least won't put the spark watermark on your images like they do there, but that also can be handled easily with a gemini watermark remover as long as you did use the download button from the image and didn't just right click save as.

[crisp basalt]

nano banana 1? it shows there without paid tag

[pulsar coral]

Yeah but it only gives few free tries daily.

[crisp basalt]

i mean yeah, im not surpirsed

[pulsar coral]

Probably I'd need some good 40 cents per session to be done in one day

[crisp basalt]

I’m still wondering why you wanted Nano Banana 1 in Flow, but thats okay. I just dont want you to start sending a whole wall of text

[pulsar coral]

Me? Don't you mean literally everyone who prefers Whisk's images?

I'm not going against windmills here

[crisp basalt]

wel its only you so far lol

everywhere i go , its only you that rants about it

and thats fine,

i guess

[pulsar coral]

That's part of the confusion yeah, and the rants are more often about me explaining what Whisk's "precise mode" is.

After people request it

[crisp basalt]

yeah and your explaination are

dead wrong lol

but thats ok ,

[pulsar coral]

You've never pointed at it even once

[crisp basalt]

i do , and u chose to ignore it lol

[pulsar coral]

And you kinda like to do that

Well why is it wrong then, I'll choose to listen

[crisp basalt]

nah lol,

[proven portal]

5 on it

[pulsar coral]

You forgot the gif.

Also seriously if it would be wrong, at least one other person would have said it, more often they do agree and have left the same kind of feedback.

[abstract needle]

Hi, I need a little help. Why is the logo/image I'm uploading to Flow being altered as if the AI ​​wants to recreate it?

[pulsar coral]

Also there's one more other thing I never say tho, did you actually have time to try NB¹ while it was enabled at Flow?

It was "Doing things" man.

[crisp basalt]

yeah , it was there until nano banana 2 was announed

nano banana 1 was always there , but people didnt know , nor even care to use it in flow specifically,

[pulsar coral]

I know I'm all over it, but since you asked, most of my prompts at Whisk have to keep some "Fix" prompt about hand anatomy and morphollogy, and only like 2 outta 4 it worked and looked fine with NB¹ .

While I had time to test at Flow tho, it worked on EVERY image perfectly

It was the best NB¹ generation from anywhere for some reason, as if it had some sort of "thinking" or "pro" mode enabled or something, kinda like what happens at Gemini with the Pro mode

And I don't think people wouldn't have cared to try NB¹ at Flow, they just wouldn't wanna use Flow in general at the time, I did also notice so very late

[crisp basalt]

nano banana 1 doesnt have thinking built in tho

[pulsar coral]

Well that's another thing you wouldn't have any place to notice with... cept while it worked at Gemini maybe

[crisp basalt]

its not a maybe lol, gemini did most of the carrying

[pulsar coral]

Eitherway, I do generate dozens of those images daily, and did notice it did finally started working properly everytime

You'd only have noticed using the same template prompt over and over. Seriously people were too entertained at Whisk to notice what was going on here.

[crisp basalt]

also "precise mode" is already on Flow.

oh wait sorry *refine mode

[pulsar coral]

Yeah or "Smear" tool... this is where the confusion starts

[crisp basalt]

i think youre confusing yourself

[pulsar coral]

Why is every Malaysian I've met so vindictive

[crisp basalt]

?

[pulsar coral]

I swear some of my worst enemies are from there

[crisp basalt]

thats a weird thing to say man lol ,

[pulsar coral]

ikr

[crisp basalt]

weirdo

anyway

[pulsar coral]

Anyway, It's not like Whisk clearly says "Nano Banana" mode when you click the Precision mode tool, and whenever people have come here asking for it or saying NB2 is not as good as what Whisk does, I've indeed had to explain what's going on

Some of us do need the tool and gotta feel empathy about it

[crisp basalt]

well go check network log lol

[pulsar coral]

So if you're gonna say "It has precision mode now" yeah we might as well just start writing a rosetta stone about this

[crisp basalt]

this is when i enabled the "precise reference" u can see it uses GEM_PIX aka nano banana 1

[crisp basalt]

thats on u lol , u thinking too much about it

[pulsar coral]

Like said it's about empathy. Some people actually use this for work, and things are not working so well atm. Not just on my side.

