#flow-general・🌊
1 messages · Page 55 of 1
If you got advice to be able to do that kind of simple modif in Gemini instead of Flow or point me to where those are, would appreciate.
I generate hundreds a day so going through Gemini is a bit 'difficult'
I reported the DDoS in the bugs room
It does generate it but...You end up having to edit the gemini star out and then resize it too - I'd prefer a Flow that works 🤷♂️
Someone spilled beer on the server.
Agree
Gemini can generate any reasonable aspect ratio by prompt, alike end your prompt with "Aspect ratio is 9:21" and it will do it. Other problem is it leaves a spark watermark so be sure you use the download button on the image not right click save as, that will give you 2K and will allow you to use this watermark remover without losing quality. https://geminiwatermark.org/batch-watermark-remover
Google Pro user here facing this issue on Veo 3 generation
yes .... DDoS ... I reported it
specifically referring to video generations
nano banana is working fine
referring to video generations only
I used to generate like a hundred images a day in Gemini before they busted the NB2/Pro models not just by copyright and censorship, they also busted the quality of the models with all the probs people are reporting. I had to downgrade to a Gemini canvas with NB1 to do anything now.
video generations
Thanks for the tips, friend! Appreciate!
Like 21:9 aspect ratio is a good size for wallpapers, Flow doesn't give it and NB2/Pro are perfectly able to do it by prompt, but they haven't enabled it for god knows what reason
People sell wallpapers on the internet, whodathunk it.
havin this since 3 hours no matter what i do or how often i reload it wont work....even restarting wifi, changing IP or changing prompts nothing works....whats going on?
same since 3h
Hello my fellow command prompt engineers, I've identified this error: sometimes the flow generates only one image, and sometimes it generates nothing, showing this error. Do you know what it could be?
folks ... there's an outage .... I can't say that enough lol
I reported it in the bugs room
Unfortunately, there has not been any official announces tho, right? Or I missed it?
I would just log it in the #1238621511401734254 channel & wait for the storm to pass.
not that I know it .... terrible on their part! When we have an outage at my cloud software company, we notify pretty quickly
My mistake is inclusive in the generation of images.
Thx
This is more or less what it'd be supposed to look like at Flow, including the aspect ratio selection for auto and other ratios, selection for the Nano Banana (Gemini 2.5 flash Image model) like Google AI studio enforces since the start but Flow disabled who knows why.... I'd like to say Google AI studio is a nice alternative but it's is very annoying to use in comparison to what Whisk was or Flow could be. And yeah I've suggested this stuff at labs feedback too.
This is a "special" advertising campaign from Google. ))
It'd probably cause less servers probs too considering how Whisk used to work with that NB1 model without probs locked at 1K resolutions
https://tenor.com/view/cat-loading-error-gif-19814836
VEO 3 on Flow be like:
... you'd think Google has probs with how NB1 is less censored, but every other Google platform still has it, even Gemini on the canvas/gem generators. It's just bad managament.
Anyway, let's get to some bugs.
Hmm yeah Flow is still locked at 8 image generations at once, nothing has changed since last week. Weird.
Allright, to all those getting the "Unusual activity error", this happens because you're generating images/videos too fast.
The limit right now is 8 generations at once per 2 minutes, as in it's better to wait till all your images/videos finish to generate or fail to do so and 2 minutes before requesting more.
If you go faster than that you'll get the "You're going too fast" error message and if you try to generate more sooner than that like twice then you'll get the "Unusual activity" error.
To fix this error you need to clear cookies, it can be every cookie or just every cookie related to labs and Google, and so on if it happens to you again.
If after that you still get the errors you do need to wait a couple minutes after clearing cookies. And if after 5/10 minutes trying you still get the "You're going too quick" message, that just means your account hit the daily limit for image/video generation attempts.
So in this case If you're generating images try swapping models from Nano Banana 2 to Pro and vice versa, you might have some more image attempts there. Otherwise yeah the account is dry for the day, it will reset not exactly after 24 hours but like 30 mins after you generated your last image/video the day prior whatever the hour was.
According to your tier plan from free to Ultra you get a set daily amount of image/video generation attempts, no matter if you still have credits or not or if you generated an image or not, the failed attempts count towards the limit. It's around 200 attempts for free then double that for each tier, I think Ultra has many more than that.
So taking all this in mind, once you start your worflow you might wanna be sure you used all your image generations for the day as best/much as possible in both models so you can start your workflow around the same hour next day, it's a routine at this point.
its not that bud, its literally an outtage
Well in some regions it might be, the "Unusual activity" one does happen in other regions where there's no outage.
I know cause I'm in south american and it's working fine here.
Maybe that's why they don't care much, Google might not be having outages at NA either.
Anyway if your region is suffering outages yeah you might wanna inform at labs-bugs-feedback. You can know that according to error message, the "your generation is taking too long" error is probably related to that one.
Cant be because i generate 1 per 10min and still doesnt work even spreading out on different accounts….
What error exactly are you getting
Yup hes got a point, I only let 1 maybe 2 run wild, 8 is wild.
yeah 8 is just the current limit, prob is not waiting some 2 minutes after all are done.
i‘ve waited 2h and tried again and notjing works
If you're getting the "You're requesting generations too quickly" error then your account might have just hit a daily limit. It all depends on the type of error.
Cause then again yeah it could be something else.
i same
its not that error
Failed
Generation may take longer than expected. Please wait a moment and then try again. Failed generations will not be charged to you.
And i tried doing everything
Waiting for 2h trying again
Nothing works
Same here
yeah this one indeed could be an outage error in your region, you do need to inform this one at labs-bugs-feedback. Idk I'm spamming images atm and it works where I am.
Is all this stuff because of Google I/O? System strain or something?
It's different from america
Latam and Brazil are the same. Lots of people around here are dealing with this too :/
Tho yeah that "Generation may take longer than expected" error also happens sometimes due to a connection prob, might wanna free all your devices resources and reset your router so you get as much bandwidth as you can. Flow eats a lot of resources. Specially from mobile.
But if that doesn't fix it either and your connection is fine, yeah it's a prob on Google's side.
ok got the sauce
i remember same issue used to happen w nanobanana generation, basically try generating 1 of each - veo fast 3 s 6 s 8 s quality ect
Working for me from Ecuador... my connection is just average tbh so it should be fine.
until u get 1, and then u can go back to doing normal like lite veo
kind of shoots u up into prio is my guess but it works
Yeah more often all these are speed and cookie issues, one can learn to get used to it... cept for the connection probs
Note, the outage is video generations
Yeah on my side the image & the video gens are working now.
bro with all due respect its just an bug that happens nothing u can do except for sometimes gaming it
idk Flow can indeed be completely random about things. These are just the things we've been seeing all week.
Ah .... video generations might finally be in the clear ... just now started working
Yup it changes by the day really.
damn well maybe my sauce wasnt saucing after all, maybe servers are back 🤣
I love how when they have an outage, there's no announcement and they just quietly fix it behind the scenes after some time
yh everytime they barely notify in chat here anyways
now that video generations are working, we can get back to seeing the "prominent people" errors 🤣
idk with all this supposed censorship Im still doing what I want with NB2/Pro... the images are still default ugly compared to NB1 tho, can also be fixed by prompting but it's ridiculous the default is like this
noice
hi
We’re back online after a temporary outage impacted Veo generations.
thanks everyone for reporting this to the team in this channel and in #1238621511401734254 
This is a total disaster. No videos have been generated in any quality for a whole day now...
What error message are you getting
When experiencing these types of issues, providing the error message usually helps at times.
Errors like “Something went wrong, please try again later” or “The flow is overloaded, please try again later”
"Queued"
Never finishes, coins gone
Maybe there's still some server issues after all. If you're still experiencing them after like 30 mins you might wanna report it at labs-bugs-feedback like Georgia asked above. Might be better to also add your current region.
