#flow-general・🌊
1 messages · Page 50 of 1
I tried everything. Clean cookies, sing out sing in, change browser, refresh ip
but on an other google account on free version it works
Well Well Well
Google classic
Google moment
Believe me you are not the only one suffering from this :/ I'm totally frustrated
Same issues
Me too, ahhh
Yeah me too doesn’t anyone have any idea what’s going on?
Same here
Me too. Crap
Meanwhile at Google. And yeah notice I had no prob to put Disney's logo there, all for nothing.
hey all, I've forwarded your messages to the team who are investigating. in the meantime, please help us better understand your situation by creating and/or adding to existing reports in #1238621511401734254. you can use the "Flow" tag to search reported posts and issues.
📣 please let us know what version you're working with, if you're on an Ultra/Pro plan, any prompts, images, etc so we can look further. thank you!
Looks like all generations timeout, I'm on the Ultra, using the 3.1 fast
Means I'll just take a break from flow .... as we say in my inner work circles: bagel run 🤣
yaw guys its been two hours since the bananas dont work ?
NBP and NB2 are working in Higgsfield
AWESOME
Oh, wait, I thought that was an image from Flow.
Veo models have had 4, 6 and 8 for a while in the API. I don't know why it hasn't come to Flow yet.
They haven't. I did some of that just this morning.
Servers top out a few hours on weekdays, usually during the workday in the EU and US.
Cyanobacteria for roughly 3 billion years? ha
Yesterday Veo, today Flow. Why can't I generate videos?
The prompt interpretation quality really does ebb and flow. I don't see a pattern to it.
You just need to Extend. You can make them as long as you want, just not one-shot.
Servers run out of compute a few hours a day most weekdays.
The API and apps have separate provisioning.
Is there a way I can fix mine not working?
The default answer is always to try resetting your browser and computer and coming back in. If you're using a different browser, use Chrome instead, and don't have any extensions running. They have the absolute most buggy web app. Very finnecky. I have a whole browser just for Flow.
Yeah I do think it could be an extension prob too, seems a little unusual.
I am using Chrome and have been. I've tried different browsers and different devices. None of them seem to work
Try running a good antivirus too, there could be some layer on your screen that you can't see and it's blocking your dragging.
And just to be clear I can drag images into flow for uploading, but I can't drag the images already in flow to organize collections or simply add them as an ingredient
This wasn't a problem a few weeks ago. I didn't add anything. It just started not working, at least on PC
It never worked on laptop and mobile for me
And if all fails, I guess you could also try checking for solutions for when drag and drop is not working on windows, maybe it's not related to Flow.
Even windows 11 can sometimes derp on that https://www.youtube.com/watch?v=lEARLwWe0d4&t=49s
Fix Drag and Drop Not Working in Windows 11/10 [Guide]
Commands Used:
DISM.exe /Online /Cleanup-image /Restorehealth
sfc /scannow
There are many ways to move a file on Windows, but one of the most common is by dragging and dropping it. However, a temporary glitch or corrupt mouse driver can sometimes prevent the drag and drop feature from wor...
we're investigating this, see more in 📌 : #flow-general・🌊 message
Yes. I had to endure that with Veo yesterday and today with Flow. 😡
guys?
Yes a lot of users are having the problem not just you don't worry, they told us staff will check on it.
but I want to vent and get mad, as I can't get the project done.
That's the least I could do 😢
If anyone else is experiencing the same issue, please comment. It happens to me when animating images. If you can share screenshots, even better.
veooo is down
yes bro
😡
ikr. I'mma just go do like Wrigley says and grass it up... and by that I mean play a game.
It'd help if they lower the censorship a little, they surely went overboard and it's not really working anyway, just annoying the wholesome users too. And the error messages need to really state what's going on, at this point the confusion about them is just too much since they didn't fix the previous ones.
My flow videos won't load pls help 😫
hi all, as a reminder, we're investigating this: #flow-general・🌊 message
feel free to add to existing bug reports in #1238621511401734254: please procide as much information as you can to help us better pinpoint. thanks all!
same
paying for flow ultra but everything dead
we all walk the same
its been 3-4 hours , wassap with that?
Curious if there's a way to get escalated to someone who can help me with a (maybe enterprise?) question? My agency is making a spot with a celebrity owned brand and would love to see if I could get a pause on content violation if I provide proof etc?
Hi
Could you please provide an update once the investigation yields further developments? Thanks
Are you using Flow in the browser directly ?
yes
looks like data center/compute issue
please extend our subs for 1 day as compensation
always trouble but no compensation, shame on you google
They don't perform miracles here.
@main saddle urgent it looks like all users here got errors, can the flow team please check this
yes i think too
team is aware and investigating: #flow-general・🌊 message
@main saddle can I get a ticket opened to explore this?
I haven't seen that issue. I checked again just now and it works on my end, both with individual and multiple images selected.
cause main gemini app is giving me issues too
i'm losing millions of dollars right now because of this
@edgy holly Try animating an image using the frame option, and let us know how it goes.
... whoever's in charge of Nano 2's censorship needs to be fired asap and last week's values restored. This is going head on to full disaster.
just run z-image locally
This is definitely going to be the year of local models. Totally. Becoming huge
what is z-image
I love nano banana and such but if you are trying to do things that keep hitting the sensors, the best bet is just to Do your own models at home. A lot of these companies are Aware that the restrictions are ridiculous but their hands are tied, It's just the way the community is but I think that eventually it'll lighten up.
I know LoRAs can do anything, but the specific prompts I use depend on listening, it really wasn't NSFW as LoRAs do but just simple fullbody posing.
I don't think anyone would use Nano Banana 2 to do actual real NSFW as a LoRA can, the Nano models don't even know how to do that.
alibaba opensourse z-image
i think most of the user came to flow for better design, better look and no errors no errors on characters look depending on content/prompts they're doing. But since the stupid censorship came through last night it cancelled out all 4 image i was doing or sometimes randomly add clothes to characters even hough there no clothes type prompt only bra and panties prompts are on there. Now i have to find a better ai site to do bra and panties content. At least they didn't effect bikini content.
The flow has been low?
Hi, I'm having trouble generating images in Flow. Yesterday it was working fine, but today I can't get it to work all day; it keeps giving me an error. I've tried with several photos, and it didn't work!
It's more like this than actual NSFW, I know LoRAs can do around the same thing, but with the Nanos it's more about following the text instruction than adding the typical prompting/negative prompting.
And this actually is original Nano, not Nano 2. I think except for few people that still know how to bypass with a LOT of patience, we might not see this type of image again soon.
is nano banana currently broken? i cant get any images to generate
Ha! Bra and panties content …. Google has always flagged that … AFAIK that has always tripped the censors
me too bro
Yeah all is broken actually and staff said they're currently addressing the issues, check the pinned posts. You might still be able to generate teddy bears and gingerbread men tho.
oo i see, even veo 3 is broken?
What's happening? The website went down.
Everything.
