#flow-general・🌊
1 messages · Page 45 of 1
Anyway main thing you gotta know about Nano 2, is it knows a lot about everything by itself and is very updated about it, that helps a lot.
"Zootopia" alone as prompt might only work if there aren't too many of those characters in the image.
When will Veo catch up to Seedance2.0?
any tips or promts for creating this type of pictures?
That's no doubt some sort of pixar style, was just doing that for some reason.
With some rattatouile in it, more likely you'd get the same by plastering it from reference image, but I think it can be prompted.
The fuzzier ones should be a little trickier
did you got similar results?
Still using korean reporter yes
Lemme get it fuzzier
this one is good can i have the promt for it?
Yeah lemme untie it
what did I do? I got Untitled Collection all over my project
Artstyle: stylized 3D animated characters in whimsical pixar-style render.
Physique of characters: Expressive face, slightly exaggerated cartoon proportions, oversized head, tiny round noses, extremely huge large round eyes with irises, slightly protruding ears, ALWAYS plush textured hands, small cartoon body, high-detail animated movie character design, detailed hair.
Skin/Face of characters: Skin ALWAYS covered in easily visible peach fuzz texture.
Clothing: Dynamic.
Composition: warm cinematic lighting, shallow depth of field, bokeh lights, ultra detailed, subsurface scattering skin, soft global illumination, 85mm lens, high-end 3D render, whimsical animated movie aesthetic.
Most of the trick seems to be in the "Peach fuzz" texture, and the composition. So you can probably adjust that, and if you use reference image change the clothing to "Accurate to reference image"
did you use any tool to create this promt or did you do this by yourself?
Yeah my exploding coffee brain.
Then again if you want a tool you can just upload the image to Chatgpt and ask it how the Nano Banana 2 prompt for creating the image would be, and it will help you pretty much samey. Tho it tends to give short prompts instead.
Let's see Korean girl with it
Most these prompts come from practicing with LoRAs before Nano Banana models were available, it uses a LOT of this
Oh yeah you can also separate "Expression:" from it if you want different faces, add "Fullbody only" to the composition if you want that and so on.
its hard to get pictures like this
Nah just save your prompts at notepads and you'll get used to it.
ok thank you
I'm hoping it's a sort of shell they aim to flesh out into more of a real video editor, like a simplefied Google Vids, but we'll see.
Actually I'll take the Google vids format in full
It's not bad!
Not quite as good as Canva's video editor, but catching up fast. All that free stock audio is nice. Super useful.
Doing that stuff locally is faster and more precise, but maybe that'll change.
Both.
I don't think many people used it, so they just dropped it. Probably why they're dropping Whisk, as well. Never got popular.
awesome typo, hahaha
Seems it was indeed Disney tho. It did get censored af.
Still getting watermarks even though I’m on the Ultra plan. This is a total scam and a direct violation of consumer rights
Added a bunch of folders in your project, apparently.
What's Disney? Whisk?
Legend says nothing is more evil. So no wonder. And it's at Nano 2 wherever.
If you try to make some Disney property images you might more often get policy content refusals.
I can think of a few... thousand, maybe million, worse things.
Maybe but if you say something is on Disney levels of evil, even AI can feel it
It did refuse once, brr
These probably pass cause it doesn't seem to be any character in particular cept Maleficent
is it possible to delete the watermark on the video?
is the extend feature in Flow broken
I haven't tried even once but tbh probably the best way is to find some "Veo watermark remover" specifically, not random watermark removers.
Google stuff is very specific.
I've tried with images and the specific Gemini spark watermark gets removed clean if it's an specific remover for them, and you did use the save button to get the image, not right click save as.
man, I feel lucky as I don't see the watermark and I have ultra
I had ultra and the watermark a few months ago, if anyone does have that issue, check if you are in a family group, for some reason by ultra subscription was being overruled by a family subscription of a lesser tier. (Went a whole month like that, using my credits when they should of been free generations and everything, and the most support gave me for the inconvience was a $10 play gift card, I was so mad, cause I needed it to not have a watermark for work. And I had contacted support multiple times about it that whole month and they kept telling me it was a temporary error)
which regions are you talking about?
Check if your account is part of a family group
US
Guido Zavagli (Google)
Google Staff — PM 12:16
Hi everyone - Flow has updated its video generation watermarking policies in certain regions to ensure compliance with regional AI content regulations. Under these updated guidelines, every video generated in these locations will feature a visible watermark, regardless of the user's subscription level. We understanding this may be unexpected and will reach out to affected folks next week with more details.
huh???
which locations
This is a deception of consumers and a clear breach of terms
yikes..
From what I know, it's happening almost everywhere except the US
They said they'll post more details in 2 weeks lol
Basically telling us to cancel and move to another platform
i was looking forward to get the ultra plan on my next payday ...
I can’t believe it… I subscribed to Ultra because of the watermark. I didn’t realize that such restrictions also apply in South Korea. Now I guess I need to think about whether I should look for another solution.
its so weird , my country didnt even announce anything involving ai regulation or something recently. suddenly google is pulling this..
Oh well, one of the USP for the ultra plan was the removal of the watermark, as stated and marketed. Since that is not the case anymore without any prior notice,
Refunds should be in order, so the affected personal can move to another platform.
for real.. this is just ridiculous..
Without the watermark removal, Veo (flow) holds no value for production, no matter how advance how capable the model becomes, it just doesn't matter.
Aside from the 0-credit Veo Relax mode, the no-watermark generation is honestly the only real reason im even considering upgrading to the Ultra plan when payday comes. Most of the other benefits that come with the plan aren’t even usable for people outside the US..
For production, ultra wont make sense outside of US, for personal projects with watermarks well sure, why not. $250 for "VEO" on every video.
Yeah it's been more than confirmed that was it, did it cross through anyone's head to try a VPN tho... then again if Google wants to I've seen them going gorilla with the fingerprinting and canvas ID so that won't always help.
Not to mention even tho they fixed Nano 2 a lot by now, the Nano 1 model from Whisk is still essential and they removed it from Flow too. There's just too many things it still does different and even better and no amount of feedback is going to bring Nano 2 to that.
So that's another thing everyone's being deceived about by being asked to move to Flow.
In light if the comments about the watermark being region dependent, I played around with the VPN and switched it to one of the countries discussed above and didn't see the watermark
I did mention this coulda been happening cause Gemini staff did say similar stuff months ago. We coulda tried that first lol.
But everyone seemed so convinced it was a bug, wasn't even just one person anyway.
if I want to avoid the watermark, do I have to change my VPN to a different country?
Which countries are allowed to create videos without watermarks?
If you have Ultra plan and getting a watermark yeah, at the very least you'll have to be in USA or maybe some south american countries, I'm there and haven't seen watermarks in mines.
No one cares for what we do here... mostly cause we don't do much.
It seems like Google wants to do something to make things more difficult. Even faking a VPN to the US couldn't make the watermark disappear.
And I just paid Google for 3 months for the Ultra plan :)))
Try different VPNs, apparently it's working for other people.
Just not Korea or China.
Invite friends, get rewards https://store.mattel163.com/invite/skipbo?code=SQMEAIVTQ&s=L8I73F
My first go to is always this VPN client. https://www.vpngate.net/en/download.aspx
You only need the one that says "Download SoftEther VPN Client + VPN Gate Client Plugin"
And you can go nuts trying different VPN countries.
You turn a VPN on then go to https://whatismyipaddress.com/ and check if it's registered as VPN and if it's not those are more likely to work cause they are still registering as residential home IPs and not VPNs
Thank you. I'll give it a try.
If none of that works, more likely Google put some canvas and fingerprinting recognition. And yeah the only way I've seen to go through that is via Tor browser, but that's really slow.
To go back to flow talk, I finished this drone from the other day
Flows mechanical logic is indeed not so good, like that user generating car images said
I know that image's blurry, but yeah that's why I'm trying to fix it
Then I tried the counterpart of "Anatomical and morphological" which should probably be "Apply mechanical and aeronautical logic."
Got bulkier but does seem like something that flies
Nano 2 likes to bulk everyone up anyway.
So then went like
"Keep it's design accurate to reference image. Do not make it airplane designs, keep it's accurate to reference image X-wing look but not from star wars, fix it into being completely symmetrical.
No cockpit it's an unmanned drone.
Apply mechanical and aeronautical professional logistics so the wings make sense. Keep it's engines accurate small and right where they are, do not add rotors.
Make it be at a hangar completed and ready."
It's official, Discords are always ran like this by many other people too anyway.
And yeah I know my prompt there is very dizzy
It's 1 am here ffs
Well let's see what it did anyway
Probably one of the most important things about cars and airplanes is at least on the outside they should be symmetrical, maybe even aerodynamic
Tho idk what that first thing is.
Well there's Google staff around here, and they did listen to some of my feedback on the models. It can only be real.
But now you mindgamed tbh. Also they're surely the ones running Flow and Whisk.
When they say updates will be done, they happen.
Eitherway, Nano 2 does seem to have some probs to take the most logical decision to build a machine.
Gotta be way more clear about it
The blurry reference image is more than a challenge too anyway so yeah that's also on me, still...