Because they're visibly on it.
there is a bug in the collection folders : when you go through differents collections you must refresh the actual location url of the collection you want to use to force the image generation inside the last collection you want to use .
hmm, interesting - can I ask you to file this in #1238621511401734254? any screenshots, additional info would be fantastic. (please do not share sensitive information on Discord)
Wish I would've seen this before posting my token key in collaboration
No idea why that’s happening
Anyone getting a "It looks like you don't have access to Flow. Check availability here and reach out to support for issues." screen.
Is it just me or does anyone else try to make a video or animate a picture and its stuck in queued and never works?
i have it 80% of all generations
Huh, odd
is flow veo 3 working now?
same
Half of my generations get stuck at 99% for several minutes now, wasn't doing this about an hour ago
failed oops something wrong. is there any system maintenance going on flow?
Google typically intentionally degrades older products before announcing new ones to create contrast.
But this can reach absurd levels: tools become practically unusable.
This practice is common throughout the company's ecosystem. Individual features simply stop working.
A prime example is voice input in Google Keyboard, which hasn't worked for about a year, and Google plans to unveil a new voice input system only now, at its next presentation.
The Google Flow interface is sorely missing a chat assistant.
Is Google ever planning to add TEAM subscription, just like heygen or kling??
I keep getting this and as you can see, I am not typing in anything against guidlines.
i want to have chat with you, can you send friend request ?
done
how can you use face images? i cant . the image generated through flow and same image if i use for next prompt as reference i get unusual activity or prominent errors. requesting too many images
oh is it that????
what?
its a server issue, not your prompt dont worry
Oh ok phew
may be it will fix after 19
I was like what did I do wrong now LOL
the only way is to just click on retry button
thats the i/o day right?
I tried and I keep getting an error....let me be more persistent
ok so i cant say much but, yall probably will get your own "agent assistant" in your flow
yes . I asked gemini about the flow situation and gemini told me that
trying consistently gives more unusual activity errors
i think flow team are prepping for i/o day
hey do you know any other website like flow where i can use nano banana 2 or pro ?
LOL it's hallucinating
every ai website ever
most of them are very restricted and limited.
what do you mean by restricted and limited if you mean like NSFW then yeah everybody is like that unless I guess if you have like a model on your own computer, but do you mean like blood and fighting?
blood, you just have to say something like there's red water on the floor, thick red water or thick, red paint and it usually does it pretty well
there's ways around it in other words... it's not completely closed off but NSFW for sure is a no
no not nsfw...lol... when gemini models starts generating nsfw?... So i am not asking for nsfw or edgy contents.
there was a post on Twitter, where the guy completely circumvented every model by just using really creative very specific terminology
you just have to have a very good handle of the English language. What I do is I use a thesaurus when I need extra words or you could always ask an llm
Oh you mean the stupid blocking....yes I hate that
like it blocked for no reason, even if it's not even a bad thing
seedance blocked my staircase video...literally a person walking down a staircase LOL
in FPV
did you use real person?
ohh
i only used seedance 2 one time normal car to modified car . because its too costly. dont know how people generating so many videos.
They want us to make movies but wont let us use characters...so we cant have consistency. It's pointless.
yeah and it's not that great
new gpt model is also very inconsistent if we use reference image.
but for text to image its good.
Really I didnt know that
it used to be really good. GPT1 was great at consistency...I dont know how they make models go backwards
they get worse and worse every time
but it was slow. gpt 2 is good this time. ( text to image.)
but currently for image to image and character consistency gemini models are good
hope they fix flow soon. and bring nb 3 asap😂
Man they have really screwed Flow up haven't they. Can't download things 90% of the time, can't generate things 90% of the time just constant errors.
It has always been a bit buggy, but has become completely unusable the last few days
I always see unusual activity on web. But it's ok on mobile flow app. Why?
I get the unusual activity error every like 5 or 6 generations, but today I can't upscale anything it just fails and it has now decided I can't generate anything either.
I can use it very well on Gemini, but not on Flow. I've given up on it because there are other better AI tools available.
if u mean u want the no restriction? in my experience, flow is the ... kinda less restrictive but if u are subbed to pro/ultra plan
u can fine tune nano banana 2, even got image search capabilities
and nano banana 1 is still availabe to use
Same for me since 2 days ago, I couldn't generate anything since Friday
Anyone with the "Queued" error?
Me
flow has any problem now? cannot generate anything
Pls notify me once it gets fixed
What is flow
I waited about 10 minutes for the first video to be generated, it worked, and for the rest of the videos it took about 2 minutes per video so at least it's working now
Can I DM you ?
yes
Queue error?
yes
Anyone having issues with slow generation image that keep getting stuck at 99% today
some times for me
Ugh, tjete was a time when updates made things better, but nowadays it's more likely to break something!
hello bro I want to talk to you about something
Hey
what have you found matchs the veo quality without the veo issues ?
Hey everyone! Just a friendly reminder to keep all conversations in the supported language for this server. This helps our entire community engage effectively, and we've gone ahead and deleted the recent posts that didn't comply. Thanks for your cooperation! @pearl vapor
pomelli
??
Hallo
Halo Georgia, is there any update for Indonesia region? I worried if I can't join
Hay daniel, did you have fill form about UXR's?
I'm not rich enough to sub to every website. I should have known about Flow long ago. I found out about the website last month, and this month it’s already buggy 😩. I always get the same feeling with these AI websites—that constant fear of missing out because it happens to me every time. Man, I always find out about things so late... especially when it comes to free stuff.
Uhhhh
yes i checked it.
If u never sub before
I can give u 3 month pro free trial
But I’m not sure if it’s still valid
Thanks a lot for the offer, bro! But I'll pass on it for now. Appreciate you checking though!
I will wait for new models for now. And search new websites in meantime
Nah still busy
Currently, there's no tool that can guarantee both the quality and the effect of Veo. I suggest you use MJ to create the characters and scenes, then use Nano Banana to create the three-view drawings of the characters, and finally use SD2.0 to create the complete video. Absolutely perfect! 💯
Se*dance 2.0 is the perfect Video Model.
seedance
why u censoring it like its a slur lmao @zenith current
SeedDance 2.0 , Midjourney
😭
I'm worried that Google won't allow these words and will kick me out.
thats on discord moderator lol not google, plus u not really breaking the #rules-and-guidelines・📋
i have to agree with u on this
🥲
Of course, using GPT Image 2 to create the three-view drawing of the figure would be better.
May everyone become a dreamer.
Also gotta agree with this tbh, GPT Image 2 is actually really good, especially with text rendering. Most image models still mess up typography pretty badly, but this one handles text/layouts surprisingly well
... yep Flow's NB2/Pro got so bugged after Google's little copyright and censorship stunt 'update' that many are migrating to GPT Image, mostly cause being out of options, that something at least works a bit is better than nothing. Even AI arena is confirming NB2/Pro being outperformed.
i was surprised how loose the restriction in aistudio when u generate using nano banana 2
Without the search It probably mostly reverts back to something alike Geminie 2.5 and imagen. But not everyone's gonna pay for Ultra to check.
Maybe later we could run that prompt I usually run, that alone would say it.
Flow doesn't even allow auto aspect ratio or ratio by prompt for the images, it ain't suprising it's lacking lots other options.
u can use pro tho
Ah man that sounds awesome, I had a small chance to try some of that just in time.
.
Knew it wasn't just my idea
are u not aware that aistudio inage generation is available for Pro And Ultra
Ah you mean the tiers, I thought you mean the Pro mode. Gemini has them named the same.
sir u need to stop confusing urself
idk what you mean by "gemini has named the same"
In fact, I have no reason to pay for a Google AI Premium because their performance has not met expectations.