Meanwhile Whisk still runs fine like a refrigerator that's been connected since the USSR existed.
yes, it is down around 4 hours
But we'll wait and see if they give us an answer.
Mercy, will it ever work again???
In fact, it’s tough even to get past the censors generating two piece swimsuits
4 ?
its down 6 hours bro
this is sad , we have work to do
really omg
It's true, I still have a few creatives to deliver.
Until about 2:00 AM, it was working normally, but around 12:00 AM it stopped working, and that message keeps appearing when I try to connect!
I just can't deal with all this creativity that the better, more powerful long-term Flow home is giving me.
Been cutting the grass outside in such creative ways waiting for Flow to get fixed and let me work again.
How what goes? What specifically are you wanting me to test?
At the soonest that year might come in 5+ years. The vast majority of people don't have a GPU at all, much less one good enough to run one of the better image generation models. It's cheaper, faster and easier to run on apps.
@edgy holly Create an image and animate it in flow, do that, will it let you?
I can run anything, and don't plan going through it again after testing the Nano models, never turning ComfyUI on again either even if it kills me. I'm not even doing the heavy NSFW those can do anyway, just basically photo shoots like that other guy.
It's never been this unstable before, it's been down for hours, I can barely generate any images, this whole situation is so frustrating!!!
Well let's just have patience anyway, I imagine if it hasn't been fixed yet is just cause they can't, they do know what's going on anyway.
The question is whether it will be resolved or if it will remain non-functional, because it has already been unstable for more than 6 hours.
Flow down?
I'm pretty sure it will be fixed like always, Google staff just works on business hours, that might be why it seems to be so hard to fix atm.
I thought about that possibility too, let's wait until tomorrow to see, or even later to see if it works, I have work to do.😭😭
Yeah tomorrow is a safer bet but could be sooner who knows.
It's like calling the cable guy
Video generation reaches 99% then fails. How long has this been going on?
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!
try lite, i usually get after 2nd try
That just means it's failing. Percentage bar Isn't actually a percentage bar. It's just the animation while you wait Period if it gets stuck at 99% ,it never loaded at all.
Some things are not changing. After not giving me any Veo access yesterday and messing my requests today; it only gave me two unusable videos and told me I reached my quota
@bleak walrus all day
I’ve been unable to generate videos on veo. What’s really going on?
Why has this been happening?
so do i
Has this been happening for you or it just started?
Right, I came in here to point that out. It's not even thinking about it anymore.
This has been happening to me too. Since last night
I have a question. But has anybody here been able to create a sequence of 40 seconds or so? How many tries did you have to do per sequence? How many credits did you have to use to be able to succeed with the whole scene? Was it realistic or fantasy?
At this point I'm just trying to figure out how worth it is Flow for what I want to do. I haven't been able to complete a video without the tool hallucinating or adding non requested elements making the videos unusable. Or as today, not working.
it is only me or you guys can't generate videos right now?
Everyone I believe
Why is there no announcement from Google? What about those who depend on their paid account for work? Is this a question of too many users but not enough resources?
i've tried
and it works using ingridients for this time
and still error while using frames
Is the error still present?
Yes
whyy??
yes i've lost a few millions because of this
Service is down, folks. Pacience. It will certainly come back sometime
Mine always keeps saying that it violates content policies, no matter what I do. My prompts are innocent.
On the day they opened to transfer the works from Whisk to Flow, I asked to be transferred, but so far nothing has happened. Can someone from the staff guarantee that everything will actually be transferred?
We can now make upto 40s?
For now, it only lets you generate images; videos aren't supported yet
This should be resolved now.
Is there any fix for being unable to remove projects ?
So video generations should be back in a working state?
Hey everyone! Has it been resolved yet?
doesn't look like it
image generation seems to be working okay ... that was a little spotty earlier today
It seems like it's still ongoing! Or has it already been resolved?
It’s STILL happening
hmm, I was just able to generate images
but video generation is still gone
hi all, thanks for your patience. this issue should be resolved. please let us know if you’re still experiencing this issue by emoji reactions 👍 👎 on my post here. please note that this is rolling out and can take a moment to reach each account.
thanks again for using #1238621511401734254 to report!
"this issue" meaning video generations?
Thumbs up is for “it’s working” or “not working?” You didn’t specify
Problem solved, I can generate images now!! 😭🤝🏻
GUYS FLOW ONLY LET YOU GENERATE WHEN YOU SELECT THE "LITE" OPTION; IT DOESN'T WORK WITH ANY OTHER OPTION.
It only works like this
Ah ... I haven't used the lite option
trie it
The image loading should be fixed, generations would be dependent on credits and service tier (ultra,pro etc)
Ultra 🙋🏻♂️
thumbs up that it’s working, thumbs down that it’s not
Ultra and no single success video gen in hours 🙁
....
so I finally was able to generate videos using the lite option 🙂
the other options are still down?
what's the difference between lite and fast ..... is lite lower quality?
Ah-ha! There's an update: Commit v-change-152148-f337ced29-260408160106 | 4/8/2026
it works now , ultra plan.
still same here and sometimes stuck at 99%
oh the ingredients works but not frame to video
Of course. Why?
Not working for you?
What I mean is that I'm trying to make a sequence that is expected to last 40 seconds or so. So far I have 3 barely usable segments and nowhere close to the 40 seconds I needed. I have seen videos with great imagery but nothing is a complete scene, only 8 seconds sections and then cut to something else.
FIX THE VEO 3
And no, no videos in Flow yet
Peace of shi....
...t
Yeah the fast option is still timing out for me and I have the update. I can generate using the lite option though ... so IMO it doesn't look like it's 100% back
Oh,I see. Apparently not all models are working. I ran the test in Lite. Quality mode is frozen at 99%.
Man, they really are bad at comms.
yeah lite appears to be working
I was able to generate this video using the fast option .... prompt A tree blowing in the wind .... as simple as you can get it 🤣
"Quality and Fast modes are currently down. We are working the problem. Until then, Fast mode in Ingredients and Lite mode are working. We will update you when the problem is solved."
That's all they'd have to say.
I only selected to generate one video that run ... so perhaps the complexity of the prompt and the number of videos per batch matter
Ah.... I did generate that tree video in fast / ingredients mode
Their default is one of the worst they could choose. It's always to pretend they dont' know whats going on, be vague if they say anything, let us nerds flail about, and a bunch of conflicting content come out that multiplies confusion, then they fix it and usually don't say anything then either.
so it's 100% the frame to frame mode that's broken .... that's the mode I use all the time lol
the fast / ingredients mode works .... I can generate videos that way
Start frame in lite worked for me.
yeah sorry let me rephrase: fast mode with frame to frame ... that's down 100% ... yeah I can generate using lite mode frame to frame
I'd probably need to use Quality mode to keep the style. It always wants to pull to realism.
This happened once right before an update. Maybe it'll be 3.5 tomorrow, or a shadow update.