I went more aeronautical there, it can sure do things
that looking fire 😄
umm well the official google top level domain links to it. so I'd say its official
i think the end result matched the reference (that blurry one right)
im DMing you I want to make sure you get this resolved check your messages
Im telling you right now, they told me innacurate information like that multiple times during the month and a half I was having those issues. (It worked without a watermark for two weeks, and then one of my family members had subscribed to plus to give the family Youtube Premium and for some reason that took priority I had no idea that was the issue, until finally a support person who knew what was going on helped me), I made a thread on here about it here multiple times, contacted support about it multiple times, they told me lies like that, like the two weeks thing, that it was something new. And I complained the same way because it was a change of terms, but the reality is that they didnt change the terms its their support just trying to get a closed ticket AND its really messed up. I bet if you look at the flow page it doesnt say ULTRA under your user account in the top right. I bet you they are charging you tokens even for the 3.1 low slow mode, (If you can even see that option) I'm litterally VJing with visuals right now that have the Veo logo on them from that time, and every time I see it it makes me mad. I'm going to DM you I want to make sure you get it resolved, cause they are lying about them changing the terms.
question: whenever i edit images a lot on flow i end up with image i like. but: its blurry, ridden with artifacts and straight up bad. the composition is nice, the angle and so on. but trying to have it clean up or re-render either just gives me the same image or does nothing at all. how do you guys deal with this? and even downloading an upscale of 2k doesnt solve this. i have an image attached to show what i mean: you can see it in the textures mostly. how to have it re-render but clean all this up (but not change the image/angle/person)?
where is the watermark? we love the watermarks, please bring watermarks to NanoBanana2!!!!
It doesn't always do it by default, at most it will give you 2K yeah, but you can still ask it to better it with prompts alike
"Change nothing in the image leave everything accurate to reference image but increase image quality.
[Image quality: Remastered 4K HD ultra high resolution, sharpest definition, denoised, unblurry, clean lines, High contrast, fix any jpg artifacts.]"
Or that's my go to, can probably be better, even if it doesn't allow you to download 4K without Ultra plan it does indeed prepare the image for it.
If you don't want to change anything in the image it usually can't go more complicated than that.
In fact even while making the image you might wanna end it with an image quality prompt, not after.
Nano 2 is all about coherent precise prompt instruction.
That'd probably be false advertising, not breach of terms, unless the terms of service specifically mentions the watermarking differences between account.
They usually can tell where you are.
Yo guys
My 20 friends and I will share one Flow account, Will this result in the account being banned?
45.000 credits
Can we get official confirmation about VEO watermark? Cause i believe for me and many others it makes flow useless
You can read this at the labs-bugs-feedback "Veo watermark on ultra plan" thread
Basically if you're seeing the watermark on your ultra plan video it's due to your country's laws about AI. Some people fixed the issue by going through residential VPNs.
U.S. residents don't seem to be getting the prob, most south american ones either.
But for sure China, South Korea, some European countries are having the prob.
I don't think they'll care for that specially if the account is in the same location like a studio, but the Google account linked to it might give you some probs cause of opening at various locations, they won't ban it but yeah you're gonna have a hassle cause it will ask for security codes or text messages often so good luck with that.
I know cause I've had secretaries opening the same accounts from asia to south america and sorts.
Got an issue guys
Flow issue here, the message keeps appearing, stating that my operation is too fast and to please try again later, but I haven't been performing any other actions. This has been going on for several hours now.
Have you tried only asking it to generate 1 or 4 images at once.
It's not a good idea to ask for more than 8 at once tbh
If not it should be resetting itself in about 24 hours.
Only 4 images at once
god flow sucks compared to img x and whisk. why do they keep getting suckier. img fx was better then whisk. whisk was better then this horrible flow. DONT GET RID OF IMG FX
at least imagefx has a good clear name
Got this email about about Whisk (also ImageFX in a seperate email) shutting down (was in my spam box, but it came 2 days ago and haven't seen anyone else share it... might have missed, but important info that people have been asking about regarding migraton of generations):
I noticed that everytime i generate an image with nano banana pro or 2 it alwyas has this yellowish temperature/hue to it.
My goal is to create images as similiar as taken from an iphone camera, does anyone here have any suggestions on what to do in this case (like adjusment settings)
I haven't tried using specific camera settings, but giving it directions in the prompt or input images for color palate should help. For consistency, once you have a prompt segment that works for you, save it and use it word for word in all the prompts you want that same look in.
just write "remove the yellowish temperature/hue to it." in the prompt for your generated image
In fact, no such laws have been revised or enforced in Korea yet. Stop talking nonsense
Well a Korean user did get the watermark, so that's what we know.
Right now since no one has given a list of countries affected we can only go by each affected user. And there's no doubt officially it's a regional thing.
I'd try "Illumination: Clear natural realistic morning light" and sort, sometimes it just gets it by the simpler things.
I have a question, can you use an external voice, like from Elevenlabs, for example, and use it to generate a lip sync video for the audio with flow.
anyone has the problem that the library doesn't load thumbnails at all now?
Greetings all, had a quick scroll and couldn't see anyone else having the issue - I'm getting "failed, oops something went wrong" on anything I try to do since 11.00 UTC (ish) - have tried in a new project and with/without ingredients on images, with nano banana pro, across various settings, any ideas?
--fixed: was my VPN, switched endpoints, now working.
veo4 when ?
nano banana 2 is having a problem following prompts as of the moment .. is it just me or is there anyone else having problem as well
nano banana pro as well, now you come to mention it.
and in the models now i can't find nano banana 2
hello weekend maintenance update I guess, cup of tea it is then 🙂
I clicked Save to project on my image by accident, but how do I remove this copy?
nvm, I found it on the sidebar with Images tab ✅
guys did they remove the delete feature on extend???? how to delete a clip
once i make it i am doomed to keep in the glitches or restart entirely
Just select the original clip in the sidebar and extend it again.
My generation are all stuck at 99
Hello, please, can you help he, how is name or tag of this? I mean "+" in left side. Have it some name? I want to write prompt, that camera must be same as on image added by "+", but I don´t know name of it.
You mean the "Plus" symbol? that's what we call the "+"
He doesn't speak english gotta cut him some slack
I add some image and I want, that generated image will be have same camera position as this image. But I don' t know write it in english to prompt.
How in the world can I tell the model to stop adding music to the generated videos!?
I don't want music and it seems like it started to want to add music lately
Try
"Camera/Perspective: Always accurate to reference image, view angle/framing/distance locked to the same as reference image."
I use angles a lot and pretty sure perspective changes it
It can also be put next to
Composition:
But that depends on how long your prompt is, sometimes short prompts have better effects on some images.
Tbh most video AI generators would listen to literal prompts about it, stuff like
Music/Background BGM: None, never add music.
As it'd be obvious, but I heard VEO can be a little stubborn about it. It should be obeying that tbh.
Just give it's own spaced slot, or put it right at the top of prompts to be sure.
I think Nano Banana pro is the unnecessary one now since Nano Banana 2 is pretty much an improved update of it. It'd be noice if they put Nanano Banana 1 instead of pro now, since that one's still way different.
It'd probably waste less resources too, and even give a better impression.
Using Ingredients
actually nano banana 2 is like , distilled version, Pro is actually really great at refining something from an image.
when it comes to prompt adherence, i think nano banana 2 generates more variety outputs
Yeah but now that the other user above made me check about camera angles, Nano 2 is obeying them much more than Nano Pro.
yeah thats why pro isreally great at doing "precise refining"
Not to mention all the anime prompts I've been trying, Nano 2 is obeying reference image more, and pro has an annoying habit of completely changing the artstyle for no reason into "PVC" figure-like.
All day Nano 2 has never done that to me with any prompt.
The difference starts being abysmal already.
Nano 2 also keeps physique more consistent than Pro atm, Pro kind like dries people up or something.
prompt issues
None are doing any of this as good as Nano 1 tho, but Nano pro is consistently falling behind even more atm.
Maybe it's my prompts yeah, but Nano 2 does get them.
I do clearly divide physique, facial structure, head to body ratio.
yeah thats why its "distilled" version
Artstyle etc, not really going random about it.
they tuned it so that user dont need to have an expertise in prompting
Pro just seems to prefer "ugly flat" for some reason... I'll try prompting better and see tho, maybe "opposite of what's considered ugly" would work cause why not.
Does the "flag output" - "inaccurate" option on the generated videos actually help to drive updates to the model? Just curious as I've been hitting that quite a bit lol
Yeah they listen, send as much as you can. Just don't expect it to be immediate tho.
I think the record is 3 working days, and if a lot of people complain about the same.
"send as much as you can" .... oh believe me, I have used that option quite a bit!
Well just don't break everything else... then again
To me is the opposite for some reason. Banana2 feels broken. It tries to improvise many times, bringing in things that have not been asked, or re-imagining things in a different way when you try to alter the viewing angle. I am doing a lot of product photography and this has been a pain with Banana2. I am mostly using Banana1 and Pro at the moment and jumping into 2 when they other two constantly fail.
Why is nano banana 2 so bad now?
Everything it generates is either distorted or completely off from the prompt
Imagen 4 works much better
Maybe I'm just being more specific with Nano 2, admittedly by itself is not at all good for taking the right decisions so I did start taking them for it.
In fact that would explain a lot.
Like those drone images I posted yesterday, they went random bad until I really started getting serious with aerodynamics and symmetry.
But then I go with Nano Pro with that and it just won't listen to me well at all.
My prompts separate into encasing, more than being a clear connected paragraph.
Also yeah I've seen it doesn't respect photography too well, gotta lock the camera perspective and things alike or it will just push and move everything all over the place and change it.
idk I just work with random anime image posing and that does go well.
... not at all as well as with Nano 1 tho.
Yes, this is something I've noticed. I have to give it the "don'ts" all the time. Don't do this, don't do that, keep it this, etc, etc... Is so frustrating at the moment. Especially when you're trying to change the viewing angle, oh man... disaster!
No wonder another user was asking about the camera perspective earlier. Well they'll read that.