Yeah some famous faces it can still recognize without search. NB1 could also do it some extent. But needs to be something like really popular.
thats nb1
Gemini calls "pro" mode the more juiced version of the models, That's why it doesn't allow to select NB Pro from the get go
sir
u need to start reading official documentation 😭
We facing generation issues everyday
dont spread misinformation 😭
All I see from you is you doing fear mongering because of your false assumption 😭
Jesus do I need to screen everything. Give me a sec.
Always Everyday for me 😭
i dont need ur screenshot
cuz ur screenshot is just u asking Gemini chat 😭
what u should be doing is go to official gemini documentation
I laughed, because it's ongoing for months bud.
Not just yesterday. You are not alone. Several bugs have been reported here.
I’m not gonna waste my energy trying to correct you cuz at the end you’re not even gonna listen to other people anyway 😭
U can use well on mobile app
Here you go. No one's arguing with you bro
When you select that, Gemini uses NB Pro
Once you run out of "Pro" mode, it swaps to NB2.
Just use it for a while and you'll see
It's another thing that also proves NB2 is just a nerfed cheaper version of NB Pro
But yeah the "Pro tier" account is a completely different thing I get it.
Not very convenient naming.
https://aistudio.google.com/prompts/new_chat?model=gemini-3.1-flash-image-preview > gemini.google.com (create image) / flow.google image generation tool
Want one more thing to prove it uses NB Pro with that?
It uses basically a photo for background.
NB Pro does that, unlike NB2 which would make it anime.
... which is just horrible tbh, that's why I don't like Pro
I tried to follow @pulsar coral and @crisp basalt , but I am lost what you both are trying to say 🙂
Can you use gemini to add your argument in bullet points 🙂
Is it a battle between NB2 and NB Pro
You see, back before NB2 was released, we only had NB Pro.
And Gemini also had that "pro" mode, which could be applied to NB1 instead.
Once you ran out of NB Pro images, again before NB2 was released.
The "Pro" feature would add more juice to NB1 instead... and it was DAMN awesome like you have no idea.
Flow had that for a couple days too, I did notice it in the images... then they added NB2 and disabled NB1.
For all intensive work, I have concluded that flow is a broken tool
But I can relate to what you are saying
So when Wrigley said Google AI studio had "Pro" I thought he was referring to that, so my bad there didn't think he was referring to the plan tiers.
... tho I dunno what "officialler" Gemini I was supposed to be trying.
It depends on the output you are aiming for.
NB2 Pro is the best model from google just now. They have the key note next week and we are likely to see new Image and Video models.
However the top tier model will come to pro and ultra users first. So it will be guarded by a pay wall.
We will get some free generations everyday and it will take a while if it hits Flow.
Need a Pro plan to process a joke? 😉
Using NB2 directly at Flow it's visible it rather draw things than make them look like stock photos realistic like NB Pro does.
To use Gemini I'd have to make sure all the "pro" mode photos are spent so I could use NB2
I was just joking about how serious you both were. Or atleast it came across like that to me 🙂
... or I can just go to Flow directly, or use Google AI studio like Wrigley does
I think we have more generations in Flow than AI Studio. Is it correct?
You give up quality to gain more generations. Flow shines at that, is not it?
For free yeah, but for free Google AI studio only gives you Nano Banana (Gemini 2.5 Flash Image). Which considering that's what Whisk used can be good or bad. More often good.
Supposedly your same Google account also links from Flow to Google Ai studio now, but I think it's only for Pro and Ultra plans. Haven't seen it working with plus yet but might be just me there.
Are you guys able to use Flow over there?
In past I had more success with Flow than AIStudio Free tier.
Most of my work was low-latency basic stuff... so quality and resolution was less important than volume.
However with recent unreliability with switching to AIStudio seems sensible.
Working on my end... in the way it usually 'works' but I did just generate stuff
I have used L**2 and that works well too if you have the hardware to support it
Another thing of the "Pro" mode in Gemini, is it doesn't only work for juicing NB2 into NB Pro. It can also make the chat model, canvas mode, video modes and music modes better.
... so yeah that'd be a good thing to have.
Ofcourse the pro model is superior.
For the anime style image, if you add a prompt - Photo-realistic scenery inspised by real beach sunset, does it still give you bad output?
I would think that a JSON style prompt with Negative Prompt: Anime Background would do the job.
I can't get anything done here since Friday. It's really tough working like this
Do you have paid plan? Pro, Ultra or whatever.
Yeah, Ultra
.... it shouldn't be doing that to you, what error messages do you get?
"We noticed some unusual activity..."
I'm not doing anything different here, nothing different from what I've been doing all these months using Flow. It's really weird
same
Try clearing cookies specially the cookies related to labs and google, and don't ask for more than 8 generations at once, wait till even the last one generates or fails to generate, or about 2 minutes before requesting more.
If you get the error message again, gotta start at all over.
I'll try to do that. Thanks for the help!
Then if you start getting the "You're requesting generations too fast" error message and it's still happening after like 5-6 minutes it just means you're out daily generating attempts for the day, gotta come back next day after like 30 mins at the same hour after you successfully generated something the previous day.
And yes it doesn't matter if you generated something or not, you only get a certain amount of daily attempts to generate something. It's according to paid plan, free plans get like 200 attempts, at least double that for each tier upgrade, I think Ultra gets much more tho.
... going at 8 images/videos at a time, Ultra would be there all day tho.
- That said, this is not that very specific error about Flow's generated content getting stuck at 99% and sometimes saying "oops something went wrong" that's something Google needs to fix on their end, if seems to affect accounts at random and at random locations. Tho maybe it's not happening so much anymore atm.
okay tell me if you start fill it
I didn't know that. Interesting, and it makes sense. Clearing the cookies worked for me here. And before that, when I generated some images/videos at first, the problem only started when I requested too many images/videos at the same time. There must be some new limit that wasn't there a few days ago. Thank you for the detailed explanation!
Just have patience and if it gets worse maybe go down to 6 at once, it helps to watch the clock as not to mess up on waiting a couple minutes. It'd be worse if not anyway.
Hello! I like to come up with fashion ideas with flow, but I'm worried that I'll accidentally generate something that it thinks is nsfw and get banned! How revealing can something be without breaking the rules? The rules I found weren't very specific at all...
it won't ban you for anything unsafe. I just cough know, specially with my pics migrated from Whisk. More than all because it has it's own built in censorship measures and it wont let you generate something very NSFW, at worse suggestive stuff will pass.
What if it's something revealing, but not suggestive? Like beach wear?
I do beachwear all the time precisely too, like a hundred images a day and no nothing happens. That said it will indeed give you a lot of error messages of all kinds, but none of the them mean it will ban you.
Okay, thanks so much!!
Actually, I can't think of anyone who got banned from Flow... but maybe just don't link stuff directly just in case, cause the Discord does have some the usual rules tho.
To be absolutely safe, you can report wrong generations. If Flow did not follow your prompt, you are not to blame unless you left it ambiguous on purpose.
Click on three dots on top right corner of the image and you can see Flag Output option.
Before they restricted to only generating 8 images i was actually able to generate some images and it didnt block me or give me error it just passed
Like a couple months ago it worked much better yes, you could even generate 25 videos/images at once. And NB2 wasn't copyrighted/censored to snowflake standards.
Is pro less censored?
i meant to say i was able to generate some nsfw images and it didnt block me or censor me
Recently yes, not by a lot tho.
All the censorship and copyright is mostly due to how NanoBanana 2/Pro and Veo perform searches, it reads Wikis taking too much info too seriously. Like Wrigley has been saying it has some sort of search feature, which only shows on paid plans at Google Ai studio and I think at Vertex, it'd be good if we could also turn it off here but we can't.
Turning it off would make it more like how the Whisk's Nano Banana (Gemini 2.5 flash Image) model is/was.
Hello. I can´t use image of boy, or girl (kids) as reference image for Imagen 4? Why? I don´t create something bad. I want only create images with same person, in other clothes. Or playing football. Nothing bad.