Also cigarettes. Veo is always bad with cigs
how i hate this moments
i have client that i need to deliver video like 3 hours ago
I cant explain them that its not my fault 🙁
Still no luck deleting old projects
and higgsfield Veo generation works (not perfect, there are few failed generation) but 3.1 fast works there
but for like 20 videos i'll spend all my credits so no way
Omg i just got in this discord, so freaking glad that i am not the only one. I was stressin out the entire day.
That's usually teh case.API and app have to use different servers.
yea veo is very expensive everywhere except Flow
Here we go.
Lite mode. Just doubled down on the style:
"Art style like textured 1950's movie poster paintings, rough brush strokes, with chalky texture."
is google back up yet?
its hard to prompt veo to create those... hand drawn 2d animation style
me Too
google pls
do it with "lite"
None of them are allowed!
yes!!
..
I don't know anything about cartoon styles, but there are tons of 1950's movie poster paintings, so that's a style Nano Banana can do easily. Veo requires more reinforcement to keep it beyond a few seconds.
Something is something
Wonder if the Gemini canvas can also do a video website that works, that's not my forte so not even gonna try
Eitherway whoever's already paying for Google plans also has access to Gemini tools with image and videos, so go nuts there too I guess.
Hey guys I think it’s back
Yeah it’s back man it works :3
Tested it out on my phone and iPad
Is this news accurate? Are all the features working?
For flow yeah the nano banana 2 is generating and not giving that error again
Haven’t tested the other ones
the monkey. We're talking about video
Idk but the image generator is working
Nah I think Nano Banana 2 is still bugged, which mostly started when some genius tweaked the censorship.
@frozen sand They’re still fixing bugs region by region. My Flow still can’t generate video from images, though it still work when using ingredients. Try switching to VEO Lite. It did work for images, though the quality is ...😬
Just tried this. Didn't work :/
Please tell me I'm not the only experiencing this
It's really frustrating to not be able to drag my images in projects anymore
why even have an ultra sub if its gonna be down all the time lmfao
bro exactly, i could understand if it was like 30min or an hour, but an entire day?? bruuh
Quick Hack: only Veo 3.1 lite is gnerating videos. 👍
really bro
Im also facing same problem, Only veo3.1 lite version is working
Any guarantee when it will be fixed? Do they have no team members here?
There's also visibly been a decrease in quality with nano 2 images... not just with art beauty also the composition coherency dropped.
How do they manage to constantly make it worse, I can't imagine what pinhead's feedback are they following on all this. Doubt it's user feedback.
Well at least Nano 1 can save the day now... not really easily or perfect, but thats what there is.
Is there any safe way to share my flow account with my editor whos located in Philippines? I'm in europe, and in order to have "family" accounts, they need to be in the same country right?
When will Veo 3 be resolved?
I think you sent too many request at a time
is there no admin here? where veo 3 fast mode back to normal?
a dedicated VPN might help
Please help me revert to Veo 3.1 Lite so I can use it. I'm currently using Veo on the web and it keeps crashing.
... whoever made all last weeks changes to Nano 2 and Flow in general should be fired tbh. Dude's about to sink the whole Nano and Veo projects.
If no one's gonna be around to fix things for hours and hours, then they should leave them as they are while they are workng properly and someone can be around. That includes Whisk and everything else that actually still works, if there's even anything else left.
previous prompt is acting weird
same
even the slightest resemblance to any famous person gets rejected now, and I didn't even use Flow for anything "spicy"
basically Flow or Banana, whatever is actually doing the moderation, is now high-end fiction generator, next thing are to go are probably brands and anything resembling a brand
flow is not working...
it was just an hour ago
still not working
Me too
just select it from the models list.
Another funny thing is my Deviant had a tier where I was selling the prompts used, I had to change them all cause of the snowflake who changed Nano 2's behavior, which again was for nothing since all prompts still work the same after tweaked. Just a waste of time.
Not really the first time (this year even) when this company arbitrarily changes what they already sold as is, and people visibly thought they could just keep using in peace cause they payed for it.
Like Whisk is so full of people's info it's not even funny, who knows if Flow will go through the same someday.
Does veo 3.1 lite start frame end frame working?
It's pretty normal
mine it’s not working not the light not the quality only the start and end frame
lite cant use extent is it true?
anyone else getting stuck at 99% to "Failed -The generation might be taking longer than expected. Please check again in a moment. You will not be charged for failed generations"
same
Good morning everyone .... still not fixed ?
@main saddle when will be fixed ?
99% error 😮💨😮💨😮💨
🤬🤬🤬🤬
Did you try with veo 3.1 lite, does it working?
Sometimes
stupid question: does flow use our pc/mobile resources to generate images/videos? because every time we get an error message we have to clear the browser cache so it can work again? Even if I only change one word in the prompt, it will trigger various error messages.
I think, No..
Hey guys
I have an error: "Failed -The generation might be taking longer than expected. Please check again in a moment. You will not be charged for failed generations"
welcome to the club! 🥳
<@&1366831844502274180> any information about the error?
idk how you guys keep having errors, they were solved long time ago for me
maybe it's a regional thing
ive never had this, this is the first time for me
Hey, if you have details about the error or think it’s a bug, feel free to open a thread in #1238621511401734254!
ok thanks Rain, posted it in the bugs channel
Being fair they did say accounts would be slowly going back to normal some before others.
But yeah the nano 2 model is still ruined
hi google maybe you should fire whoever made these changes on flow. everything is good yesterday and all, of sudden we are getting error.
Hi @random temple, May I know what region you are from?
People lost so much money today and yesterday, these are indeed used as tools for work not just to play, so there really should be consequences.
They should really be more serious and professional.
deadline waving goodbye like
Now I think the Ingredients mode is working?
bro, please fix Flow
i cannot generate any videos there now!! it's been more than a week now
What the hell is the google team doing now? Why they cannot fix this?
if you are having issues with the 99% bug please upvote my bug report in the #1238621511401734254 channel for visibility
Hello Guys, am i the only one, who has problems with the Image -> Video generation?
mine is not 99% bug
Oh yeah i see, other people here have the same error :/
it just fails on every generations..
what is your error message? failing because of what?
unusual activity error
Are they any alternative tools?
kling?
hows kling for images?
some admin open a slowmode for...... idk
lol
what do you suggest for images other than banana?
long time ago, I using midjourney.... but I don't know it's still open or not
btw, thanks for veo down today, that's make me have nothing to do in my work time lol
stil down?
for me , yes
but my friend said, can using an ingredient mode.... but I need to use frame start in my work lol
@main saddle @tough yacht @shell gazelle when will we get flow back ? we wait since 1 day almost
we still cant create our videos
please fix this
please compensate, free flow subs for 1 month for everyone
10k tokens would be nice
Damn
I thought not being able to restore deleted content was already bad enough.
did they fixed the censorship issue yet?