Anyway from the majority cause I did see that happen all week, when generating the image it's probably better if Nano 2 doesn't have creativity freedom for moving the composition of the image's angles and perspectives or move/push props and characters around the background unless user asks for it, it'd be much easier to do it the other way around anyway
And the looser the better, everyone probably recognizes it as "camera" I sometimes use "Viewer perspective"
I'm using Ultra, but a watermark keeps appearing. Where can I check for notifications about the watermark?
You can read this at the labs-bugs-feedback "Veo watermark on ultra plan" thread
Basically if you're seeing the watermark on your ultra plan video it's due to your country's laws about AI. Some people fixed the issue by going through residential VPNs.
U.S. residents don't seem to be getting the prob, most south american ones either.
But for sure China, South Korea, some European countries are having the prob.
Also apparently Labs staff should pronounce themselves more about this in few hours, it's probably still Sunday for most of them so yeah, business hours.
Maybe some users also already went for a "Veo watermark" specific remover. Cause what else, no idea how's that going.
idk even if countries put laws against AI videos without watermarks, tbh all that's on the internet, countries that don't run their own servers shouldn't really have a right to regulate what's on the internet, it's their prob. Most people working on AI distribute their stuff on the internet too, not as physical stuff around countries.
Cause it's weird mostly USA is free of this.
I’m an Ultra subscriber. Why is the Veo watermark suddenly appearing on my videos? @real wyvern @cedar bloom @unreal swift
You can read more in the “Veo watermark on ultra plan” thread in labs-bugs-feedback.
Basically, if your Ultra plan videos still show a watermark, it’s likely due to your country’s AI regulations. Some users have fixed it by using residential VPNs.
Users in the U.S. don’t seem to have the issue, and most in South America don’t either. But it’s definitely affecting China, South Korea, and some European countries.
Labs staff are also expected to comment on it in a few hours. It’s probably still not working hours for most of them, so yeah gotta wait on that.
Also some of the countries don't seem to really have regulations on this matter yet they're still having watermarks on the VEO vids so dunno there. Gonna have to wait on it.
I honestly cannot understand this situation.
Are you suggesting that I should use a VPN? I don’t understand why I would need to do that. Using a VPN can introduce significant security risks, and reliable, secure VPN services are quite expensive. That would mean I am expected to pay not only for the Ultra subscription, but also for an additional VPN service. This makes no sense.
More importantly, adding a watermark without any prior notice is extremely unreasonable. If Ultra subscribers are still going to receive videos with a watermark, this should have been clearly disclosed before the subscription. There should have been an explicit statement such as: “Even with an Ultra subscription, a watermark may still appear.” Without that transparency, this situation feels very unfair.
I understand that there may be regulatory considerations involved, but to my knowledge, my country is not currently part of any regulatory region that requires this watermark. It seems unreasonable to apply such restrictions broadly without properly verifying the actual regulatory status of each region.
I am currently working under a tight deadline and urgently need to produce videos. I genuinely do not understand why I should be forced to generate watermarked videos after paying for an expensive Ultra subscription.
I would appreciate a clear explanation and an immediate resolution to this issue. @inner ice
Well it's just an option that worked for some people until. But yeah we can only wait till staff replies and I imagine for some of them it's like 2 AM atm cause murica'.
We don't really know what else is gonna happen yet.
But more likely those watermarks won't be going away, I had a similar prob with them at Gemini with the images not the videos, and yeah even paying the watermarks never went away, so I just adapted and found a watermark remover specifically for the spark watermark thingie and been fine still making money.
Idk if there's one around for the VEO videos specifically tho, seen them but never tried them so idk.
Cause for videos it was a little too expensive at flow so I just had to use other video tools that responded the same to my prompts, I think even better but gotta really put some dam effort in the prompting, but wierdly those don't apply watermark even free.
But I do totally need a Nano models image to start the videos, it's awesome really.
Working fine on my end, location south america.
mine isnt working properly
sometimes it just says error occured
Also just in case Google staff never said to use VPNs, it's just a thing people do everywhere. TBH residential free VPN ips work better than payed ones because they aren't flagged as VPN everywhere.
And this trick ain't working for everyone anyway.
To go back on topic with Flow, here's an example of how real images get changed into something else by Nano Banana 2, I tried the obvious prompts alike "Keep the cat accurate to reference image." and the obvious things.
The real cat is the blurry one with 1:1 aspect ratio.
The lack of aspect ratios might also have to do with the changes, in some of the images it flattened the poor cat when trying to keep it accurate, I also tried to unblurr the image simply.
It's not that easy really.
Can't prompt "Never change the aspect ratio" which is something that would work at Gemini cause it allows manual aspect ratio for Nano Banana 2 and Pro to crazy extents.
This one's probably as close as I got, but it did move the camera a little.
can anyone explain why im encountering this issue?
Other simple prompts in other images failed too?
yep some are failing and some are generated
Well if the site does work, sometimes it's just your image or your prompt might have a conflict. Might wanna check if different ones are working.
And if you do need the reference image, sometimes upscaling it will make it pass better.
i checked with other prompts now no prompt is working
how do i fix it?
Was it working fine some hours ago?
yep last night when i opened it was working fine
Cause if so it's probably just a matter of restarting your router/wifi and your pc, updating your browser. Or simply waiting, I did get that prob when my connection slowed down the other day.
Needs to be dramatic tho
im using web app
VPN wont work
Well like said for some did for others didn't.
I think it's mostly about using residential VPN or not, retail vpn are useless, except for watching netflix.
Once it's flagged it's good for not much else.
All those VPNs that youtubers advertise are the first ones to get tapped for sure.
Mini Shredder got caught for using one of those see
Nano 2 does listen if you correct the scales tho, ain't hard really.
Simple terms alike "Gigantic, bigger than, taller than"
eh, mostly ok, not even trying
Always error ? flow please
A beautiful woman is dancing
Why are we getting "Veo" watermark on the ultra subscriptions?
You can read more in the “Veo watermark on ultra plan” thread in labs-bugs-feedback.
Basically, if your Ultra plan videos still show a watermark, it’s likely due to your country’s AI regulations. Some users have fixed it by using residential VPNs, but this solution ain't being endorsed by Google.
Users in the U.S. don’t seem to have the issue, and most in South America don’t either. But it’s definitely affecting China, South Korea, and some European countries.
Labs staff are also expected to comment on it later, it’s probably not working hours for most of them yet atm, so yeah gotta wait on that.
Thanks for letting me know, friend.
Hey! This particular error is due to policy: we don't allow uploading images of photorealistic minors.
When will you correct the bug not allowing to use frames from flow video with minor in prompt in nano banana, or even modify frame from nanobana with minors? It been 3-4 months you said you'll do it
Ok, so I was enjoying Flow very much after the latest release. Great improvements and the guard rails seemed to work a lot better. I was able to upload images of children without them being flagged. Now all of a sudden everything gets flagged. Even if I try to use an image of a child that I just generated as an ingredient it gets flagged. I hope this is a temporary hickup.
We also can modify (for example make it cartoon) images with minors in Gemini (Nanobana there is painfully slow) and any other tools using Google API. The only place we can't do that is Flow : the official place from Google to use NanoBanana. I understand it for video, but for images made with nanobana, it's a weird policy that it works everywhere but not is the official place.
It worked like literaly 15 minutes ago. I uploaded some MidJourney images and it accepted them without a problem... Now it stopped. I know it was like this before, and I aslo noticed that you can do it through API, but after the latest update to Flow it also worked there. I thought they had finaly figured out that is was totaly weird that it works everywhere else except in flow....
This is the image I uploaded without a problem. Now if I want to use it as an ingredient it gets flagged... That's just stupid.
Same for this image which I just generated with Nano Banan 2 in flow.... Can't use it as an ingredient....
When veo 4
If some other images passed and one didn't sometimes it's just an error with that one image. and upscaling it to %400 here would sometimes let it pass https://imgupscaler.com/ even Whisk also had that error sometimes. But if it wont pass even after that yeah its more likely not passing due to policies.
Hello, everyone.
Has anyone tried to get “slow shutter videos”? I haven't tried to put it in Promt yet—I only get time-lapse and some ridiculous blur...
But I want to get something like this around the hero...
I'm talking specifically about videos, about the effect of shooting with a “long shutter” like 1/3, 1/10....
bro it doesnt let me to download 4k vids or 1080p vids or 720p visd
like it only let me download gifs bro im paying 200$ per month for ts and it still aint working
bruh help me please
Everything is working for me.
when i press on 1080p for example nothing happens
i can see the options
but it aint workin
like it genuinly aint
help me out chat
Give it about 30 seconds to prepare the file, it'll show up automatically in your Downloads folder once it's ready!
When you use this image as Ingredient, you can´t generate images? If I understand right.
Yes. These were actually generated with Nano...
Or you can´t add image as Ingredient?
I don't see a new update since the 3rd, but something updated ... I no longer have that flickering on the flow page anymore! 🙂
Tail is misplaced
I've just been doing the janky thing of uploading my clips to drive and editing with the TikTok editor on my phone. Whats the next step up for video editing?
Can Nano Banana 2 understand proper sizes, measurement, distance, and other real-world physics application?
I'm trying to use real-life measurement to describe an obj
not really. I played around with that a lot. It only knows what relative proportions should be for normal objects
i'm trying to put this damn thing on my avatar
what do you want it to look like?
i see, now i get it. That's why I describe the enamel pin on the collar as size as the US dime, it gets it
but not the real-life dimensions of that dji mic 3
oh ic
I wanna make it look like it's wearing a legit dji mic 3 just like in the amateur photo (guy wearing blue shirt)
for normal objects, do we have a list or just any common object?
can you use the amateur photo as reference? Give it your photo and tell it to replicate the size and placement of the mic in the reference photo?
tried it but doesn't yield good results, will try to use the standard object like a 2 british pound coin
Understand Nano Banana 2 hex/rgb codes? If I write example "Recolor wall to 250 250 0." It recolor wall to correct colour?