And seeing Google Ai studio doesn't care idk what's the big deal with Flow.
Okay- and I'd also like to know, is pro worse than 2, or just in some ways? And when you generate videos, fast takes 20 credits while lite takes 10. Does that mean I'm paying extra for it to generate faster, or is it a way better model?
Flow's Imagen4 is bugged and when prompts are too long sometimes it thinks it's unsafe, try making your prompts not too long and less detailed so it doesn't misinterpret you.
As in shorter yes, as compact as you can.
Oh! So if I have problems making certain outfits, is that something I can use instead?
I only written. "Girl has tuxedo". I mean, it is short prompt, no?
You could try yes. Specially the Nano Banana (Gemini 2,5 flash image) model. Flow and Google Ai studio are linked to your same Google account, but I think only Pro plan is giving any benefits there atm. Free accounts only get like 10 images.
If its that short and it still fails maybe you're gonna have to change tuxedo for some other word. But yeah supposedly that'd be short enough.
It write something about age problem and all generated images are failed.
Idk maybe the prompt is longer than that, or you're having some other error message related to something else.
But you haven´t image of girl as reference image. 😄
There we go. It's the reference image then, it's rejecting it. Idk man maybe try at Google AI studio with Nano Banana (Gemini 2.5 flash image) instead, it's the same thing but done better.
I want to say make a Gemini canvas with the model, but I cant explain all that mess atm.
Yes, when I generate some image of girl in t-shirt example, and I set this image as reference and after I rewrite prompt: "Girl has tuxedo". Reject it.
So in sweden there is an election this year for the parlament, i decided to make an irony video for each and every big political party in sweden this is the first one , in english https://youtu.be/LDz46mwGGlA?si=-kGHAyxj5xVuEfXq
Swedish politics Ironic takes on the upcoming elections.This film is satire, parody and pure fictional chaos.Any resemblance to real politicians, activists, protests, meetings, slogans, coffee addictions, ideological contradictions or public transport complaints is entirely coincidental.No LEGO revolutionaries were harmed during production.
@gloomy verge I think it's because Girl is interpreted as not an adult.
Try replacing it with X-year (20+) old adult male/ female
Anyone know what the "create hub" is?
Is google ai pro free trial ended?
The song's good. What AI service are you using?
Can I DM you just to ask a few things?
arent you scared of violating copyright? theres a copyright on lego
Hello I’m using Suno 5.5 but I’m very, very picky with my prompts, the lyrics are rendered by ChatGPT after feeding it with fun situations happening inside every party. I also have the Swedish democrats ready
Nah, there is tons of videos out there with this figures and actually, itś not me making the pictures I instruct it to be LEGO like mini figures, not LEGO figures soooo…..it’s of course a grey zone, but if they would complain I will remove the videos, not more to that. 🙏
at the end its not how you instructed the AI, it is how the models look like.
i would be scared of getting a letter of a lawyer tomorrow, demanding 10.000€
I take my chances on this one…to fun not to do🙏 🙏
And the Lego it’s self has no copyright any more, the mini figures does do, they are newer than the Lego it self
I checked the copyright, and how Lego reacts to videos like this: The LEGO Group embraces and encourages user-generated videos starring minifigures, often reposting creative stop-motions and fan films on official channels. However, to protect their intellectual property, they strictly prohibit commercial use of their trademarks, unauthorized third-party brand logos on minifigures, and any content depicting extreme violence, drugs, or adult themes
So I think I’m in the clear
Sorry was out with the grass, and sure.
NB2/Pro's and I imagine VEO's searches are probably linked to the Gemini model and tend to block copyrighted realistic faces (alike those from real actors from movies and series) more than whatever that's unrealistic cartoonish/anime abstract, because it can read from Wikipedia articles if a person or brand is prominent enough or not.
Lego stuff might pass, specially if it's semi-realistic or cartoony. Stuff alike from Marvel and DC can be the same, I did put a bunch of experiments here.
Characters like Wolverine for example are very related to Hugh Jackman so yeah it's much harder to make it pass but being ambiguous enough it can
Then there's Disney stuff for another example, that can pass too unless it's volume of popularity is immense atm like with Zootopia, but there's also workarounds for that... but yeah that one specially is hard to crack.
Anyway probably gotta stay away from photorealistic cinematic style in those cases, much easier. Like try to make Captain America from those MCU movies, yeah can be done but it's a headache, comics cap does pass like butter tho.
Yeah i tried to look up in chat to reply to those examples you put out in here, but I cant find it.
Look for captain america next to my name, some image examples will pop up
Like with the search filter.
I've also done Shrek with Disney characters lol
Lol yeah i think i joined on that particular chat string that day.
Please I'm trying to create a movie on flow but my videos prompts are been rejected. The movie is about street survival and who will become the next street lord but all my prompts are been rejected even though I'm using ultra. What might be the issues and I have used all precautions
I am wondering if the reason for increasingly strict censorship is because the EU. Musk just said on X that censorship and takedowns are all required by the EU. Google's censorship may also be for the same reason.
Policy orientations in some parts of Europe appear contradictory: on the one hand, they adopt an extremely permissive attitude towards excessive entertainment (even the sensory stimulation industry), while on the other hand, they impose strict restrictions on political freedom of speech, promoting a particular extreme line. This phenomenon inevitably prompts reflection on whether its essence is to divert public attention from core political issues by indulging the masses in hedonism and "tittytainment," thereby substantially suppressing the space for public discussion in order to consolidate their control and authority.
It'd be good if anything that made sense like that was the reason for the patterns they chose for the censorship and copyright blocks of Google's image/video models. But it's obvious someone just let AI choose what to do there, pushed a button and went for coffee never looking back.
Kinda obvious when you can't easily generate something with the word "Wolverine", as if the word was only related to the marvel character.
It was just poorly handled
Elon has indeed publicly positioned himself as against the EU’s approach to online regulation and moderation, especially parts of the Digital Services Act. So no wonder Grok is the most uncensored AI platform, tho it does have it's limits.
So considering one can have a stand against it and even laugh at em', the real problem with Google's generation tools probably isn't that, but instead that they've attached them to public access services alike Google lens, the QR code readers and others.
Grok stuff doesn't have that problem because they keep image generation separate from their public access platforms at paid access websites. Can't have more permissive image generation/video models when everyone and their parents are also gonna use them for whatever while sitting playing at their phones. Sites alike Flow here would obviously need less limited and separate artistry models... maybe even age verification.
And then if a country doesn't like, they can just ban the page/website instead of disabling everyone's phones.
That's actually how legality works.
No, Google has operations in Europe. Europe has 700 million people. Google Chrome has a market share of over 90% worldwide. Europe, in reality, lacks freedom of speech. Therefore, we must yield to Europe. I use Microsoft Edge on my computer and Baidu on my phone for this reason. In the UK and Germany, expressing political opinions online can lead to imprisonment. This is true.
Yeah Google is everywhere and in everything so that's their problem. Like the play store in every android device.
And the image/video models we use at Flow are the exact same ones every device uses atm. It can't work like that if they're also gonna run an artistry site like Flow, needs to be something different.
Maybe that's why Google AI studio hasn't disabled Nano Banana (Gemini 2.5 flash image). it's the closest to that they have atm, and also works as separated closed site. Also since you can turn the model's search functions off there, the other models can become more permissive like Wrigley's been showing.
Google's revenue in Europe is as follows: According to its 2024 financial report, Google achieved revenue of $350.02 billion for the year, of which $100.12 billion was net profit, demonstrating strong financial performance [1].
Yeah and that's not always a good thing, like that they don't need to listen much to other people.
That doesn't mean they don't have a lot of failed projects, it does happen to them over and over.
And if Flow doesn't take a similar approach to Google Ai studio by adding those options too or even better gets a separate image/video model that isn't also used in every other platform, idk really. How far more can people rage.