[crisp basalt]

the refine mode , with precise ref. on , uses gem_pix aka nb1

so if u really want to use nb1 , just use whisk , for now

[pulsar coral]

I don't need json, math or alchemy to see that screencap of Google AI studio with NB¹, go there, run my template prompts and see it creates the same image as Whisk or better

I'm not even the only one who's done it by now

[crisp basalt]

see lol , youre doing it again

🤷‍♂️

[pulsar coral]

But it's the truth tho, I've also posted images here side by side proving it.

[crisp basalt]

cool.

lol , anyway moving on

[pulsar coral]

In fact if you don't believe me, just try it yourself.

[crisp basalt]

sure, ill try if u stop being so dismissive towards people

[pulsar coral]

Well what is it gonna be then.

[crisp basalt]

google staff quietly watching this argument prolly laughing their butt right now

[pulsar coral]

That'd be something if they'd at least watch

[crisp basalt]

yeah go ahead and be a weirdo again by pinging the whole staff and mod team

[pulsar coral]

Bold of you to assume

So, let's go on with it

[crisp basalt]

https://tenor.com/view/donald-duck-sleep-gif-12073235593935901891

[pulsar coral]

That's one funny thing to ask

[crisp basalt]

ok give me the step by step

[pulsar coral]

Ok then, here's reporter girl

Prompt

[crisp basalt]

https://tenor.com/view/donald-duck-sleep-gif-12073235593935901891

ok im waiting sorry

[crisp basalt]

https://tenor.com/view/mts-meg-squinting-meg-camera-megan-thee-stallion-megan-near-camera-gif-24541289

sir or ma'am what am i looking at

[pulsar coral]

It's a random style prompt from Whisk's Nano 1.

[crisp basalt]

.

[pulsar coral]

Can't be posted at Discord. But yeah this is my example here.

[crisp basalt]

whisk nano 1 doesnt generate "text caption:"

[pulsar coral]

So let's see what it's doing

[crisp basalt]

nano banana model does not output a text prompt, only images

that caption was generated by imagen 4

[pulsar coral]

Notice all the results samey

It's REALLY specific, unlike something I'd prompt, that's why I didn't write a prompt

[crisp basalt]

???

[pulsar coral]

Only the same model would get this same type of image

[crisp basalt]

bro what

[pulsar coral]

Now let's run it at Google AI studio's Nano

[crisp basalt]

first of all, pls screenshot the whole thing and not half like this

[pulsar coral]

Ok then

[crisp basalt]

i want to see these too

and show this to see if u enable or not

[pulsar coral]

idk man, all I want is the same everyone wants, to get these images back

Anyway that specific prompt would do that, I'll put in a notepad later

Going to Google AI studio now

[sonic hawk]

Guys, how do I get this? I don’t have this in my flow. Please help.

[pulsar coral]

How many of these do you need to be convinced it's the same exact thing, 20 or 30

Or do i need to generate some at Vertex too

Not even using "Seed" here, it just does the same image by itself with the same prompt at a completely different site

Seriously, he is probably the smartest guy here and he didn't see this either, no wonder almost no one else does either

Here's the Whisk style prompt, go nuts

[crisp basalt]

lol

[sonic hawk]

help

[crisp basalt]

im still lost with what you are trying to explain here , @pulsar coral

[pulsar coral]

I said Whisk is using the original Nano Banana for it's precise mode, you said I choose to ignore your code explanation on why that's not true

[crisp basalt]

i dont wanna ask any further question cuz you will end up being dismissive anyway so

[pulsar coral]

idk I might still be wrong in some way or the other, but the eyes don't lie here

[crisp basalt]

thats on you, good luck

[pulsar coral]

Still waiting on him to explain what was I wrong about tho

Anyway those AI studio images have the Gemini watermark

[crisp basalt]

i already did few days ago,

[pulsar coral]

The same experiment can be reproduced with any other Whisk styles prompt, it's very effective to prove it's the Nano Banana model itself.

Doesn't need to be the anime girl one, can be done... until Whisk passes away

[frosty niche]

What is going on here lol

[frosty niche]

Whisk, no image res/ingredients.

Flow no image res/ingredients.