??
is there a way to keep the avatar in my videos using the same voice between different scenes?
The images were in the project already. I dragged them to the prompt and started the generation. Than it gets flagged.
After you added image to prompt and start generation, you can' t generate new images, yes?
I feel like it has less of an opinion.
Can't. Just have to wait and hope it works later.
Yes
It is writing some problem?
Yes, that they don't allow to upload content with minors.
Okay. So, I know where is problem.
Do we have a date of when it will be possible to transfer my creations of whisk to flow?
No news on it yet, but should be before the month ends.
Hi @green needle regarding the transferring of your creations from Whisk to Flow. In the upcoming weeks you should see an opt-in button in Whisk that will let you migrate your library directly to Flow. You can also download your content any time before April 30 but please be rest assured your creations are safe. Just be sure to do so before the above mentioned date because after that any remaining content will be permanently deleted.
I hope this answers your question☺️
Good morning/evening everyone ☕️
2D cartoon honey badger standing on two legs in a green bamboo forest background, the honey badger looks strong and brave, slightly moving and blinking, cartoon animation style, cinematic lighting, smooth animation, text "HONEY BADGER" written at the top of the screen, vertical video, 4–5 seconds animation
!!!!!
What is this
can you guys rate it out of 10 and give points for why did you rated so and give suggestions to improve.
Flow is detecting an uploaded photo as a prominent person, but it is not!!!! How do I resolve this?
Sometimes image recognition models detect a face and assume it’s a known person if the dataset has similar patterns. Maybe try cropping the image or using a different photo angle to see if the detection changes.
I’d say around 8/10. The environment and colors look really nice, especially the lakeside vibe... it feels calm and relaxing. The character walking animation also looks pretty smooth.
Maybe adding a bit more movement in the environment could make it feel more alive... like subtle water ripples or slight motion in the plants and trees, as if there’s a light breeze. That might make the scene look more realistic. Overall it’s really nice!
Hello. I want to ask. When I create image in Flow, can see my images other users too, or not? If no, how they can see my creations? Except Discord. Is other way? Or only place it on Discord? Thank you.
Sorry how many Nano Banana Pro and 2 generations does a Pro sub have daily?
hello, when i add videos to scene builder its auto muted, and i have to unmute every 8 sec clip seperately, and when i rewind to watch the full scene its muted again. how do i fit it ?
Hello fellow Flow users! Reaching out to ask if anyone else has seen this issue in Flow. I attempted to get rid of all of my Flow/Whisk data (just to start fresh with new projects), and I'm not sure if deleting everything at once freaked out the UI. Some projects that were supposed to be deleted are "zombie" projects where it says it can't load any data from them.
I can create new projects, and make new images, upload assets and all of that. But going back to Flow's home screen and re-entering the project, it didn't save any of the generations i made or assets. In addition, clicking on "delete all" in the Flow history library causes an error saying "Something went wrong." I'm assuming this is some sort of backend issue where Flow is haivng trouble deleting data at one time. Has anyone else ever run into this issue? And if so, any fixes you'd suggest? Thank you all
🙂
FYI-I've tried signing in and out of Flow, different web browsers (primarily use Chrome) and clearing history, cache, and cookies from said Chrome browser.
sorry mate not tonight
google basically said it depends on the country you are in
@tacit halo Hi guys, in Flow i'd be super nice if we could delete edits from the bar on the right. Once you're working with edits, sometimes you don't regenerate over an edit but over an original frame so the image doesn't degrade. The possibility to delete edits on right thumbnails bar would mean we can have a more tidy "history" of the edits that work.
"prompt must be provided" even with a prompt?
sometimes I work for a prominent person and there isn't a way to tell google I have permision from that person, most of the time I have to faceswap the face, make the image and them faceswap the result with the face of the person back
where did the expand button go in ingredients to video section ? i want it back
is there
it comes after generating video
what are you talkin about ?
there is or was a feature when selecting ingredients to video before the new ui was implemented , you could inputa prompt , then click expand , select cinematic film noir , or action figure , i am trying to look for that option bc it can takea basic prompt and convert it into prompt magic before genning your video ,
try after creating video
extend and expand are two totally sepereat features
the old "Expand" button and those specific style tiles (like Film Noir and Action Figure) aren't visible as separate buttons in this specific view anymore.
The UI has shifted toward a "natural language" approach
i asked gemini it says its gone
how to make video without watermark?
can you really not do a square generation ??
Holy moly WHY do we have to constantly re enter prompts? I make one image and I can't recreate it without typing the whole prompt again. This is beyond obnoxious.
No redo button either...
Is anyone able to provide the limits for NB2 with Pro vs Ultra?
no
can't
Well you could move to USA, they don't get watermarks there. I bet there's a way or two as to replicate that, even if VPNs and proxies didn't work for it for some people.
I was making a scene of a man opening a shipping container, but the container looked off.
⭐ still giving praise for the atmosphere it delivered
That lighting is awesome!
I never thought I'd complain about image detail, but I've noticed that Nano Banana 2 has been exaggerating the amount of detail in images, often making them cluttered.
Thank you!
Who else was seeing this
Oh so that's where the bone was
Trying to get the accurate 1998 Godzilla but Google's like "This generation might go against our policies"
Something's missing here
There we go
Looking for feedback on a next scene, something to introduce the character.
Which is better or what should I change?
Second one, something's off with the paper doors in the first one
That or she lives in a half sized house
Oh I see it now. The corridor seems so narrow lol
Maybe it's cheaper rent, idk with azns
Ah well she does live with her parents
just ranked up to
I also made this version with the old character design. It gave this room
Yas bish, slay
What you think this looks better overall?
Yeah surely does, dunno how you fixed the perspective
Some Japanese do somewhat live at floor level like that
a photograph from a distance of this woman sitting on tatami floor in a japanese living room that is mixed with western design characteristics. she sits behind the wooden desk to write down notes, she looks focused, low view angle, highly detailed
_
A bit of everything
Ah nice
Yes I wanted to capture a certain vibe.
I remember being in several living rooms there that had a mix between sitting on tatami floor and some western furniture
Also random ✨ Nano Banana 2 is still good ✨ moment
Thx
A vibrant freshwater aquarium close-up featuring a group of Otocinclus catfish and bright red Cherry Shrimp busily cleaning bright green algae off a smooth river stone. The image is split—one side of the stone is fuzzy with algae, the other side is sparkling clean. Bright studio aquarium lighting, shallow depth of field, 4k, macro photography.
is there no search function for projects like there is with whisk?
there's a setting you need to adjust: clear prompt on subnmit, turn that off
I noticed that the message at the top of the screen in Whisk changed today. Before, it said it would be possible soon, but now it says there's an option to make the transfer. Is it already available?
you can change it in the settings
Please help me
create professional shots for the product. focus on the product. smooth scene transitions. realistic background. product design unchanged.
Now that Whisk has been Deprecated. Will Google Flow get a 1:1 aspect ratio in images?
I know this has been asked a fair bit, but is the only reason its not implemented in the platform is to push users over to using the API?
1:1 Aspect Ratio, used for product images
Is there any updates about the watermark??? All of a sudden started appearing in my videos. 😠
Curious, anyone got a list of media art styles that work?
How much can i enchance an image
The lighting here helps alot more.
Is anyone else unable to upscale videos to 4K? It keeps saying “Upscaling Failed.”
im also getting "Upscaling Failed!" errors too for a day now. wonder when i can download again
been facing this since the last week and I'm on the Ultra Plan. how am I supposed to use this for commercial work?
man ... prompts that are in no way sexual are getting flagged. even taking it down to "they talk about the day they had" still gets flagged with 2 fully clothed subjects
what are they wearing, anything that implies see thru, transparent, etc, is going to get flagged just need to reword
pajamas. like a pajama shirt and pajama trousers.. id know if was even nearly suggestive.. nothing like that. id send it but the subjects are of myself and some irl friends
Where is Veo 4?

and my initial prompt was a comical scene where 4 adults (the subjects) start a disproportionately heated argument about some niche fanfic disagreement regarding the sonic franchise. that kept getting rejected as sexual content.
apparently arguing over sonic the hedgehog is sex
i know. im reporting it, not complaining
i even thought.. maybe its interpreting heated argument as potential violence (even though the denial speicifcally stated sexual content) so i added that the argument is purely verbal, with no physicl contact happening whatsoever.
It's kinda annoying nowadays
Got it, thanks
There is a way to delet the image adjustments i dont want, and keep the ones i want?
i have an ultra account and still have the watermark any fix?
You can read more about the reasons in the “Veo watermark on ultra plan” thread in labs-bugs-feedback.
Basically, if your Ultra plan videos show a watermark, it’s likely due to your country’s AI regulations about AI. Users in the U.S. don’t seem to be getting the VEO watermark, and most in South America don’t either. But it’s definitely affecting China, South Korea, and some European countries.
Labs staff are also expected to comment on it soon.
It didn't work for everyone, but some users fixed it by using residential proxies, watermark removers specifically for the VEO watermark, prompting a black bar at the bottom to crop that part of the video with the watermark etc.
For the past two days I have been experiencing an issue with my Google Ultra subscription. I cannot upscale or download my video scenes in 4K, and I keep getting the error: “Upscaling Failed! Google Ultra 4K.”
Any help would be appreciated.
Not only that, it will not allow ingredients to be added. Also, all the prompt results are in some other lang
Someone can guide me? how i Can extract the frame of a video in scenebuilder?