Eitherway I'm probably gonna pay a Pro plan again just for a month to see what's going on at Google Ai studio now.
Flow, Gemini and Google AI studio's same Google account link with that plan.
is anyone experience more errors on flow today?? i got error on first attempt.
the prompt was about dacia car 24hr race. and it got error saying unusual activity. until yesterday i got errors after generating at least 4 images. but today i got error at first attempt. i even tried simple hello prompt but still not working.
Yep, same here. Logged in after 2 days, tried one prompt and I get the same message.
Create image of car
Hey everyone! Just a friendly reminder to keep all conversations in the supported language for this server. This helps our entire community engage effectively, and we've gone ahead and deleted the recent posts that didn't comply. Thanks for your cooperation!
No begging, are you create or not?
flow is much slower now, but not as many errors, good to see!
how are you adjust an elemant of the picture like person or object to move them higher and lower?
I have to says flow has been worse then whisk, at least whisk can keep high resolution image unlike flow
can someone please help on veo 3.1 Ingredients I give it my start frame image and then it chanhes my start frame pose with completely different pose. How can I fix this?
how you fix this error "Promt need to be Provided"
Guys, when UXR's will started?
what are the daily limits of nano banana 2 and pro for pro accounts?
i had my limit reached quicker than usual today
Can i talk with someone?
hi all, please use #1238621511401734254 to submit any bugs! these threads are actively monitored by the team and so may be seen faster!
You guys HAD to kill whisk so we cant edit imagen 4 images proeprly anymore, and never even gave us a functioning search in flow
Hi, don’t use ingredients. Ingredients are for references that means that it can infer what you want, but it doesn’t need to use the images that you give it. Use start and end frame
Start an end frame will lock in the frame that you want
Thank you so much for seeing my message but let’s fix example I want my video to be a handout shake, I’ve tried frames and I’ve also had friends here on Discord try to help me out and quite honestly we struggled and failed and wasted many credits using the frame starting in frame option on VE03 but with ingredients it worked but did not manage with the post which I wanted
Maybe change the perspective or camera of the photo??
I’m just thinking outloud. Not sure it would work but it’s always a possibility. You have to try everything sometimes.
I was attempting to get a monster with an exceptionally large mouth to eat someone, but I was unable to do this. I had tried everything and was at my wit’s end when one of my friends unexpectedly generated it on the very first try. Honestly, it’s all about gaining a fresh perspective and exploring various approaches.
And other times it’s honestly just luck you catch the model on a good day and it does it for you. Don’t ask me why..
Ai is annoying like that.
does anyone of you know if i bulk download videos on flow for example lets say i create 100 videos & i download them all by marking them & use the bulk download function, are they kinda sortet by creation time/date or are they all random in the zip file?
because i work alot with csv files & i use the creation time to figure out which videos will get which file name for the csv
✨ Seriously, this needs to be fixed before Veo 4. It's incredibly frustrating that we can't delete a video or image while watching it. Once we click a Igm or Video we created in FLOW -->> In the upper right corner, we can - Favorite - Download - Share - Hide History. We urgently need a Delete key added there.
We routinely have over 100 videos/images in a single project with extremely subtle detail changes (and I'm sure other creators have even more). Once we leave the viewing screen, there is absolutely no way to figure out which specific version we were just looking at so we can delete the correct one in the overview. Please fix this.
Guessing they nerfed the daily limit to NB Pro on Flow after recently increasing it?
Wating for Google I/O, after that , let's see if they will do anything . Guess if they will simply abandon FLOW and create a new platform called SINK
What this messages means I can't generate the video you requested right now due to interests of third-party content providers. Please edit your prompt and try again. why is flow giving me this error?

oh, fixed
now keeps giving me error again
okay, changed browser and now its working
now it's not again, don't even created 80 images with my pro account what happened to image creation limit cap?
hello, do you guys can download your videos rn in flow?
This is fantastic! This is so exciting! The Google I/O conference is about to begin! I hope we'll see VEO4 and Banana 3!
I hope to have the best quality model currently available, surpassing Seedance 2.0, and I also hope for more lenient policies.
hi all, please use #1238621511401734254 to submit any bugs! these threads are actively monitored by the team and so may be seen faster!
i cant create more images with nano banana pro now, anyone knows why?
omni is live omg
Gemini app or elsewhere
it's happening right now?
My limit is 99 Images 🙁
gemini app
Oh? How can you tell? Upgrade membership or?
theres no indication that says its omni video but , i checked network log and its actually omni video that was doing it
idk im on ultra right now, the only way for u to find out is that when the video done generated , it should give u 10 second video
rather than 8 second
omg omni is gonna be huge game changer for ai video-
saw the thumbnail of my generated video
Nvm it's on other plans too
Lol why a head stuck out the steering wheel 😆
https://tenor.com/view/hans-hans-niemann-disappointed-disappointment-sigh-gif-18055457960592821606
you know what , ill try to have positive mindset about this whole omni video hullabaloo, surely its becuz of my prompt but either way have this omni video generated
i seem to be having an issue with Flow image generator. im trying to make some images on there, but when i hit submit, 2 seconds later every try keeps telling me this. i try new prompts and images, old prompts and images, and this thing doesn't even try.
even when i hit retry, it doesn't even try and says the same thing
They could be updating the image model. Patience my friend!
if thats the issue, then ok, i just needed to know why. thanks
For NB 3 or something?
Anything is possible
I'm getting the same thing?
Patience is key
New day, new issues, do you experience this error:
It's gotta be an outage...
I'm trying to generate videos, but after a few seconds it goes back to the main page. Is this happening to you too?
i am getting this errror as well...Is flow gone?
same here
i have ultra :/
same for me
Since we are all experience this issue. This is a good news, that means it will be fixed soon 🙂
hi all, please use #1238621511401734254 to submit any bugs! these threads are actively monitored by the team and so may be seen faster!
Finally the Archive is now reverted back as Trash.
Done
I hope this is related to veo4 release tonight 😁
I was wrong, we have new UI 🙁
But also the 'same number of request exceeds the quota limit' error also happened to mine too.
is it happening on every browser?
I’ve only tried to generate two videos, and it automatically redirected me to the main website. Now, out of nowhere, this message appears.
oh so everyone cant open flow?
i guess they are giving up updates for the next i/o? but now we cant use it 🙁
its back
Don't wory guys there is a relase, but it is a UI change
Please refresh and check it out!
Still down
I can't use FLOW why?
eyooon new UI\
time to give grok my money lol
its back but i am getting logged out constantly ..why?
Wait, I see a different error.
same here, error 253
Storyboard Studio❤️🔥
flow didint work ?
The agent & studio ✨ ❤️🔥
I try generate, but automatically redirected me to the main website.
@restive marsh me to
🧐
Try refresh.
It's a new update
The number of requests sent exceeds the quota limit. Error code 253
flow isn't working for me. is there a new site?
Give it few minutes, they applying the update.
kk
We’ll have to try again later, in several hours.
at least give an annoucement team
FLOW AGENT
Give it a few minutes, we're ramping up the traffic
yes, now appear
Top-notch
what is this?? flow??
okay I see the new update ... I cleared my browser cache and now I get all the failures to load media .... is this what everybody else is seeing? I also occasionally saw the error about quota limit being reached
"The number of requests sent exceeds the quota limit. Error code 253"
AAA is this likeness
The number of requests sent exceeds the quota limit. Error code 253
i was hoping flow will get better and that unusual activity and requesting to many images will gone after 19 may but i think with this new update the traffic will increase like hell and they will limit usage more 🥲
for me still getting errors but seeing this new updates are craze
crazy
Now we talking❤️🔥
flow agent 🥹
Please remains calm Google updating the website, just come back later for the best experience! - and yes, Obviously the next few days it's going to lagging/updating. 👌
yep yep this is a very much huge QoL update for me
A giant leap forward.
no, it got worse, cause now its gone when i enter flow, and when i try going to it on google, it gives me this
wait for some time.