Video generation still works in the chat app though
but only 3 videos for the pro plan
mine too isnt there any estimated time for updates?
scene builder not works
😡 😡 😡 😡 😡 😡
Hope they fix it asap
cant i just use google vids and its veo portion instead of veo in flow im not sure if the veo in the vids is unlimited tho
which chat app? mine keep failing
จะได้กี่โมง ตั้งแต่เช้าแล้วพี่
You're requesting generations too quickly. Please wait a moment and try again.
i'm gettting this error for some reason and as soon as it comes up it will never let me generate again
ok now upscaling doesnt work gg
everything is melting down
Howdy friends, while veo seems to be having a lot of complaints I've downed tools and started writing some chapters and scenes and making plans for settings and stories. I know it sucks but it could be fun to write some stuff in the mean time.
Wow, it's working now!
you sure?
i think its coming and going, veo lite seems to be working. might be a case of overload idk
It’s working now
hi all, we’ve issued some more fixes for these errors. please note that they are rolling out and can take some time to implement.
please let me know if things are working for you (👍) or not working for you (👎 ) by reacting to my post here.
@frigid piersame
Hi Georgia, I've had none of these today! (at least not yet). THANKS!!!!
🤮👎
this is so stupid
Same
Would be nice to have a TEAM subscription, where more seats could be added, can't believe it's still not a thing
het guys ...
"Google Flow" is really messing up and doesn't stick to prompts.
Even upscaling and improving the image quality doesn't always work.
Meanwhile, editing in "Google Gemini" chat works exactly as I asked, it instantly understands and does what I need...
What's going on ? Im working at "AI Pro", but the quality is worse than in free Gemini chat... And the Flow interface won't help here.
What should I do ?
Use Gemini
@woeful void Hey there! the google labs team are aware of issues happening, please refer to georgia's post here. You may check the pinned messages as well to see if your issue has been mentioned, if not, check #1238621511401734254 and if your issue is not in there either, feel free to make a post following the guidelines pinned in that channel
it has watermark and has no 2k quality
use em
When you create your image at Gemini ensure you download the image with it's download button, not right save click as, then this watermark remover can take the gemini watermark off without losing quality. https://geminiwatermark.org/batch-watermark-remover
It can do it really fast and in batch.
Downloading it that way will also give 2K quality, I just tried, tho you can also use prompts to improve it, but by default the quality won't increase more than that.
Generations appear to be working for me now 🙂
Experimentally, the Gemini canvas mode can also create a 4K upscaler for you and it will indeed work, you just gotta ask it what you want. Sadly I haven't been able to make the canvas mode reproduce a proper Nano Banana 2 interface, it keeps switching to Imagen 4 with it no matter how I set it.
But it for sure dam worked to create an original Nano Banana and Imagen 3 UI. F the complaints about it.
The canvas mode images don't have watermark either... so yeah that'd be nice to have.
More likely that's also the secret why Whisk and Flow images don't have watermarks, once you create an image generation website with Gemini the images don't have watermark anymore.
What is this model doing here?
I think you can only select it from the main page... must be good ol' Gemini.
0 credits😄 💯
Also I don't know if you know this, but Gemini's Nano Banana 2/Pro can also control the aspect ratio by prompt (must be reasonable aspect ratios tho, nothing like over 9000)
You can end your prompts in something like this and it will obey,
Aspect ratio is 3:4
Etc.
That'd be nice to have at Flow but could be a little complicated to separate it from the video interface.... that's what I wanna tell myself as why they don't
But yeah if the Flow website was created with Gemini which more likely was, it's even sadder they don't just update by prompt the necessary things or at least the ones people ask for, or change the error messages so they state their cause.
It's totally doable.
Hey bro what is your alternative tools for google flow? What website?
anyone know if there will be a veo 4 anytime soon? 😏
Probaby the only contenders are Gemini and Google AI studio. it's less known but Gemini and Google Ai studio can also generate more images at once by prompt. But Google AI studio can't generate Nano Banana 2 images without buying an API key which is a torture to do, it's default is the original Nano Banana that Whisk uses for precise mode.
What's happening with FLOW nowadays? - always BUG
Also yeah both give images the Gemini spark watermark, read above on how to solve that.
The other way is to create a website with Gemini or it's canvas mode, but if you want Nano Banana 2 is a little more complicated, I haven't gotten it right. NB1 works tho, so that's a bless already.
hmmm I hit extend on a video and I expect it to generate a video containing the previous one and the newly generated one ... i.e. making a video longer than 8 seconds ... the extend generation only generates another 8 second video that starts at the end of the last one ... did I miss a memo on the scene building or making videos longer than 8 seconds?
And yeah then there's the tons of image generation websites that charge image generation with credits, but I haven't found a single worth one, but if you test one and works yeah you can pay there if you think the cost is worth it, more often it isn't.
Oh right, gotta remember if you already payed for a Google plan to use Flow, it already connects your same Google account to Gemini, don't need to pay there again.
How about for image to video? Any recommendation tool that have a start and end frame?
<@&1173646484311384114> IS EVERYONE HAVING TROUBLE CREATING VIDEOS IN FLOW?
being honest and considering all the current problems really don't give much other options but to say it, some people just seem to end up using other video tools alike kling...
but if it's for VEO I guess the recommendations are around the same. I haven't played much with VEO video generation with Gemini's canvas mode, but I'm sure it can create a very similar or better website if you spend some time trying and you're detailed patient about it.
I just generated this at Gemini's Veo which is the same as Flow's really, and this kind of turn animation is one of the most accursed things to prompt correctly in an image generator, and yeah VEO can do it.
"Make her naturally turn 360 degrees only to the right until completion. Dont add more items or props."
However a plus account won't give you more than a couple vids there at a time or takes it really slowly or something, you better be on Ultra at Flow too or whatever
im thinking about to switch to kling ai, just because of the downtime / errors here,,,
Does any one experience degraded image quality on editing an image more than once? The more edits, the worse it gets. I don't mean, I attached the image. I mean, I use the pen tool to mark an area I need changed, then prompt for the change. It changes the image, but the quality degrades. Each additional edit, it looks more and more like a AI Images from 3+ years ago.
In whisk I could use the seed number and prompt for changes. The quality remained the same each time. It was easy to prompt image sequences for first and last frames this way. but now, any image beyond the first is nearly useless. Any suggestions for making things better in flow?
agent...???
To the Google devs (Speaking as long time cloud software engineer myself): can we get an RCA (Root Cause Analysis, including the 5 why's) on this?
nope it works fine
They should ... they're a cloud provider ... most cloud providers do provide RCAs for widespread outages like this
I've written RCAs for the platform that I develop for
lol I got this too it's odd
Too many users not enough resources. Not just Flow, even LLM backend is down: Engine Error: Engine Error: [GoogleGenerativeAI Error]: Error fetching from https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-pro:generateContent: [503 Service Unavailable] This model is currently experiencing high demand. Spikes in demand are usually temporary. Please try again later." Has Google gotten too greedy? Anyone else paying Ultra and getting this?