Hey, I have a question on Google flow, how would I use my own personal voice on veo 3? And when are they going to do clips longer than 8 seconds, or extend to 3 minutes
You can't
They won't

holy LORD WHY do we have to reenter the prompts every single time? why cant you keep them like whisk
Go to the cog icon, then turn off "Clear prompt on submit". And it'll stop erasing.
thank you very much
Then on the last image you generated you can also right click and pick "Reuse prompt" and it'll write and reuse the prompt and reference images from that one automatically.
It's actually a little better than Whisk when you get used to it
Tho yeah clear prompt probably would be better in "off" by default, but then you wouldn't lurk for options would you.
Reuse prompt is useful if you have to refresh the page too.
where can i find updated video tutorial on how to create a motion sequence video on google flow ?
im stuck, ive created 5x images using whisk> downloaded onto my laptop> then started create new project > simply trying t create a animated youtube shorts video
A cool bot for your server this is a moderation bot + fun bot too this will make your cool server looks more cool /invite to add this bot to your server and /help to get help hhoe to use this bot this is a built propose bot use /commands to get see the all commands
Developer - @trail solstice/samarpro
Info_Rug
https://youtu.be/2VsURIZr19E?is=S01RUurovSHmvTiZ hiya👾 i have creaded a framework build up on 5 logic modules creating artificial dna memory and reconstruction as well as character stabilisation🧬 if ur new to google flow and struggle with deformations, texture changes, unnatural joint movement or other error corruptions feel free to ping or dm me ill may can help u out 
a video i made with google gemini using personal modified parameters
the character design, landscape, and scenario is made by me too
more info:
Character and Weapon Design:
-the weapon is inspired by the VSD Marksman Rifle
-her body is inspired by analysing anime sniper art
-her body was transformed and heavly modified with biont implantates...
Anyone care to speculate why Flow is telling me that I may be violating its policies with clean head shots of a woman in a Cyberpunk setting? Nothing untoward whatsoever and yet "oh noez you're being naughty!"
it may be a copyright issue. u can paste the error to google gemini for aditional analysis.
Gemini knows whether it's a copyright issue? Irritating when it's a screenshot of my own custom character 😒
gemini can analyse google flow errors effective as it has a similar logic design. paste a screenshot from ur error and may add ur prompt after that.
Will try, thx
Can you imagine if people making real films had to deal with these idiotic filters every time they tried to film a scene?
not realy as in the industry ai is heavily modified with structural logic to perform on specific tasks with specific structures preventing errors (most of the time)
I think it'd irritate me less if Google wasn't so bad at UX. "Oh your video was rejected, let's just go ahead and also delete your prompt for you... there you go! Have fun typing it all again!"
ur also need to modifie google flow otherways with the base logic without any structure it will take ages to fix ur errors with prompts (i talk bout flows instruction options)
That's the really annoying thing. Not only do they misclassify your inputs, but they have no concept of customer support, and deliberately leave you guessing as to what the problem might be.
I can’t use my own voice on veo 3? So how does it voice sync
“Anyone here experienced with Google Flow? I generated images in Whisk and need help turning them into a sequence video. Happy to pay for a 30-minute walkthrough.
pm me directly. thanks
I generated the image, clicked it, added another prompt, and generated. Now it shows the second image in the library/grid-view. It does show the stack sign, but it'd be better if it showed the first image when hovering over it and could be chosen.
Hey guys, does anyone know if Veo 4 is on its way?
Anyone know how to get the generated ai to talk slower/normal? I put it in the prompt and getting the same speed. Mines talking too fast
I think we need more mods to answer/help 'em
We need people that’s experienced with veo 3 and can answer questions
Hi All,
I am new to discord as well as new to AI Google flow. I have pro plan of Google Gemini and Google flow. I want help from the master who knows Google flow. Can I share screen shot of Google flow I am am not getting options in timeline to add scene to complete my video.
Please help.
Thank you
can anyone tell me how to generate a veo3 animation base on the prompts that I placed?
I've been very descriptive about my prompts but Flow seems to be incapable of reading them
Hi try to create the scenes with short texts, not more than 13-15 words
then combine them in a video editing tool
Thank you, what about keeping the same voice over 3-5 different scenes?
its really challenging, I use a voice changing AI tool after I finish and export the final version of the full video
Can I message you and pick your brain a a little?
When veo update ?
Hey, not sure what prompts you are using, but I recommend you generate the character/ environment using nanobanana pro, and add it to the flow prompt as an ingredient
Take advantage of the big complaint of Antigravity users and do yours too. It’s incredible that one being a Pro cannot develop their creative and development processes because for them being “Pro” is “Plus”
can we please get the ability to press left and right or up down arrow keys or even just click a button to see the next image or video generated?
Right now I have to click back each time to click another image. It would be great if you could just click an image and press next to see other images genrerated in the same prompt.
This lets you see the minute changes between images, which you wouldn't notice if you back out each time
i can created here?
yes i create image using nano. any way someone can do a live screenshare video frm start to finish ? maybe something as simple as an object talk style video for shorts?
or video tutorial would help out
Maybe it's not in the dialogue but in the animation, so try putting this at the start of the prompt:
"[Pointers: Make the animation coherent natural and unhurried with enough frames to be fluid.]"
(If doesn't work for some reason do combine it in dialogue prompt with something alike)
"[Character dialogue: Natural coherent and unhurried enough to be fluid.]"
I've seen the "Unhurried" prompt work to fix video/dialogue speed not just with VEO but worse AI video generators, at dialogue you can also add the language and tones the character will use or what it will say, the order must be arranged correctly tho.
Is there away to transfer motion to video? Or is that something on the roadmap this can already be done on luma
That pointers thing at the top of the prompt is overall a good way to start a video wherever
Ok then, this should be more than a challenge for this, I think it did well enough
Ok again, ref:
And video
I failed on the scale prompt and eyes closing, but yeah can be done too, and at first the animation pretty much was doing whatever it wanted, had to prompt all the way like this
"[Pointers: Make the animation coherent natural and unhurried with enough frames to be fluid.]
[Character's physique/clothes/face/props: Completely true accurate to reference image without any creativity freedom]
[Artstyle: Completely true accurate to reference image without any creativity freedom]
[Camera/perspective/composition: Locked to reference image, no angle panning, no transitions.]
[Dialogue: Language is slightly Mexican girly accent English. In tsundere condescending natural tone she says " "My name is Teodora Vicenta de la Purísima Concepción de la Inmaculada Trinidad Villavicencio Duquesa de Oraverás, Marquesa del Jujuy y niña de la Condesa", she also adds some vocal distress interjections where coherent to this dialogue.]
[Scene: While talking she first watches her nails in a girly way then closes her eyes while flicking her hair then opens her eyes narrowed and takes her turquoise iphone checking on it with her eyes not minding viewer. The market background has natural activity and movement and sounds.]
[Video Quality: Remastered 4K HD ultra high resolution, sharpest definition, denoised, unblurry, clean lines, High contrast, fix any jpg artifacts.]
[Fixers: Never change the character's clothing or face from reference image, keep them true to reference image.]"
Why all that, because without that it was doing this
Now you know.
At first I only added scene and the pointers
People have indeed been complaining about random angle and character changes, but it can sure be handled
Also yeah that long af name should prove it can do the voicing well.
Also it seems it's better to encase things, specially the scene, that way it won't mix things all over the place.
And gotta put the scene in the most likely correct order, requieres cooking your brain tho, it hurts man
Also don't you love how it could do the Mexican-american accent, that's awesome really. Can totally be intentional.
It's all in how americans pronounce the "R" letter. Mexicans do it whithout reverbering.
Also the "Camera/perspective/composition: " does lock the screen or moves it
Might move that to tips when I polish it more, 60 credits a month hurt man
Made my first Flow Video
Man that's too neat
Also I don't think Flow knows what "Interjection vocal distress sounds" are, other AI video generators don't seem to have a prob with it. Things alike "hmph" so it seems those need to be prompted directly.
Really?
It was my first one
I just generated two videos
And combined them
Added a lil sound
Well when there are no weird things going on the vid can be called good, and that looks natural
I saw a quick tutorial on Instagram, though of trying it 😁
Got Prompts from Gemini
Did you prompt for it to transition scenes? Flow seems to do that sometimes by itself
It was in the prompt
I am gonna try the transition of building a stat wars ship now
Let's see how it goes
Does flow avoid generating copyrighted stuff?
More often, but I've seen if you're ambiguous enough about what it is, the model might still know and generate it
I see
Like I was trying with Zootopia stuff the other day and said "That recently very popular movie with talking animals" and it knew what it is more often, cause it sure wouldn't accept the words "Disney" or "Zootopia"
The models do have some of sort of built in search function, similar to what Google can do.
Not as regularly updated tho, but I'd say it knows things from around a month ago
When do the credits reset ?
Ohh
I used too much 😭
Used 4 frames
It seems to vary with account and location, for some people they reset daily, for some once a month, and for some never do until you pay and that one happens to me sadly.
You'll know when you do
Ohh
Is paste (ctrl +v) intentionally blocked in the prompt area?
congrats demon cat lol
Thanks
A realistic interior construction installation scene filmed from a completely fixed camera angle on a tripod. The camera does not move at all.
The video begins with a completely empty modern living room with wooden floor, white walls, ceiling LED strip lighting and a large window with curtains.
Five professional interior installers (age 25–30, medium athletic body build) wearing grey work pants, black t-shirts and brown safety shoes enter the room carrying wooden panels, shelves and tools.