[pulsar coral]

Apparently we both don't know what was going on either. But that's on me.

lol

Now run that prompt at Google Ai studio's Nano 1, get in

[crisp basalt]

basically ai hallucination is catching up to him but thats fine

[pulsar coral]

And I thought that was turdman, wriggly turdman

Anyway idc, just want to be stated that Whisk is using NB1 for precise mode, I can be a cornchip about the rest for what I care

And that everyone asking to have Whisk quality images at Flow needs the original Nano Banana model turned back on, I'm only one of them.. sadly the one that noticed

[crisp basalt]

looks good 👍

[pulsar coral]

In these you posted Flow's Nano isn't aiming for "Vintage" that much and the prompt literally has that word, Whisk's original Nano is doing it much better than Flow's Nano 2. Flow's looks like something cel-shaded yesterday, not classice cellulose acetate/nitrate film like Whisk's Nano does.

[pulsar coral]

Yeah but these are missing that too, look closer

[crisp basalt]

cool

[pulsar coral]

The cel-shading ruined the acetate film quality that Whisk's Nano did get correctly

[crisp basalt]

thanks for sharing with us

[pulsar coral]

.. that's also why I picked this, because you'd come out with this. It's the perfect example.

Ever seen a classic somewhat blurry 1940's Tom & Jerry cartoon? That's acetate nitrate film. Or really old anime

Not even that old, Evangelion also looks like this, Flow's Nano 2 totally missed the point

[frosty niche]

Lol i ran that as a prompt in GAIS & it made an anime image generator. 😂

[pulsar coral]

then again being fair Nano 1 does also miss the point sometimes too

Let's see.. what looks like this, sort of

https://www.youtube.com/watch?v=JyR51psl15c

That sorta captures the style, even in HD Flow's Nano 2 shouldn't have been making it look so new with a prompt that says "Vintage"

... again, reasons to need the thing

In all the anime Flow Nano 2 images you both posted it can be seen the cel-shading is imitating new tools, not vintage ones. From comparing in the anime from that video, Whisk's Nano 1 got it much closer to older stuff.

Literally unplayable.

Or more like, Google Ai studio's Nano 1 did cause the watermark is there, both are the same thing... which is pretty much how this started.

So yeah this is just one or two things, there must be a thousand things Nano 2 ain't doing as good as Nano 1.

Anyway for any more doubts, or if used as feedback, I did leave the prompt up there on a notepad. It clearly says "Vintage" nano 2 didn't know what that is.

[pulsar coral]

Well yeah Nano 2 doesn't make them ugly, but I think only the first one was close enough to the actual prompt compared to these Nano 1 images.

Night gents

[shrewd vault]

“हर दिन लंबा सफर तय करता हूँ, बिना शिकायत के। माइलेज मेरा हथियार है… और भरोसा मेरी पहचान। दिखावा नहीं करता, काम हमेशा पूरा करता हूँ।”

[tiny canyon]

@main saddle how much images can i create on pro and how much on ai plus plan cuz nano pro and nano 2 costs same the 0 credits.

Also 4k upscalling is only avaiable on ultra plan?

[frosty niche]

Made a G-Doc, with reference links to help you out with your concern @tiny canyon

https://docs.google.com/document/d/1fFd2PiqB2BS2jSJvKF9kBfpTdwx5qj_OuV3eptW2S4g/edit?usp=sharing

[tiny canyon]

THANK YOU VERY MUCH SIR FOR HAVING TIME TO MAKE THIS.

I just wanna say that I almost never create images in gemini, less control.

I create almost everything in flow.

Also do you maybe know when google new videi model is expected?

[frosty niche]

Naw, just helping out where allotted is all.
No news on that from me atm.

I'm sure the Google team will announce it when the time is right.

Thx!

[gloomy verge]

Hello. When I have some image, example, Female with white t-shirt is in flow possible recolor without creating new images by complete new prompt? From white t-shirt to yellow t-shirt, but image is still same? Thanks for help

[tardy rose]

Hello guys, sorry for this question.. But where can i change the background of images?

[burnt canopy]

Hi guys, has your content generations been failing in the last 30 minutes?

[tardy rose]

on Whisk??

[tiny canyon]

Do anybody know how much credits per second cost on kling official site.

On higgsfield is someting around 2 credits per second

[wet flame]

Which version of Kling?

I used Kling 2.6 and Kling 3.0, they both have different prices on length, and with sound or not.