Some accounts and locations seem to have gotten the video generation fixed and others not, they'll eventually get to everyone, but yeah if it isn't fixed yet might wanna go to labs-bugs-feedback and state your region there.
Flow just seems to keep insisting in biting more than it can chew
"You're requesting generations too quickly. Please wait a moment and try again."
I waited an hour to create 4 new images and this pops up. Been like that for last few hours.
If it was working fine some hours ago maybe you just hit a daily quota for images, Flow sends the wrong error message for that, it should be saying something like "You ran out of image generations for today, come back tomorrow 🍌 " or something alike but instead more often says you're going too fast. You should try again tomorrow around the same hour it was working fine the last time.
Also you can check if swapping from Nano Banana 2 to Pro or vice versa gives you some more images. Your Flow account runs out of Pro images much faster cause they cost more.
Yeah at this point I'm just pasting this, however with all the error messages sometimes even this is not it lately.
is it indicated anywhere or is it just more or less feeling?
Nope, you'd know according to your Google plan, for free accounts the images can be more or less 30-40 for Nano Banana 2 and like 10 for Nano Banana Pro, according to the day's demand.
Plus gets double at least for NB2 or even 100 sometimes.
That if you even manage to survive all the current bugs and the new levels of censorship, but not all accounts seem to be suffering that much.
I would go pro if they fixed Antigravity and promised not to kill the limits
Also remember your same Flow Google account links to Gemini, you can scroll up for all the chat about how remove watermarks there and all that.
It's one payment in both
You get some more images and really few vids there if nothing else to do atm
right, ty for the information
But the real fun part is the canvas mode, you can use it to try and generate the exact website you need alike Flow or Whisk, 4K upscalers and other things, it's really underrated.
Whatever prompts were used to generate these Whisk and Flow websites surely work there too and one can keep building on it. If even interested.
Like yeah I'm pretty sure my wonky NB1 canvas works.
Hey all. Question: I'm making images and then making videos out of them. But Nano Banana Pro when it makes the image, where I save my last frame as the thumbnail, it ends with a .PNG format which is great but my DaVinci Resolve Studio won't accept the file unless I change the end of the extension to .JPG so is anybody come across that before and you know if there's a fix I guess I'm trying to figure out if it's something with flow or I need to contact black magic for da Vinci. Thanks.
Happens to me all the time too, best you can do is get an image editor to re-save the file as PNG or whatever, this one's lightweight easy. https://www.irfanview.com/ It also has a batch mode to change them all at once by dragging and dropping them to some window, it's not that easy to use but like you get used to it.
IrfanView ... one of the most popular viewers worldwide.
That's my reason for the question. Thanks. chatgpt said there's a browser extension that will do it as a save as but I'm kind of leery of extensions
Flow seems to be saving the wrong extension after upscaling, so yeah that's a bug everyone's too lazy to report, me too.
Thanks. Yeah, Paint does it as well. This machine is running on Windows. My other 1 runs on Linux.
There's the batch mode thing somewhere, saves time
faststone image viewer has batch too
too quick generations. please wait again! cant even create 1 image and video per hour as ultra user 😡 😡 😡 😡 😡 😡 😡 😡 😡 😡 😡
I'll try that
Thanks, and yes, But I'm mixing assets in a folder before producing a video, and I can go through terminal and have them change the PNGs over, but then some of those don't work either if I took screenshots of something. It's kind of goofy. I don't understand whether it's Flow or if it's DaVinci Resolve. It seems like something's missing from the file, which leaves me more to think Flow. I haven't reached out to Blackmagic for DaVinci yet.
I can do powershell: Dir *.png | rename-item -newname { $_.name -replace ".png",".jpg" }
Yeah I've seen it, it just saves the wrong extension, it creates some sort of hybrid invisible jpgng or whatever. But resaving fixed it for me, might work.
Flow's bugs are comical, shoulda gone for a human created website not something generated, pretty sure that's what's going on here.
Well, I did a search in the bug group and couldn't find anything. I'm throwing several terms in there, so... I created a post. We'll see. Thanks, everyone. Appreciate you.
Like irfanview does warn you that this is going on, this is the error message a 2k image upscaled from Flow gives at irfanview. It's almost undetectable if you don't have an image editor.
Says it's a jpg saved as png or whatever in the hulkamania is going on here
That's the image unchanged, says it's a png but it really isn't, but it is.
DaVinci has a community forum, but apparently no Discord. This is after I change it to a .jpg and then changing it back to a .png still won't import. 2nd image is changing it back to .jpg It's a minor annoyance when I'm doing the time-lapse ones, but on my long form videos where I might have a 100 some assets to create, it can be burdensome. So I'll just make a change in my workflow and add a second folder just for flow, do the batch conversion in the terminal and call it a day, I guess. We'll see if Google gets back to me about the bug report.Thanks again.
haha... Gemini figured it out. So, yeah, Flow is creating a .jpg file, but it's calling it a PNG. That is definitely a bug.
yeah but still resaving them as JPG should have fixed it, so that's weird
Who do we even ping for this, not gonna rewrite it all
No, renaming them 100% does fix it. That's the second image. So when I imported it, the bad file appended to the end of the clip. And then when I changed it and recopied it over into the timeline in its correct position, it was fine, which is the very first clip on that timeline with no sound. I need to have that file because it's a frame I can choose as a thumbnail on these.
I posted a new thread in the Labs bugs feedback channel, but who knows
He's the one I've seen tweaking things around @iron matrix , sorry if I bother you man, but might wanna look at this, it's at bugs feedback too already. Happens when upscaling or seems when downloading images too.
thanks
hey guys whats the limit of videos per day in gemini with ultra plan
laughs as free user
Hi everyone, is everything okay today? Has the Labs team fixed the Veo3 issue yet?
no
https://www.facebook.com/profile.php?id=61573258261962 hello help me please follow my page please 🥺🙏
See posts, photos and more on Facebook.
no
no, too
It's an AI page, be good with your brother/sister.
Like everyone hates Ai, we need to be together we only have ourselves.
what is AI page
Can an AI page do kickflips?
is flow generating videos now?
for me, I got 10% chance that generating and 10% of that is good.... so I can use 1 output from 100 generations
🫠
AI means love in Chinese. Why are you into AI and don't know Chinese.
And yes this is relevant to Flow because Chinese are kicking our derriers in AI.
I think Nano Banana 2's censorship might have gone back down to tolerable levels.
thank you for this knowledge shifu
Or at least it's leaning more towards doing what it wants instead of what it's told to
Lemme go try some Disney stuff
Nah just my idea, Nano Banana 2 is still as censored almost unusable as 2 days ago, got confused cause my new prompts still do the exact same.
But I think pretty much anyone else also figured out how to reword things back, the censorship is only trolling the users who're just trying to do everyday images.
It really doesn't help with all the other error message problems.
Things so quiet now, I imagine half the Flow users got the video generation problems fixed, and the other half who didn't get them fixed just went to find some other generator.