Because it's updating
now i have double the agent
what is this mean?
Just be patient
ok then
veo4 update???
Lol
im gagged , the scenebuilder
I'm on mobile i can't get a double screen
yeah i realized flow agent doesnt show on movile at all 😭
I completely forgot to check that with so many new features dropping lol.
They do have a mobile app though in beta
no "literal" new model it appears but
if u use flow agent, to generate image or video, it seesm to be using a new model
ok??? we publishing to youtube
okay. cant wait to use .
if anything, close all your bworsers altogether, then open brwoser again, load up flow,google
cuz right now my other browser shows 423 error
but this one browser i have it opened seems fine
CTRL + F5 (mash it like 20x)
Lol I will publish everything now
are you working right now?? do you face previous unusual activity or requesting too many or ooops something wrong errors?
Nope! no more unusual activity!
bro why i can access my flow back but still the old one??
but i might spoke too soon 🙂
no ...i tried. i guess i will wait for 1-2 hrs .
T.T
oh not the toolbuilder! ❤️ love it
try spam 25 times with image generations ..😂 then we will get answer
seems like the character thingy is basically the same as creating a person in nano banana and manually uploading the image to the video generation?
Agents are everywhere now, and they’re top-notch. I’m done with google studio.❤️🔥
yall literally bringing whole aistudio vibe coding feature to flow , i am sat
Ok yeah Google is cooking
well we've got a new round of outages 🤦🏻
Need her to sit with me
i turn off agent now
Agent Smith is your intern
Obviously, Flow has been updated. 👍
wow can we automate generations with this now?
Wait, very useful for my ZVL Characters.
I'm getting an error on the page. Can someone tell me how to fix it or is there a page problem?
it depends on what kind of automation u want agent to do for u! eventually it;ll deliver!
so no longer prominent errors for generated images ?
ctrl + f5
Haha, it literally rebuilt the [UI] interface just to add the model dropdown for me, what a dream.❤️🔥
its lagging or just me
Is anyone having trouble generating videos on Flow today?
The number of requests sent exceeds the quota limit. Error code 253
ctrl + f5
okis
oh no its cameo/character creation
@flow team You Rock✨ 💯
its like sora 2 cameo character
Having fun 🚀
still not even 1% close to seedance 2
u know i was scared if they gonna make me look like im in jail mugshot but this is nice
the mugshot in question :
Thank ya thank ya. Lots of work went into this one
This is useful.
@fierce nimbus u better hav those huge changelog ready! 👀
with all this new tool do we have a new model of veo?
I think I’m in it hahaha
You might want to do a complete refresh and try again.
Flow has been updated... 👌
now automatically redirected me to the main website. Now, out of nowhere, this message appears
3d animation got a huge update with only 2 lines prompts
you can create your own tools now.
yep im gagged again
new video model?
yeh omni flash
wow this is amazing
Sometimes it sends me to the official website every now and then, that's a page error problem or what the hell
This is far from ordinary work, there’s some seriously high level thinking behind it.
It’s the first time I’ve touched a platform that no longer feels like just a platform, because I can now reshape almost every part of it.
Honestly, it feels like a completely new concept.
wow finally consistency with voice??!!! thats literal dream
Im just waiting for my flow to update it is still not updating
same for me
veo 3.1 fast use 40 credit oww sht...
and qua?
200
remembering 3.1 fast was free not long ago 😄
👀
gonna test stress this out
yesss do it! lets us know what you think!!
Can anyone download videos in 1080p? I’m getting an error message.
we're working on a fix for the upsampler. stay tuned
thx
Make sure it goes under “my tools.” With this concept, it feels like you’re basically adding skills to the agent, so I’d assume openclaw is no longer necessary.❤️🔥
@midnight cedar will there be ways to backup / export everything that was meant to be consistent? For example a created voice or a created character? i guess exporting a character is just exporting the image? but the voice? saving the voice description is good enough?
whats the different between veo and omni ?
Does anyone else getting a "throw-back" to the starting page of Flow while still being in a project (according to the url)? 🤔
same
i can't see omi? i see only veo 3.1 how?
Thanks, bud, for clarifying. 🙏 Good to know... it's not just me... 🤣
Long live Whisk!!
Omni is the latest video model. Longer durations, more capabilities (video-to-video, more references, etc)
ah this better not be hallucinating!
flow is completely unusable with this update!
theb limit error I believe has taken the service down
even when you get in and try to generate videos you get the failure to load new media which is probably related to the limit error
What’s going on flow is crashing
mine is working now. and the errors are there like unusual activity , requesting too much and all..
Ooh flow is more than a model..
It's a Runway lane of opportunities
All
78 , 98 , 27 ,55 ,45 69, 69 xD
I just built an agent within Flow to create superheroes
things are getting more and more expensive. 3.1 fast = 40 credits now and omni = 50 credits.
not getting far anymore. i guess you need 3 ultra subscriptions a month nowadays
is only i am getting this errors? or everyone facing same as before like past days?
Great work, there’s so much here, it’s going to take way more than a few hours to explore it all.✨
Veo is 10 credits for Ultra. You might be requesting a batch of videos.
Nece job guys, so much cool stuff again, you rock!
whisk doesnt work anymore it just redirects you to flow
work for me 😼
its there
@inland pier I want to build a custom tool that takes a list of ~100 prompts and generates product videos/images in bulk, auto-saving to the Media Gallery. A few things I can't figure out:
Reference image / start frame: Does Flow.generate.image() accept a reference image (to keep a real product looking identical)? And does Flow.generate.video() accept a start frame / first frame image? What's the exact parameter name (e.g. referenceImages, firstFrame, mediaId, base64)?
Uploading my own images: How do I get my own local images into the tool to use as references — upload to the gallery first and pass a mediaId? Or can I pass base64 directly to Flow.generate.*?
Avatar + voice: Is there an avatar / voice (TTS) capability in flow-sdk for a consistent on-screen presenter and narration? What are the functions?
Batch loop: Can a tool loop and call Flow.generate.video() many times (100+) in one run without hitting a rate/quota wall inside the tool runtime? Any per-run limits?
Feeding data in: What's the recommended way to pass a large JSON list into a tool — a textarea/paste, file upload, or fetch()?
Bulk export: Once everything is in the Media Gallery, is there a way to download all of it at once, or get the file via the SDK from a mediaId?
Docs: Is there an API reference / docs for flow-sdk anywhere?
Thanks!
i dont see it on list
now i have yet to try all these COOL WHACKY LOOKING TOOLS , im overstimulated omg
its in tools page
let us know what you think!
Upload to your gallery first, then tell the Tool agent that you want to pull media from the gallery. While you will be able to do some stuff in bulk, there will still be rate limiting.
AGAIN it automatically redirected me to the main website. 😔
also
Try visit the "Tools" on the left side bar of a Flow project, you can find Whisk there
When the tool pulls a reference from the gallery, how does it target a specific item — by filename, an id, a picker? (I have ~31 distinct product photos and each job needs its own specific reference, not a random one.)
What's the actual rate limit (e.g. N generations per minute/hour)? Need it to set the right delay/backoff.
Is there an avatar + voice/TTS capability in the SDK for a consistent presenter/narration?
Any bulk export from the gallery, or download per item via mediaId?
I'm trying to generate videos in flow, but after a few seconds it goes back to the main page.
I don’t see any cost or credit usage info in storyboard studio yet [hopefully it’s free lol], but honestly, every other platform I’m beta testing with the same concept is in trouble right now.
you select which one you want through the media picker within each tool!
tools do not yet cost AI credits, but there is a daily quota~
hahahahaah to show you the power of Flow's new applets, we made brought you all back Whisk inside of Flow 😄
Gtsy, Anika:)
looks like flow got all the features lol
For a Tool Builder tool: is there a programmatic way to reference gallery media by id or name from code (e.g. Flow.gallery.list() or get-by-name), instead of the interactive media picker? I'm batching ~30 products / ~180 generations and want it to run unattended.