[wet flame]

Hope nobody is dumb enough for that 👆 Thanks Mods

[frosty niche]

Hope is just another set of pixels in this context.

[mild swallow]

！！！！

[vague island]

so does this just mean u have hit ur rate limit for the day?

[wet flame]

The worst part is that, you can't batch select to delete them, but you can refresh the page for them to be gone.

it's literally gambling at this point to see if it can successful generate or not

[static dock]

hi! Can't upscale to 4K this morning, it just keeps failing, do you guys know why?

[brittle pelican]

well well well...
google classic

[wet flame]

Whisk is still superior

[brittle pelican]

3 days in a row, non stop errors

[sick sparrow]

upscaling images to 4k keeps on failing?

anyone else?

[brittle pelican]

yes faild

[woeful void]

Why does the standard work window disappear when I drag elements into it,
turning into such a small and inconvenient plus sign ?
Doesn't this actually make things more difficult ?

🫡 Please submit this to the bugs and fixes section. I've been blocked from accessing it ))

[quiet kindle]

image 4k not working now, anybody?

[static dock]

yeah seems to be a general pb

[woeful void]

4k is "ultra" tiers video or foto gen ?

[tiny canyon]

3.0

[austere zodiac]

yes

[tiny canyon]

I don’t undersrand

[wet flame]

Not sure, but wanna take it to dm? I'm also calculating the cost

[wet flame]

It was a scam spam

[tiny canyon]

Ofc

[quiet kindle]

yap, but now not working

[dense raft]

totally dead...

[broken kettle]

4k ded 🙁

[fossil python]

Upscaling dead?

[vague island]

yep

[burnt canopy]

In mine, I can generate images but I cant seem to download any of them

[bright heron]

Seems like all my generations are failing right now

[meager socket]

We're down right?

[potent patrol]

https://tenor.com/view/zara-larsson-ew-tbygui-gif-24349575

OH GOSH

[winter dawn]

just $250 a month btw

[dusty aspen]

yes yes and yes

Google Flow Stop working Brazil

[warm orbit]

Wth it’s never been this long💀

[dusty aspen]

im nervous

https://tenor.com/view/scared-scary-water-water-shaking-shaking-gif-17829914067447638174

[winter dawn]

can you please type here when it works again? anyone pls

[stone flax]

Any one else is experiencing problems generating with Flow? It doesn't working properly since yesterday for me ...

[barren mulch]

works fine for me

[dusty aspen]

Please grant us access through your VIP account.

[dusty aspen]

BUG

[raven siren]

Most of my generations are failing...

[thick matrix]

mine too

is there a solution?

[winter dawn]

yea... wait

[raven siren]

How long? It started yesterday...

[thick matrix]

😭

[winter dawn]

last time i had theese problems like two months ago it was buggy for like a week

[raven siren]

If you have acces to an API key you can try that. That''s still working.

[lyric dust]

Is Veo 3.1 unable to create pixel art style videos?

I tried generating some content, but the pixels tend to heavily merge or glitch during movement or high-speed sequences

[dusty aspen]

Bug solved!

[thick matrix]

really?

[dusty aspen]

yes

work for me

[winter dawn]

thanks

[dusty aspen]

OMG fail again!

[warm orbit]

Veo is working okay-ish, but nano banana is completely down

[thick matrix]

yes nano very bad

[odd turtle]

Is the problem only with me?

[unborn bolt]

cant generate anything

[warm orbit]

Why isn’t no one from google team answering what’s happening 💀

[wide ether]

Yeah, nano banana is struggling at the moment

[spring kite]

no mee to

[winter dawn]

ANY MODS HAVE ANYTHINH TO SAY HALOOO?

[bright heron]

Yep, getting zero generations. I'm trying on a higgsfield pro plan and totally down on there too.

Atleast on higgsfield there is a message saying that gens are slow but "due to google's infrastructure"

[winter dawn]

Are we really paying 250 a month for not knowing whats up @tacit halo

[zenith forum]

me too

[unborn bolt]

I just discovered that if you spam 50 generations request you can get at least 2 images

[gloomy verge]

Hello. I want buy some membership, probably, maybe. But I haven t ideas, what to create, how videos. Can you give me some tips/tricks, if it is good idea, buy it, or not buy?