Laughs in original Nano Banana canvas still doing even better NSFW than NB2
Yeah just tested the canvas prompt and still works fine. I think one can even add a gallery to it.
Gonna go update it a little
Flow kept showing this and didn't allow me to generate any single image from yesterday. Previously this problem was solved using a vpn but now vpn even can't avoid this. Any solution?
I really haven't experienced this one yet, is your account free, plus or whatever?
It's pro
my nsfw content generation fails at nano banana
In what website
I have 3 pro accounts and have same issue in every accounts. Is this server issue?
gemini app
dude
Gemini uses Nano Banana 2/Pro and it got censored to the point most likely we should be forgetting about it.
Exactly
You will have to reword everything from 0 again if you want the same prompts to work, and it can be done for sure.
Tho it seems much easier to just Nano Banana 1
then what i use shifu
Like said start the prompt from 0 and instead of what you always use, use a similar term. First try one sentence, don't try all at once.
Then build from there
manipulating gemini i love that
I just got tired from it anyway and just made one of these like Yaska does, I'm ready for when Whisk is taken down, don't need it anymore
Nano Banana 2 doesn't really work well on it, but where does it atm.
They really destroyed the model
This below is NB1 tho.
Man it really makes them 4K really fast
This is why we can't have nice things
Google why don't we have nice things, this is right there.
Listen to feedback. Flirt back goddamnit.
😄
Really when you post a gif it's like picking the X on Tic-Tac-Toe. What is one supposed to do man. Nvm it's gone.
Well this is just to play, gotta try real prompts
jfc
asdsfafasgfasgaf
It's aweseom what the hell.
lemme try and make a safe one to show
... it's so much better than what NB2 was giving me even before the censorship
the flow is goods now? or sill not working?
I have a "workflow" Flow question:
you are working a poster with a character depicted on that poster
but every time I do "add to prompt" and "say make his face took into the distance"
it generates new pictures
the thing is the pictures have a slight variation to them
so the question is:
if you found the result you're looking for but it needs a little tweak, do you edit the one picture or do you make the tweak by generating new ones based on the old ones?
is there a rule of thumb here?
any update?
Edit the outcome, instead regenerate a complete difference to the new one. You can go back to the original anytime with history.
right, so like a backup in case the new ones don't generate the desired result?
They had a history tool, you can instantly walk back to it. No need to regenerate a total difference one, just edit the one you think needs a bit of work.
ty
With the red marker annotations tool, you can tweak each detail to your perfection.
I think that's what the "seed" option from Whisk would be useful for... even tho it really wasn't actually seed but some prompts.
is there something like a "system prompt" or "this-project-only rules" analogue for flow? the coding agents have these in the system prompt and in AGENTS.md
I read something about putting .json files into somewhere, but the only place you can put them is prompt
oh my god the flow is still not working until now?
if tell Flow to generate 2 images, but one of them "might violate our policies" (on an 100% SFW image) - does that rejected generation count toward limit?
Hi guys, just a quick question: does the Ultra plan still include unlimited Veo 3.1 generations in low-priority mode?
Why do I keep getting error messages?:Failed
We noticed some unusual activity.
then you should check the log first
Hello, I’m currently seeing the following error message: “We noticed some unusual activity. Please visit the Help Center for more information.”
I was wondering if anyone might know what typically causes this issue.
I contacted Google Support, and they told me that Flow does not have a dedicated support center and suggested that I ask the developers on Discord instead. However, I’m not entirely sure whether that is the correct process.
If anyone has experienced this before or has any information, I would really appreciate your help. Thank you.
search this dicord for "unusual activity", maybe you are doing something they don't like maybe this is just a bug
Thanks
Someone said that logging out and back in made generations work again, so I tried that, but it still didn’t work. It’s strange.
I haven’t used a proxy or VPN, I haven’t been generating excessively, and I haven’t violated any guidelines.
I don't have the "suspicious activity" but most of my generations fail, all of them SFW, god forbid a woman on the poster has skirt showing a knee, heresy of highest order
over-censoring leads to people spamming requests which puts a strain on the infra
I haven’t used a proxy or VPN, I haven’t been generating excessively, and I haven’t violated any guidelines.
same here, there's talk of a lot of bugs, there's talk of the censorship, there's talk of decreased (again) limits
similar type of problems are happaning in Antigravity for like last 2-3 weeks
when that type of disruptions happened on GLM (back in 4.7 times) they were training GLM5, at least that was the consensus
sometimes they just warn you about suspicious activities, nothing to worry about.
Thank you all very much for your input. After logging out, completely closing the window, and then logging back in, it seems to be working again.
It’s really unfortunate that even as a paying user, the service can be this difficult to use.
Hello. I write only Prompt: "Christmas party in villa and generate 6 images = 2x3 images and I get on 5 images this information. Only "1" image was created. What does it mean? (I used Imagen 4 and NB2 too. Both generators do it)
To fix this and successfully get your full set of images, I recommend trying the following adjustments to your prompt to make it more specific and to clarify the "G-rated" nature of the scene. You can try resubmitting with more explicit details like:
"Christmas party in villa, group of diverse friends laughing, fully dressed in cozy winter sweaters, drinking cocoa from mugs by a decorated tree."
"Close-up photo of a Christmas party in a villa, people giving thumbs-up, wearing Santa hats, by a dinner table with a turkey."
This time I created again 2x3 images and now I have 3 images created, but still 3 images are same information problem.
Before I normally, created 9 images.
I tried 1st prompt and i got 10 different images
funny though
well the pricing page doesnt say anything about generating NEW pictures, lol
I have this same issue
When I use proxy servers it works for sometime then again it shows the message in all connection
idk why this is happening
with u
I thought I am alone with this problem.
proxy you mean vpn or third party api resellers?
I was wonerding if maybe the resellers have better results
vpn
I had this kind of similar problem long ago and back then the only solution was vpn so I thought it will work this time too but I was wrong
vpn helps u sometimes, u can't use it regularly.
I tried now by Imagen 4 and with Imagen 4 are completelly all images with this problem.
NB2 create minimal 1, sometimes, but Imagen nothing
ohhooo
just hope for the settlement of the problem
I never got this kind of problem on Whisk, even it generates images faster than Flow
Interesting, now I write only "Christmas tree" and created 2x3 images and only 1 has this problem. I don´t understand it.
Flow sucks
By Imagen 4.
bad day for u 😆
Is flow working good on your end?
yes perfectly..
Umm if you generate images constantly like 4x then again 4x, don't you get any error or any failure?
nop, may be that's because I'm not using any proxy servers or vpn
I see
My day is also bad, TBH
Yes, I get information, that I generate image quickly and I can´t create images some time. I get it yesterday, before day etc.
any updates to flow guys?
Same goes for me too, I have to wait for like 2-3h to work again
feel for u
I thought that this problem only I have too.
great then enjoy that waiting time 
You haven´t this problem?