🙂
no programatic way yet, you need to manually select (but you can multi select by bringing up the media picker and hitting command and clicking)
we're working on a fix for this
When you multi-select media in the picker, does each selected item expose its filename/name to the tool code? I need the tool to match a specific reference to a specific job by name (my files are named ref_<product>). If names aren't exposed, is there any identifier per selected item I can map on?
i dont see flow on here but it was there yesterday anyone know why
you can't ground your media in tools yet, just put them in as inputs when you are using it
I still don't see that option
i had to ,, go back to main page then create new project again
I assume it's not on the actual website
Just app
I created a new character and it was for only 8 seconds
looking into this!
what do you mean? you can reuse the character for subsequent generations
You can access Flow directly at: flow.google (https://labs.google/fx/tools/flow)
It was an error. I'm using multiple places from mobile and the beta app
hope this will fix soon.
When are we sharing characters?
Yeah its a lot happening. I can't get a video made now
did flow get a new update?
Could you tell me where I can find documentation on Google Flow updates to read, or do I need to practice?
we will have a help center article live for you to read closer to the event in a few hours
thanks
the fact that you can co-code your own tools with veo tools is surreal
i just added a button I need it on a UI
Does it also cover the difference between the lasso ("Select") and the pencil tool ("Draw")?
Still don't get the difference... 🫣
I already made a superhero generator
I won't show the images until the event starts
excited to seeee!
It keeps logging me out tho ^^" But it works well. Omni video gens are very nice looking.
okay how in the world do you bring back the repeat option on the video playback?
when I play the generated videos it only plays once and I have to hit the play button again
I don 't see a setting for that, but could merely be as browser setting
can you use tools on a free account?
my first try with using flow agent and omni flash video model
still need to explore more ! one at a time c:
Sharing a generation made with a character shares the character as well!
It seems like a failed generation does consume the credits tho.
Why is the flow website not working
You should be able to use tools, but not make them
It feels super agentic.
The upload option is gone
"generate the first 10 shots and cut them together like a proper short film, with strong creative direction in mind!" Hey G does this worked for you?
Character works quite well ngl. That's one serious improvment, gg on omni
Guys, I'm trying to generate from the agent instructions and this happened
It was doing it with veo 3; after three shots the service went into an update and kept refreshing, so I couldn’t finish the run, but it was working before that.
still getting this. I'm guessing they're still updating?
Does Omni not count as Veo for extension?
Stuff are being updated and ready for premier launch
same issue : oops something went wrong, unusual activity and even this for the last 3 days
Even after updating the version, the error is still occurring.
Does this still happen to anyone?
delete cookies
Only 1 hour left
try clearing the browser data history it would work temporary
I don’t know. Because I’m getting the same thing too as well
I do, but still getting
?
Also curious, right? 🤯
Why can't I upload images
Omni. No reference frame, only ingredients. And the result is.... 🙁
I've been with Google for awhile next to another prominent company. Never doubted
100% logic wow , lol
Yeah, it works temporarily, but after a few generations, I get this error again.
I'm more interested with the coding abilities within Flow itself
What's your prompt?
did they remove reuse prompt feature ?
Keynotes has begun
is Omni a agent or model
??
unless it's veo 3.1 references
All will be explained
already?
Yes
what do u mean by keynotes
Check Google live stream
the main one at 10am pt
I like to follow all notes
share it with us pal?
on yt
It's time to I/O! Tune in to learn the latest news, announcements, and AI updates from Google.
To watch this keynote interpreted in American Sign Language (ASL), please click here: https://youtube.com/watch?v=vw73r_xaeX0
#GoogleIO
Subscribe to our Channel: https://www.youtube.com/google
Find us on X: https://twitter.com/google
Watch us on T...
starts in 55 min I think
Get hyped guys/gals!!
Yeh, the friend above had likely meant that the stream was activated.
Flow is killing it. Kudos to the team.
But lately, google models have been so disappointing.
I get your standpoint but this integration in Flow is something very needed
Almost near.
I meant that: Flow is the GOAT.
But Nano 2 and Veo lite/omni are lacking
Time will tell
I see your cooking something
how can i do to solve the message: you re generating images too quickly. It doesn't dissappear
I'd wish for an Ultra+ sub with unlimited access to full power.
From what I discovered, if you're using the pro model, that just means you hit your limit and should switch to nano banana 2
im on free model
mostly due to censorship
Yeah that and reducing compute power.
wow great new features for what i can see
also able to upload videos, finally
looking forward to the new tool section
woah is this true
Hey
yea. i can barely use pro model. limit is too low
up to 1 gb it says
Google Flow stops working and exit the project. Anyway to fix?
could u put the vids as reference
Used Omni with one of the Nano genrated images and got this error while trying to make a video with Omni "can't make vids because of celebrities" which still makes no sense since it was an image generated via Nano..
Tried it with and without using the " + " sign : /
trying at the moment, but still uploading the initial video
Reload previous page
THe reason is lots of sessions imo, haven't faced this problem since today
(Google Flow stops working and exit the project. Anyway to fix?)
I've managed to create better short films back in October, pre Veo3.1 than now. It's not Flow's fault, again. Just wished Google had not gone full reverse on the progress of their models.
maybe omni would be better
it is flow's fault. they cant make a new model. but they do control the parameters it operates.
First thing I tried today with omni.. Come on.
Honestly what the heck was google doing for a whole year???
this omni model is horrible man
Back in October November, I made short films that got me into ÉCU, WAIFF, and many other festivals, but not, I hardly see it possible.
THIS is what theyre showing to the world? gotta be kidding
in that sense, yes you're right.
Problem from your prompt maybe?
My results for tests were correct
I've been prompting successfully for months. I really don't think prompting is my issue unfortunately.
As I said, I clearly notice the drop in quality over the recent updates.
I'm liking Omni, and was able what I might normally have done in three shots into a single 10 second omni shot with 3 different scenes.
I had to include a "character" reference image, and specific descriptive language for the in each section of the prompt for each scene, or the costume would change in each scene.
Does any of you guys experiencing this too?
Yeah, randomly.
So do I
Yeh
Yep. Also keeps just randomly refreshing back to the main Flow page - completely losing anything I was in the middle of.
Anyone know how to fix it?
Same here
Same here.
Same here
I'm still getting this lame bug
same
Same
same
OK, a theory that would explain a lot:
omni is only a sort of agent that uses Veo lite in the background...
its looks like limit cause new models
Still gets the same problem as before with "celebrities"... I added it into the labs bug -
Also added the bug about being sent back to main page over there.
OK I love the omni uality but its not responsive to my queries when i say "no dolly shots, scene remains the same all the time." it creates me multiple transitions... I'm striving for the handheld camera recording "one raw candid shot". THIS WAS MY PROMPT: "American nun in her 60s says: "What we eat inside these walls is one of the reasons we are the happiest on earth. We never go to the supermarket and we grow every single thing we eat" tone: Content / Serene. action: she puts the metal bucket on grass and Her left hand holds the cow's teat, performing a slow, steady pulling and squeezing motion to express milk. her right hand stays grip holding the camera. the sisters in the background are yelling "yessss sisterrrr!!!!" from the behind field, running towards hugging the cow, yelling and cheering. candid handheld-style recording with slight natural shakiness."
would love anyone insight on this Im a bit newbie, like that transition kills it
yoo boy doing its best lol
I kept on getting this error:
The number of requests sent exceeds the quota limit. Error code 253
Should I just wait? Anyone can answer please? <@&1173646484311384114> Sorry for the ping
Seriously, this error again?
Yeah, it's annoying. It's been like this for like 5hrs already, and I have an Ultra Plan
Same for me
we're investigating!