They begin installing a modern vertical wooden slat wall panel system on the large wall. Two workers measure and align the panels while another uses a drill machine to fix the vertical wooden slats to the wall. Another worker installs the built-in shelf structure on both sides of the wall. One worker installs warm LED strip lights inside the shelves.
During the installation, wooden wall panels gradually cover the wall exactly like the reference image. The shelves become fully built and the warm LED lighting turns on.
The final frame reveals a completed modern wooden slat feature wall with built-in shelves and warm lighting.
Audio: only realistic construction sounds such as drilling, wood fixing, tool movement and footsteps. No background music.
Style: ultra realistic interior construction, cinematic lighting, natural movements, highly detailed, indistinguishable from real video.
/Videos
Hi! Using Flow for first time with google pro plan. Why my video has VEO watermark?
Just wanted to do a loop of this thumbnail
the thing that i want for a veo3 video is a smartphone where inside there's a conversation between an AI chatbot and a user.
use nano banana to create the start image
edit until its good, then use as start frame
It seems like editing images, like making comics/added changes make the image quality really blurry and smudgy. any way to stop this?
Anyone who knows this and can solve my problem it will be very helpful 🙏 Thank you
well i need a right prompts. i think just simply create image wont really fix this
can u elaborate, is it a screen recording, over the shoulder video, etc
nano banana can make very precise things but for veo3 limit to zoom ins etc
you see I'm trying to create a scenario where the human is chatting with the AI chabot named Jarvis and the user asked this AI to create video scripts, video thumbnails and does a deep seo research
and these all happening in a smartphone with conversations
plus the texts in the chats are all gibberish instead of right words
yeah can you replicate what i said here using nanobanana?
Hi, Is it just me, or is there a degradation in the quality of the images generated with nano banana 2, making them look pixelated...? 🤔
You sure you used the image download buttons and didn't just right-click save the image as? Cause if you do that its only previews, I think in jpg even.
Is it like a known issue that the ingredients archive doesnt work properly? If I select most recent, it gives me images I used 2 months ago.
It looks like you don't have access to Flow. Check availability here and reach out to support for issues.
Hey guys, I have a Pro account and acnt use Flow
I’m reaching out because I've encountered an issue with the video outputs in FLOW. I am currently on a Google Ultra subscription, which should provide watermark-free video generation.
Everything was working fine until last week, but now every video I generate via Veo includes a watermark.
yo, u do that actually in flow? its looks like soo cool
Flow should have a stop generating button that we couls stop the generation in the first 5 seconds at least when we make a mistake...
That cookie looks great... would you share the prompt?
Has the "extender prompt" option been removed for veo 3.1?
@fierce nimbus hey Nick, can we have Styles on Flow already? is it the Ingredients Archive? something like that?
Which model is best for character consistency?
Theres a bug where flow damages certain PNGs
I literally don't know why
It adds some colors, some light corruption or idk
Its consistent, I uploaded the same png multiple times and the damage or modification is consistent
The same exact thing happened, like theres some colored box or some discoloration
Is there any new information regarding the problem of
Upscaling Failed! Google Ultra 4K?
Is there any new information regarding
Does flow allow unlimited generations like Whisk did?
Google, Update VEO please...Seedance and Kling are better right now...I really want to make my ultra subscription worthwhile
hey, now there is a limitation to creating images on nano banana pro on flow? I created about 10 images and now appear this error, I have the Plus version
I am also getting the same error as of yesterday. It does not even allow a single generation now unless I use Nano Banana 2.
do you mean for free? no paid Pro or Ultra ?
Also, I recommend downloading anything you want to keep, Flow has lost all of my images and videos from the past 2 weeks all of the sudden.
I am on Gemini Plus
@austere heart I’ve just posted in the labs-bugs-feedback section. I’d appreciate it if you could add your comments to help bring more attention to this issue
hey, now there is a limitation to
Short feedback needed on this scene. I'm trying to generate the scene from the point of view of the man, but it keeps doing the wrong thing.
Should I cut the idea also?
First-person is best described as just the camera, not what the camera is representing
Yes
That over the shoulder shot is pretty awesome NGL
it just became call of duty
Thanks
Try two hands.
But he's supposed to discover something more disappointing, an animal not well taken care of
Oh mybad I didnt get the script in time.
If I just can get the first-person into the container right, just snow
Then I can add in the animal myself and stuff
Just keep tinkering away figuring out how to dictate your vision. I'm sure you're closer than you think 😄
Yeah, Tobes you got this man!
Btw, I like talking in this server a lot. Good to see the creativity flying around (as long as its not another dinosaur)
@dry frost@frosty niche thanks guys 
I could be a dinosaur. :3
Yes blaster master
I get that from some of the other servers I'm a part of 
I genuinely kept that image in the project for the sake of this Discord server's tweakers
The subtitle is so strange on this one
"Dinosaur go rawr"
Somebody needs some milk should animate that Google Dino game as a remaster and go viral
Prob already done but yeah, just crossed my mind. 🆗 back to work
This is kind of interesting i asked & went for a 5th person perspective & this is what i got.
A storyboard basically..
Makes sense. One of the concepts for 5th dimension would be a line where you could see yourself in the future and past at the same time
I could be wrong but I smell a sense of the mind that is supposed to be the 5th person perspective
Factz Dinosaur, Factz.
Dangit toby, I got to get back to work.
So who gon' add the Yeti to those
That'd probably be as close as humans could understand how the fifth dimension works from a picture's perspective. We just can't only with our eyes really.
what happened to Veo3? I can't get anything out of it since the update. Characters are inconsistent even with seed images. Does anyone have any idea what I can do now? Is there a way to get the previous version?
why can i only make 9 second videos usinf flow?
Hello
Does anyone know if it's possible in Flow to split a project into two separate projects? Or simply move media from one project to another?
I got you fam trying to hit the 6th perspective before midnight.
Wish me well!
Scroll down to the bottom at the "Review our privacy policy" window and the continue button will enable. If it doesn't for some reason clear the site's cookies and try again.
Kk
not yet
yooo guys
can any1 let me know the exact way to create these ai brainrot videos using google flow and this object talk gpt ?
HELP! = I deleted an image by mistake from a collection, is there any way I can retrieve it?
How do I get it so it does not generate any background music?
Why flow is kind of dumb when it comes to texts? I even put it on the prompt that spell (the word) correctly but then still butchers it
the ai can't even spell Operating System. instead it use Operring lol
wasted my credits
i think there might be a problem with google flow rn, since whisk migration to flow
What do you guys think? what are you currently experiencing with the platform?
Yo you cooked while I was asleep lol
I did, I used the button, and I used the same prompt I´ve been using since 2 weeks ago, but suddenly I saw a quality degradation....
and I have tested another paid platform and got the same result.
having the option to place camera in a particular place and rotation would be pretty cool
its really hard to get the angle right of the images no matter how hard I try
its gets it very wrong even with reference images and very simple clear prompts
Can someone tell me if we'll have more aspect ratios for creating images? Like 1:1 or 4:5?
How can I complaint about something on flow... is eating all my credits and I can't make it work... I have done a video, in Portuguese PT, the first shot went very well but when I do extend it never speaks Portuguese PT again, it changes to Portuguese BR (which is not the same) and doesn't get out of there. I tried to make a new video form last frame inside the same project but still BR.
I'm spending all my credits on this... how can I work this out... why, if it makes one well doesn't make the others???
For real, you'd think they would've covered this a while ago.
We cooking all weekend toby. Buckle up.
Don't do an Extend, use the last frame of your first clip, as a seed or Start Frame on your 2nd. I have not had consistent results with Extend in Flow. So, I just start a new project each time. Its a small pain, but worth it.
I did that as well, it's on my post...
Sorry, mis-read your original post. Sounds like the AI is hallucinating. Not sure what else to say?
Thank you anyway... this is a google issue, not prompting, not first/last frame... Portuguese PT and Poeruguese BR are similar but there's a whole lot more of BR speakers. That shouldn't justify the models switching autonomously to the BR Portuguese but that's wjat happen
I agree with you
Thanks! I'll keep up the energy to develop my Flow film
Short film tbh
"No ifs, ands, or butts, only musts."
convert video
Good evening, could someone explain to me what to do when this icon appears that prevents the creation of projects? I apologize for the English, I'm using a translator.
how to make flow generate 1 min video?
why am i getting a video when i have the ultra plan
its horrible. so many gibberish and mispellings despite a proper prompt. Waste of money man. Sora might be better when it comes to accuracy
problema de dublagem? pq não faz o video sem audio e depois dubla ele com outra IA?
Is there a way to maintain audio/music consistency when using the 'Extend' tool? I want the new segment to have the exact same musical vibe as the first part of the video. How do I achieve this?
Have you tried connecting your prompts with specific sequencing? It helps when i do it.
The last frame is a perfect example of non consistency, this 1 is a bit better, same should be for videos as well.
why am i getting a VEO watermark on videos when i have the ultra plan?
Is there a way in flow to reduce the credit usage by reducing the length of the generated video?
Hello from the Flow team! No, generations are currently all the same length.
Hey Mo, Ultra subscribers won't have a visible watermark on their videos unless it's required by local laws; you may be in a country where it's required.
when will we get access to labs in eu officially?
Hey Louie, which products are you referring to? Flow is officially available across the EU: https://support.google.com/flow/answer/16353544
We’re working to bring Flow to as many people as possible. See below for a list of supported languages, as well as countries and territories where Flow is currently available to users ages 18+. Suppo
Is it possible to assign multiple promoters during the review phase? In the previous version, before the update, it was possible 🙁@tough yacht
stitch particularly, uk, thanks
just realised im in wrong space
Ah, no prob, here's the Stitch channel: https://discord.com/channels/1172568727942860810/1374414152721895584
Thank you! I just wanted 3 - 4 second clips that's why I checked. Are there any future plans for a feature like that? 🙂
I had tried before, but thanks to your suggestion I put more effort into it. I uploaded the video file to Gemini and asked for detailed descriptions, paying special attention to the sound, and for a continuation of the next sequence. Even though I was aiming for a slightly different effect, the result turned out great better than I expected. The transition is smooth, and the audio still maintains the atmosphere I was aiming for.