It's not funny, Only 1-2 generation between 2-3h? This happened before using of proxy, I used vpn to find solution. This problem showed on first phrase before using proxy
nooo
you are lucky man
I'm not using any proxy servers
I think of all us are lucky except me and Pat222
Yes. 🙁
u mean except ...
U need more patience
hahaha
😆
Hello, not sure if this is the right place for this, but I posted in Labs-bugs yesterday and had no answers.
Is anyone else having problems downloading clips?
It seems like the button is not working on desktop chrome or edge, yet it works on my phone.
Any tips?
I am using other browser than Chrome and Edge, but I tried it on the Edge and Edge download my image.
I'm also using edge and its working perfectly.
Yes, it does download images. just not the videos. 🙁
sometimes chrome creates problem.
I am not using Chrome. Other broswer I use.
you can try private browser in edge it will help you.
i think...
I want to ask that the last pictures of the cup with eyes were not created in Flow, right? In flow-gallery. Because on the right down corner is white star and images in Flow haven´t this "star".
nothing is working suka blyat
Still not working. I've also tried to clear the cache, but nothing. It's so weird that it does work on my phone.
which image u r talking about ..
I have created these 6 images on flow with that prompt i shared to u .
I can´t. I still have same problem.
give me the prompt for this let me try..
I haven´t prompt, it created other user. I am only asking, why there is star. 😄
wow ...
Hey guys, anyone also getting ''You're asking generations 2 fast,wait a moment and try again''
I've waited 24 hours 😆
why u all getting so much of problems ?
I believe it really is a cookie or cache issue, just cleaned all that and it worked
congrats u got the key's 😆
are they gonna fix the censorship, any news on that?
any updates ?
Man those are some creepy uncanny-causing netflix cg animated woke show styled people. Why are the women as flat as the men, a sweater doesn't cause that.
Nano Banana 2/Pro is ruined anyway, not worth the time
And it's believable two of those dudes are twins, but that NB2 would cause that by default is just sloppy
I swear it wasn't that bad last week before whatever goofball tampered with NB2, it would at least listen to prompts about esthetics
They really need to roll it back to how it was last week before it gets worse
So google are leasing a significant amount of their processors to other companies, some of the majors being apple and anthropic, even amazon. Recently there were articles about the load being extremely high meaning they throttle usage and prioritize the requests based on services. Also if you're using image refs of known IP's like gladiator or Frank Millers: 300 or even screen shotting youtube content you might get flagged for copyright infringement as we're not allowed to use existing IP for training models or using them. Even things like stock libraries and sound effects or music libraries have these clauses. Using words that may be delicate like FBI agents or federal bureau of president or even company or organisations like NASA have copyright over logos. These are all some factors to consider when renders fail. Sometimes they also fail if theres potential for violent or socially sensitive subjects.
Didn't cross my mind it was affecting the audio from vids too... that would explain so much since VEO adds music where it pleases... it's just pulling it's own rung
I think if veo generates the sound and music its cleared for copyright, but certainly if we're uploading references it will get flagged.
I bet at least at random it would still get something it shouldn't, it really ain't beyond Google atm to just press a button, not to check the side effects and just go for coffee from friday till monday evening...
True enough, Its a big machine with millions of users. A lot of processing power needed to run automated scans and catch everything. Considering how much youtube content gets uploaded every day, there arent enough people to watch and review everything.
Have you ever had this problem? For 2 days bananapro has been telling me that I ask for generations too quickly, please wait, then try again
I feel for you man, try funning one single one and see if it does just one?
Same problem
damn
Have you ever had this problem?
I usually take a break and come back after a while. I've had it often enough.
I think the general chat had a thumbs up and thumbs down from one of the devs/mods to see who's having trouble with their renders yesterday. Scroll up and maybe check with them?
I have not been able to find official guidance on resets/usage limits/knowing when usage is re-allowed. Has anyone found it?
Should be able to reroll any time so long as you've got the credits. Failed renders shouldnt use credits.
My god didn’t find any alternative for 1:1 pictures x4 at the same time
Gemini is 1 by 1…
Flow is free and require no credit for image…
I think theres high usage and not enough processor power or other reasons I mentioned above.
Pop it in the feedback and incase theres something sensitive in your prompt put your prompt into gemini and ask if theres any reason this prompt might fail to generate in google veo/flow
Its been good at flagging things like that for me.
I think I maxed out usage on Pro and 2 (Nano Banana), just trying to figure out when it will reset
Where can I ask/comment for my issue ?
next billing cycle, though if they take the money on the 1st of the month youll get the credit refresh ont he date you actualy subscribed.
It seems like it maxes out from what I've seen/heard. The "You're generating too quickly" goes away after ~24 hours in accordance with the reset idea
Yes, except it's not attached to credit usage
through many errors
Exactly why I don’t understand
Also I've seen when generating the unsolicited music for the videos, VEO uses the prompts for the lyrics or tunes too, if the prompt has something about "Turning around" the lyrics are related... so there's so many words that could link together and cause things
oh right, yeah not sure on pro, I used to get that. Maybe ask the question online.
Nano banana 2 is so slow compared to Pro
Where ?
Google search uses AI mode, should probably find the answer for you pretty quickly
How to do that ????
standard internet browser search
IA is on ?
might depend on your location, but yes, google search comes with AI mode now
I think this says it's the "you're generating too fast" problem, was it working fine recently?
Yes it was working perfectly before that
Reddit comment :
Please help, I have problems with Google Labs Flow AI (Nano Banana 2 and PRO), I always get this error
"Sholed You ask for generations too quickly. Please wait a moment and try again."
No, I wasn't asking for generations too quickly.
I waited a whole day and the message does not disappear.
I have already tried to clear the cache / cookies and switch to incognito mode, it did not work.
I already tried to use another browser, it didn't work.
I was using a Google PRO account.
I have already deleted ALL the projects, there is nothing left.
If I try to use another Gmail account (free), the problem disappears, which makes NO SENSE and makes me think it's an error related to the account/server.
This first happened for Nano Banana Pro and then it started happening for Nano Banana 2, now I can't generate anything.
I already contacted Google support but it was completely useless, the guy told me "to wait" because it could be a "temporary problem". I don't want to open a new Gmail account/pay every time this happens.
MAKE AN EFFORT GOOGLE
Just gonna paste my thing then
If it was working fine some hours ago maybe you just hit a daily quota for images, Flow sends the wrong error message for that, it should be saying something like "You ran out of image generations for today, come back tomorrow 🍌 " or something alike but instead more often says you're going too fast. You should try again tomorrow around the same hour it was working fine the last time.
Also you can check if swapping from Nano Banana 2 to Pro or vice versa gives you some more images. Your Flow account runs out of Pro images much faster cause they cost more.
Have you tried cleaning the cookies and cache? Also try private chrome/edge browsing, had the same problem
Nah if you translate that message it says he's going too fast, it's not that other new message about suspicious activity
@zenith thistle
My bad
It doesn't matter if you come back in 4 hours or 5, it usually takes a whole day
If I pay for pro it should change something ? I think no because image use 0 credits
So there's no coming back now, huh? I think I kinda fked by buying ultra solely for nana banana 2.