I was in the middle of making characters.
Won't you need the email I'm using? It's delaying my work tbh
Can you please address the fact that we cannot generate more than 8 consecutive images at a time when two weeks ago it wasn't like this? Or the "prominent person" filter turned up WAY too high?
thisss
I cant access flow anymore, whats hapnneing
Is this because of the new update??
😭
10 tests and ...The number of requests sent exceeds the quota limit. Error code 253 ???
Yeh, you were right, my bad
Half an hour later, I admit my mistake x')
Veo 3.1 sucks at this kind of shot
Pretty audacious for this to be going on when they have a keynote happening
im just not gonna use flow until they fix the number of requests sent exceeds the quota limit. Error code 253
The number of requests sent exceeds the quota limit. Error code 253
limit 253 bro
Everything was good in the morning
It's time to I/O! Tune in to learn the latest news, announcements, and AI updates from Google.
To watch this keynote interpreted in American Sign Language (ASL), please click here: https://youtube.com/watch?v=vw73r_xaeX0
#GoogleIO
Subscribe to our Channel: https://www.youtube.com/google
Find us on X: https://twitter.com/google
Watch us on T...
AAHHH Same... it's not just me right?
I was able to use it fine an hour ago, but now this error is appearing. I assume, and from what I can see, it's a widespread issue.
we're investigating: "The number of requests sent exceeds the quota limit. Error code 253"
this gettin outta hand
Where did the upload botton go?
Btw, how much are you guys earning per month with AI?
like 0
so i tried geminin omni. my initial impressions is that it feels like veo 3.1.2
Me.. roughly 4.5k USD with two clients, UGC style ads
Where did the upload botton go?
mind telling me how?
I don't think i like this new model, at least so far, i think other models that i wont talk about are just better at dinamism
im not feeling it either, im actually disappointed.
First create a portfolio, create sample AI videos, go to Upwork, create a profile, and find clients, it's been working well so far. But I feel like, there's more in there...
This error hasn't appeared for awhile:
The number of requests sent exceeds the quota limit. Error code 253
Is it officially fixed?
I think it's not a new model. It's only a sort of agent using Veo lite in the background. Hence the poor results.
must be because this is one of my outputs
oh thought u were doing ai model hahaha
Nah there are different ways to earn. I'll actually try FB and YT automation again, with 30s explainer videos.
"failed.
You're requesting generation too quickly. Please wait a moment and try again"
help i been getting this for full day, and i did try again later after some wait it still the same, it happen like 2 pm today till now
Jesus. We're doomed.
oh ok i see... it's too long hahaha
Yeah but it's the sweet spot for the length
Exactly whats happening
The Team is currently investigating: "The number of requests sent exceeds the quota limit. Error code 253"
we ask for your patience
left is veo 3.1 lite low priority. right is omni
🤔 Can we generate 1080p yet via Flow platform or Omni still at 720p?
you can upscale buty omni can do 10 seconds
I think omni is like leaning towards Kling quality
i can tell you right now that its is defenitely not as good as kling 3
Yeah but obviously different from Veo 3.1
i think it captures the amatuer look better than veo 3.1 but from what ive tried, its only very slightly better than veo 3.1 like veo 3.1.2
bruh, i cant get to flow. i click "Crreate with google flow" and it just reloads the site
wth
same here, a lot of us are having Error code 253
can still use it on my phone oddly (free user)
are you european by chance?
UK
but what they are showing right now, it can do some cool stuff
yeah germany here
Yeah from the video you shared, it's better than the prvious version but I think it's lacking frames
they say they are investigating it
so best just sit back and wait (or play HOI4 in my case)
this a spammer above me?
Guess i'll do the same, and play Millenium Dawn beta update
amazing
oh no, not MD, i cant play that XD too complex and too bloated. i prefer Novum.
that or the fallout mod.
side question, has Omni also increased image quality/ prompt adherence?
HOI4 dont need an economy system :p
fair haha
i dunno what Omni is XD
new google model apparently
i only ever used Flow Ai as a free user for this year
in what?
i think omni might be like a video editing model more so than a big quality upgrade to veo
Yeah, you're right. Just like Kling 3.0 omni
@humble dock Hey everyone! Just a friendly reminder to keep our discussions focused and to avoid excessive emojis, spam, or content that doesn't make sense. We've deleted the recent messages that didn't comply. Thanks for helping us keep the chat clean and engaging for all members!
no. google arent that competent enough. its just a gemini 3 propmt making skills and tool calls usage
how do i get to that? cant seem to find it on google search
its on flow which rn doesnt work
oh well that makes sense 😆
all this new stuff i didnt know about i cant even try it :p
Like bad or good limits?
bad
Source: https://support.google.com/gemini/answer/17004136 Before May 17, the Gemini 3.1 Pro daily limits were 3/30/100/50…
ultra plan going from 250 to 200? is thatt what he just said?
will ultra keep the unilimited relaxed generations?
would be cool if you could enlarge the video size when you go click on the video, rn to actually inspect it upclose you gotta make an additional step by clicking full screen. just a suggestion 👍
this should be fixed now, let us know if you're seeing issues still!
Google Omni Flash is insane! 👀
This is happening to me too, I can't open flow
anyone else?
you sure thats omni? its only 10 seconds
like it works but then it just stops working
dang, omni cooked bro 💀
isnt it 10 only?
this reminds me of brothers in arms videogame
Probably that's seedance 2 lmao
nah bro thats seedance
How do I bypass waiting?
yoo CC gang
yeah its only 10 seconds with no extend option right now, thats why i doubt thats omni
Nvm, wrong video that was Seedance, this is what Omni Flash generated oops
Hi everyone, please use the #1238621511401734254 to submit any bugs! these threads are actively monitored by the team and so may be seen faster!
omni model is very censored. no point of using it, if you are paying for it
omni is insane bro
whats good bro 👀
chasing grace
bro yeah right you knew that wasnt omni when you posted it. idk what you were trying to prove lol
is it still updating with other people here?
damn so it cant compete with seedance (yet)
Seedance 2 really cooked tho
how do i bypass waiting limits for nanobanana pro
dang bro omni is other level!! wth
This generation is great, ngl
oh nvm, this was seedance 2, this is veo omni
bruh, why are yall doing this? lol
mods wake up answer
LMAO!
bruh i knew that wasnt omni, ive been trying it out and i knew for a fact that it was not capable of doing that
This is not something we can answer, if you are experiencing product issues please use #1238621511401734254 channel
you got me for a moment lol
veo will always be funny
Flow iOS app when ??
yeah, very dissapointing
cause unless its still updating, Flow still doing this to me
not een letting me go thru
Wait, so you dont even know what causes this error?
so im not the only one
whats the promtp??
❌ Hollywood is dead. ✅ Veo is dead.
Thaat's not veo tho
this is not omni
Amazing !!!!
We are here for in Discord related issues so if you see someone breaking any of the server rules listed here :#rules-and-guidelines・📋 please ping us. For bugs and feedback please use #1238621511401734254 these threads are actively monitored by the various product team!
As for this particular issue the team is aware and are actively investigating so please be patient!
same for me. been like this since last night
yeah man, same very disappointed
I think people fail to realize the omni model isn't an upgrade to Veo. Most likely a newer model is coming too
Hello, VEO 3.1 Quality doesn´t support reference images?
Wow... omni kinda sucks lol
yeah we wanted an upgrade to veo though lol
There are more stuff to do in Flow than just veo...
ofcourse, i dont use it for veo because veo isnt good
SO veo is dead now ?
I sure hope not. Results from omni have been trash so far
even by mistake you send I see Seedance is much better visuals and controlling it
Other than that Flow is absolutely amazing with the integration of coding
i really just use it for nano
Looks great
but i also disagree i do think omni is an upgrade
its not a huge upgrade but it is an upgrade
Yeah an upgrade to a certain extent