Yes indeed, we'll share more when we know the timing.
Hello everyone can anyone help me set up a seedance 2.0 account it's urgent apparently it's been banned in us I don't mind compassion
today i learned that you can upscale extended clip
also another today i learned- ..
Hi @tough yacht , can you clarify my question? How do I continue projects after that red icon appears?
Red icon means you’re out of ai credit
Just top up your ai credit
Okay, I subscribed to the Plus plan. Can you explain how this plan works? Regarding AI credits... how many videos can I create with this Plus plan?
Idk about plus plan but
Without a subscription, you get a one-time credit of 100 credits and 50 free credits daily to try Flow.
Google AI Pro subscribers receive 1,000 AI Credits monthly, while
Google AI Ultra subscribers receive 25,000 AI Credits monthly
The 50 free credits daily are only for non subscriber
I understand. Do the plans also receive daily credits?
Unfortunately nope
the Google AI Plus plan gives you 200 AI credits each month. How many videos you can make depends on the model and quality you pick:
- Veo 3.1 Fast: 20 credits per video (up to 10 videos)
- Veo 3.1 Quality: 100 credits per video (up to 2 videos)
- Video Edits: 20 credits each
So, you can create between 2 to 20 videos a month with your Plus plan
I understand. Thank you for your help!
If you haven’t or never subscribe to pro at all, I can give you 4 month of pro for free
@limpid ice
Actually, I just signed up for Plus through my account, could I use this offer on another account?
If your other account has never subscribed to the pro plan, then yes, you can
Okay. Can you send me the offer privately?
Sure I’ll dm you
I see your prompting skills are coming alive, he'll yeah.
It feels like edge computing when your trying to get everything just right, got to ❤️ it. 🙌
Hmmm, spray can or laser beam? Hmmm
How do I move the girl and the entire store shelf down by 100 pixels? I've tried many prompts but it didn't work.
Like this?
I copied your question and input it in prompt box 
Is it just me, or has Banana 2 been having a harder time capturing the characters' styles and understanding prompts for the past two days?
It's as if they made Banana 2 even worse. It used to be on par with banana pro, but now it's noticeably worse.
Yeah the consistency seems like it got nerfed a bit imo.
Maybe the transition from whisk is causing this, either way its noticable.
This has been especially noticeable over the past two days. I'm going to run some tests now to compare the older versions of Banana 2 with the newer ones. It's basically become just like Banana 1.
I was testing this morning and the bg, just unraveled automatically on me as well. I suppose more prompting. I know theres some prompt libraries that released, I'll dabble and make my own framework later on for it.
That's too bad. I needed to replace the faces of the characters in the illustrations with different ones—the ones in the attached files. Banana 2 couldn't handle it, even though it had handled the same task perfectly fine before...
Yeah i know whisk is sunsetting so their probably bursting some pipes in the meanwhile, is my context at least. Hopefully they do it quickly & professionally of course.
Yeah. It's just at banana level 1 right now...
At least we spotted it together, now how about that bug bounty. Oh wait were still stuck in 2006 with that process. Fascinating.
I think it would be a good idea to award at least 1,000 credits to accounts for identifying significant defects in models. But I’ve never heard of a reward system in Flow. So 2006 is still relevant.
Did u try in whisk
Now were talking and actually implementing some sense/cents. I mean c'mon its overdue. But hey lets prop up some more ads in the meanwhile. Gotta love it. They probably never introduced 1 imo, not yet, the idea probably just came up now for all we know.
.
I don't like Whisk; Banana 1 is already outdated and doesn't work well with character placement in the frame.
Is it me the only person loking for a "Add to Collection' button on the images and also missing the "Add to Prompt" on the images in the multi edit editor
Did u try using nano banana pro
Instead of nano banana 2
Is that stick of chewing gum talking?
I'm more of a big red guy.
yes
So anyway
Whats on your mind?
That's exactly what I'm doing right now. Banana Pro handles my work perfectly.
banana 2 isn’t really ideal for doing precise editing like that , try the nb pro
It used to capture the style of the source material very well. But something has gone wrong with it over the last two days, and it has become less accurate. It has started to look like a banana 1
I’ve noticed a decline in user prompt adherence lately too
Yeah, same
Look into prompt libraries, I'm going to start working on mine later in the week. I've been seeing alot of frameworks being released in that area as of late.
On brand for sure.
I was thinking more of designer prompt templates based on a users storyboard possibly, but yeah on par towards that direction. fs.
I Don’t think this part applies to nano banana in Flow right?
Can confirm it does not use real time web
Hmm, interesting.
I'm over here looking at this instead. lol
I saw the example in product hunt acouple of weeks ago.
Very helpful.
...they seem to be picking up on the feedback about keeping Nano Banana 1, 2 and Pro as separate models at Flow, but instead of keeping them as separate models they're adding it all to Nano Banana 2, and that's just making things way worse.
Reminds me of COD.
hopefully not
Need keyframes for that.
It probably doesn't know the difference.
shady dealings
is is me or 4k looks all pixelated or over sharpened?
Was happening to me the past 24 hours. On & off. not just you.
:(( it made it unusable and it was a good gen
When generating videos with Flow, a Veo watermark appears even though I have an Ultra subscription. What should I do?
watermark --> thanks google you now have watermark on your videos!!!
congratulations!
You paid for ultra?
Hello everyone, I’m new here. I’m trying to make my model move her hair while dancing. She is my digital model, but Flow keeps insisting that this violates the rules, blah blah blah… Why??? Can someone explain what’s wrong with the prompt?
anyone got any tiktok videos for me
It doesn’t continue the video, even when I ask it to suggest something. I’m forced to finish the video at 8 seconds because I can’t extend it the AI just went crazy.
NanoBanana is over-censored garbage for making realistic photos of women. The slightest exposure of cleavage gets censored. Don't waste your time with it. Just to give you an idea, is considering this costume NSFW. 😂
Thanks but it not exactly 100 pixel, hard to control position
Is a costume, mate. And the exposure is nothing more than what you see when you go to the beach.
Maybe we should forbid people going to the beach too. Google logic 
Argh, yeah umm just be straight up ask for nudity if that what your trying to unsuspiciously(cough cough).
😂
do.
.
I could've sworn I got a 🔵 - Chew add as soon as you came through.
Straight 🧢
✌️
ENGLISH, mate. I don't understand what you are saying. And also that gives away your age, which makes me have no interest in discussing any further. I made my point clearly above. You either can't read, can't see or cannot understand. I am unable to help with any of the three.
Nano Banana could actually do it no prob. Nano Banana 2 is the one that mostly has no idea about female form, wouldn't even call it censorship, it just has no idea on how to do it as if it was some sort of incel, it's awful at it and no amount of feedback is going to fix it.
The only possible fix would be to re-enable Whisk’s Nano Banana model in Flow, like it was before.
The idea that there’s something wrong with the female form and that it should be flattened is mostly a Western misconception, and it really shouldn’t apply to the internet.
In fact, denying a woman of the natural shape of her body is arguably a form of discrimination. Treating normal feminine features as something that must be reduced, hidden, or flattened implies that those traits are inherently inappropriate, which reflects a biased standard.
And all this is painfully visibly being applied to Flow's Nano 2.
And tbh, it's just one more of the the things that pretty much has destroyed the Nano 2 model into being unusable, like many people have said happens to it.
Go on.
if you keep editing images in flow images, like making comics/added changes make the image quality really blurry and smudgy. any way to stop this?
Try always ending your prompts with stuff like
Image quality: Accurate reference image but remastered 4K HD ultra high resolution, sharpest definition, denoised, unblurry, no jpg artifacts, clean lines, High contrast.
Or rerun that after the image was done, but like if the faces started changing too much, yeah you're gonna have to fix it in extra steps like reference images and sort.
I'll try that thanks. This can't be intended right? it's sooooo bad
Yeah Nano models always degrade images a little after each edit, that's not perfect but should keep it well enough, it does know what "Remaster" is.
Anyway talk is cheap so let's have another little test atm.
And here's Whisk, giving the exact same quality and kind of image of that model at Google Ai studio, proving it's Nano 1/Gemini 2.5 Flash image.
Good morning house please, I have serious issue here and I don't know if something can be done. I registered for ultra and decided to share the login to some people but someone betrayed and change my email password and put number and 2 authentication. Can something be done?
You better message @real wyvern with all the issue and your data, maybe they can do something by purchase history and location.
Going back to this, gotta love the little message at the top of the Whisk website, promising that same website will be to be moving to Flow. Ok let's see what Flow is doing with that prompt
Ehh gotta admit it's mostly ok, but I had to cloth the character up because it WASN'T frigging allowing the image.
It also did visibly unnaturally shorten and thicken the character's proportions up in some images, in others got 3 arms, which happens I guess.
It might not be so visible at first, but Whisk's images did keep the character tall and slim.
So again, Google AI studio and Whisk, it's easy to know cause Google AI studio has the spark watermark. They're the same Nano Banana model same picture.
And best image I got from Flow's Nano Banana 2, couldn't get it at 2K cause it kept saying "Upscaling failed".
Being fair it's not ugly, but it sure is a different thing.
Just from there it's easy to check the scale ratios, Nano 2's woman is shorter with slightly wider waist.
even 1080p does big bro
please explain saar
.
And yeah here's the screen with Google AI studio using original Nano Banana/Gemini 2.5 flash image, proving that's also Whisk's precise mode. As it is side by side there.
Nano 2 and Pro are good, but they are not Whisk's model.
Why doesnt it work
Wow, what trash. That AI is really lame then. It really seems like they actually created artificial stupidity. I added the cleavage because it’s part of my character. So will I have to abandon this tool?
Decent start for character development.
Even though theres alot of characters in here, that dont have any.
Whats going on with Nano Banana 2
can anyone help me i want to delete a particular extention from my video because it is inacurate
Heads up: video generation might be having an issue. When creating around 5–10–20 videos, it sometimes returns "Sorry, an error occurred".
From the logs, the system response shows "Media not found". I'm currently investigating.
He'll yeah hunt that baddie down @arctic kelp 🐛
#2006
@tough yachtHi, could you please check the message I sent you? It might be related to the "Media not found" error in video generation. Thanks 🙏
Hello. Can I ask, when I create some image, can other user of Flow/Whisk see my created images? But by other way then post to flow-gallery section? In Whisk was some code, or what. Is this option in Flow too? Thanks for help.
@lyric deltaHey, just a heads-up: there’s no way to generate images or videos on this server. You’ll need to choose one of the Labs experiments, li
I love that sometimes when my base image for i2v is not good enough, using the base image as ingrendient makes a different but very similar image and i work with that
Thank you! I'll respond to your DM and investigate from there.
hi @pulsar coral can I dm you by anychance regarding your flow workflow?
Is there a new problem preventing the downloading of 4K videos?
have you tried generating a few and see if one goes through or they all fail?
Whisk model is just nano banana 1
THIS!! 💯
But for real we need the Veo 3.1 Ingredients Quality model 😭
Do we have any news of when it will be possible transfer projects from Whisky to Flow?
cant wait to see what veo4 will have to offer
is 4k ver native or upscaled?
Hmm, is flow multiple accounts accessible yet? I havent checked but this is a good question!
someone created this amazing video with nano banana. insane..
1M views in just 4 days
How do you guys deal with terrible dialogue sync
It mixes and changes the voices between characters
It's possible to cut and edit the video to zoom in on the character we want to speak, but there should be a more direct method, please let us know if you find it
replace it
I made a google doc helping clarify some experiences you may be experiencing:
https://docs.google.com/document/d/1b4HdSLFVl0DrXyfYfKS6YIxpsFeN6soxkLWT5hocYNY/edit?usp=sharing
Heres the main area of the doc, hopefully it helps.
FYI, not a bot btw.
Do you think the Veo watermark might be removed again for Ultra subscribers in the future, or is it now required permanently in some countries?"
Why my Daily credits aren't refreshing it's been 50 for the past 5 days and i need to use flow to make videos
It is now required indefinitely by some countries, so there's no end date as far as I know.
Has the Bugs/Suggestions thread been nerfed?
Do you have a list of countries where VEO watermark is required and countries where there is no watermark? On Ultra plan
Where is this at?
I got an assortment of them, their all over the place.
Weird. This isn’t on my list of threads.
(cough cough) 2006.
Hi, I'm an AI Ultra subscriber from South Korea. I'm having a watermark issue specifically in Google Labs Flow. When I generate videos, a 'veo' watermark appears. But interestingly, when I generate images in Flow or use Gemini for videos/images, there is NO watermark. It's only happening in Flow video generation. Could you check if my Ultra subscription is properly synced with the Flow engine?
it happens for me too
unfortunately
@dusty pivot this is what google staff said about your issue
Google must remove VEO watermark from Ultra plan.
Its work with me
@tough yacht I sent a message to you , please check your inbox
Please, all users who have the Ultra plan on Flow, can you tell me: who does NOT have the VEO watermark on their videos, and what country are you from?
me. im from malaysia
hi, im ultra subscriber from philippines and after i subscribe, this happens, i can't generate from my flow, can someone know what to do?
unlucky
is ur acc part of workspace acc ?
was just gonna start using flow then that happen
huh
because based on ur screenshot
it shows upgrade button
and "credits refresh daily"
usually havin this two means your acc is not subbed to any plan
becuz credit refresh daily is only for account with no subscription at all
but somehow u have 25k credits
I don't know what happen to it and i already paid for it
ok thank you
see ur country even included in supported countries
are u 18+
yeah i check it too
yes
are u connected to any vpn?
no i didn't use any vpn
ok wait
Flow is an AI filmmaking tool that lets you seamlessly create cinematic clips and scenes using Google DeepMind's most capable generative video model, Veo
it's the same as my pc
ok uhh if u go https://labs.google/fx/tools/whisk/
A new experimental tool that lets you use images as prompts to visualize your ideas and tell your story.
does it show that youre on ultra plan on top right in ur screen
yah it's really weird
at the app, it says refresh per day but upon clicking the credits, it refresh per month
Try another Google account, fastest way to confirm rollout limitation. If another account works it’s not country related at all
Just to be sure because lately Google's docs say something but then they do something different. We should know around here.
Also even in supported countries like the Philippines, only some accounts are enabled. Even if a country is “supported,” it doesn’t guarantee your account has access yet, Google rolls these out gradually.
yah already try my other gmail that was not subscribe and it's working, but this subscribe one was not working
It might also be some cache issue, try opening from some incognito window.
Tho like the screenshot says there, you better not have payed from an account stated as less than 18 years old, cause they don't seem to allow those.
it's still the same
Also the mail can't be something alike "@university.edu" as in from school or like wrigley said can't be workspace alike "@business.org" "@company.com"
Cuz of this
Yeah and if all that can be scratched out more likely you have a rollout issue and need to wait till the account is enabled
This is very common with new Google tools
Hey! I'm having the same issue as @lucid tide, I could access Flow before subscribing to Google ultra, then now I can't access Flow anymore
it was 1999
Either way you two better just message @real wyvern and say something like “I purchased Ultra plan, credits are visible, but I still get ‘no access’.” and give it your mails, it could take few minutes to few hours but they'd fix it.
It's surely some rollout issue or bug in the other bro's case.
They can fix the issue manually.
Questionf or experts here: How to ensure consistency in character placement in a location across multiple images? Example: Character A & B in specific location. Camera position is defined. Now unless character position changes, these location must be set and not deviate
modmail is only for discord related
you need to contact google one support team
Yeah but staff is there
Then again they better just also use the built in functions
There we go
Mine fixed now, I contacted One Google support, I'm not sure if they're the ones who fixed mine
idk it does also sometimes fix by itself.
yah mine was fix already too
thank you
I think we're being watched
is the option gone, that I can eject out a frame from a gen in the scene editor????
Flow AI keep changing the image style i gave him a cartoon image, in the video, The style became Realistic with real people can someone help
Is nanobanana2pro next?
Try giving the model strict orders clear pointers and visual cues as to how to recognize what you want it to do and the image you want it from. It all depends on the uploaded images, let's say I'd upload these two then the prompt could be something like
[Art style for the whole video: Consistently plaster the style from the second uploaded image with the anime girls with horns onto the first uploaded image of the blonde girl with shorts.]
if it needs specific cartoon style you'd need to be even more precise and just swap the reference image into the cartoon style you want, that'd help it more, the "plaster" prompt helps a lot on that too.
... keeping the character stable in that pose and same clothing is a whole other different soup tho, but it's surely doable.
What is happening here? I'm a pro user and have been using flow to edit my own images for a while now and never saw this
When is flow/veo going to do videos longer than 8 seconds?
All good. I managed to do it. After a lot of frustration, and by using other generators as well. This is the 'shocking' NSFW content that it couldn't do.
Sorry for the thighs and cleavage. Please look the other way if you don't to end up in hell. 😂
Ah she indeed was an ice character. Looks good.
T&A or huuhhhhh
😂 Looks good man.
Tysmm
@swift robin I have Same Problem
please use #1238621511401734254 to submit any bugs you find!!
Good question
I'm getting requesting too quickly even after I last used it yesterday.
Same
Is this problem only with me?
Yeah, since the latest update.
That's what extend is for.
can i only extend on the last frame?.. I would like o extend after 5 seconds
Extend??
how can i get this view
i dont see this
scene builder possible to upscale?
the aspect ratios in flow has finally come
very nice
yeah
new update today
whats the diff between ingredients and frames
does ingreddients just put the picture at some point in the video vs frame puts it either at the beginning or end
Someone have like that problem give me answer
You're using Wifi, more likely it's just taking time to load, you might wanna revise your connection or better yet use a device with ethernet.
When you open the Flow page for the first time in the day, it does take some time to load thumbnails until it saves it on cache, even with decent connections.
I have a question, I used the new collection feature and made one with 14 pictures in it. But when I click on download collection I only have a zip with 3 files. What might be the problem there ? Of course I can download manually but when collections getting larger this download thing would be highly appreciated
no
I knew they could do it.
Go on...
... you did seriously just remove all the new aspect ratios. I hope you're kidding, I was already using that.
No man bring it back what the hell
Give it a little longer, maybe hes moving stuff around. I'm sure he'll get back to you when they finished.
Yeah did come to mind. However.
it was working already!
Tho I imagine it's just an issue with videos not being ready for that type of aspect ratio yet.
.... Flow is before anything else, a video generation site.
That or like I'm saying they're just pulling legs here.
Meanwhile the Flow browser window I still have open still has the aspect ratios and they work. Now all we need to do is keep this PC working till the next Russia is founded.
If the fridge has been running since the last time that happened, the PC can do it too.
i think i understand why u guys removed em for now-
i was trying to use a 4 : 3 ingredient for video