Do you know an alternative to flow? Who generates several images at the same time with the same prompt using nanobananpro?
I dont think so either, prob is a bug or server issue going on
Payed accounts do get some more priority yes. But it's not unlimited either or fail proof either, Plus gives around 40 Pro images and 100 NB2 images. Pro shoud give around the double more or less.
Once the account runs out of quota again you will still get that same error message again for the day.
I generate more than 40 images per day in Free
Yeah NB2 can do that, the daily limit changes according the day's demand
But it's around that
If it's getting 100 for free accounts atm, Pro will surely get more before you hit that invisible wall
Where can I find the daily limit for each subscription plan ?
They never said it, we all just have to guess everyday
But every plan have more image right ?
It happens when it does, you just gotta know that's the cause and take it easy
Yes for sure it does
Because I had only the 8€ plan before, ended 2 days ago
I do have just a $5 plan too and haven't hit limits yesterday
Higgsfield ?
I can imagine it changes enough that they don't want wrong expectation setting and/or to be "held to it"
Maybe they did increase it
True, just go for stable difussion localy for that. I dont see issues with flow restricting this kind of content
Also there's this new "Suspicious activity" error message that might happen when you're generating too fast, not even payed plans are free from that, you might not want to generate more than 12 images at once to avoid that.
Some people might be using batch image makers or autoclickers, but I doubt it's everyone
I've gotten my questions answered, had plenty of interaction over my time with people here. You expect more that you aren't getting but thats life bud.
and most people in the world don't use discord it would be silly for them to make their customer service network on a platform people don't use
Why does nano banana pro and 2 for that matter simply restrict you generating images of someone in a two piece swimsuit? It's a swimsuit for crying out loud lol
if you add in image ingredients, even if they are nano banana pro generated images, in the image generation and you put in "bikini" or "two piece swimsuit" terminology in the prompt, that's totally a no go 🤦🏻 .... it's not p*rn lol!
I now can't generate a dog or a cat picture ... the censorship they turned on this week is actually insane I'm getting "Policy" error every single time
every i generate image i get oops something went wrong. i clear cache. delete all cookie and do it again i can generate 2-3 image but after get wrong again. firefox user
I'm just prompting a simple golden retriever playing with water in the garden ... and I'm getting "policy violation" ??? This is insane for a product we are paying big money for.
Try bra and panties content it doing the same to that one. I wish they fix and remove this. This getting ridiculous.
Post the prompt
Tried a bunch of prompts ..... the last one I tried right now is " a golden retriever dog playing with water in the garden" and I still got "policy violation" I'm sure I'm not the only one suffering from this for the past 4 days
Can I see the full actual prompt
Or is that really the prompt that you entered
Hard to believe lol
Yeah exactly that nothing more nothing less I can assure you that
Wait hold on
It will work for you I'm sure, I'm telling you I tried so many different prompts ... then tried this last one and confirmed that the problem isn't from my prompt since even a simple one doesn't want to generate
Guys im trying to generate an image in banana pro, and it keeps saying this error, why is that?
limits reached.
What was the actual prompt you entered? I’m so confused
How do i know my limit is?
By getting that error u showed
Try counting how many images u gen today
Til u got that error
Since when was this a thing? i used to generate as many as I wanted
I thoguh there is no credits needed
Always has been
But
I use these for my graphic designs, my boss will not like this
tell your boss to cry me a river ,
Jokes aside, go create a new disposable Google acc if you really needed to use the image generator urgently
Good idea
Are you primarily focused on generating images or editing an images
generating stock images
got it
How to get fired speed run
so as an alternative way for now, try https://docs.google.com/videos
Google Vids is an AI-powered video creation and editing tool. Use it to generate Veo clips, record your screen, convert your presentations to videos, and more. Access Vids with your Workspace account or your personal Google account.
on the right side, click image
or you can use Google Slide to generate image
is that using banana pro or 2?
Access Google Slides with a personal Google account or Google Workspace account (for business use).
actually for google vids im not sure
but google slide uses nano banana pro (according to screenshot above)
yea, its getting on my nerves now
Hello everyone, I hope you're all doing well.
I just want to know as much as possible, just your honest feedback on my video. I know that you can say that it's good but I want to top it up a notch. I want to really make this even better. I think it's based on my image generation to be ultra realistic. I feel as though that this is the best that I can produce and there are maybe other things that are missing to include with my prompting to help me amplify more of a realistic ultra realistic image. Please let me know. I'm going to share with you my video, plus my prompt, and my image as well.
I've filed multiple concerns about this on labs-bugs-feedback because after hundreds, if not thousands, of prompts, clips, stills, etc, I've never seen the policy filter so broken. The policy filter team needs to resolve this because, as creators, we cannot create without knowing what the mysterious word in our prompt (or combinations) sets the thing off. I even have a Gemini AI assistant (same company, right?) that can't figure out the reasons for the rejections (not that it hasn't come up with every idea in the book to try and get a simple prompt through-> I have tried everything it recommended, and it is stumped too :).
(and I only ask for a generation maybe every 5 or 10 minutes... so it's can't be for asking for too many generations. Maybe I have asked for 15 today?
and i pay for ultra just to play cat and mouse games with words with this stupid game with thier filter, and now it wont even genreate what i want.
What are you trying to generate?
Hi Georgia....This is back in spades again today.... 🙁
Just assume Nano Banana 2/ Pro is only for cute safe chibi character tea party images now... like there's this user around here who literally only uses it for that tho lol.
Even if they're probably trying to please certain public, it doesn't seem they should worry about them as much as they should worry for their actual customers.
One would think they might have concerns about getting sued and alike, but that kind of stuff doesn't seem to worry them that much either, alike when they suddenly placed watermarks on the ultra plan videos.... which also was a contradictory thing.
Who knows who they even actually listen to
Everyone's asking them not to take Whisk offline or at least add features from it to Flow, everyone complains about the filters being too exaggerated, everyone complains about the error messages, and it's been days of the same already.
The video generation is at least working better for some people now.
Hey I need information in the rate limit of Image generation on the google flow for both pro and ultra account . Can someone please helpo me to find that
There's not an official count from Google but it isn't unlimited for sure, also seems to vary according to account and locations daily demand and currently available resources. My guess for Pro account is around 200 Nano Banana pro images and like 500 for Nano Banana 2. Ultra gets at least double that or way more. You'll just know that you've hit your daily limit as you still hit the "You're generating images too fast" error after like 5 mins trying.
Also if you run out of Nano Banana Pro images you'll also hit the "you're going too fast" message, swap to Nano Banana 2 and might still have some.
ok thnaks
Well that sucks 😕
Also another problem here is that they're defaulting Nano Banana 2 for all sorts of different Google platforms alike Google ads, firebase, google lens, some search modes etc. So it's understandable they must increase the censorship there.
Sooo probably the best solution here is that they add a third model for artistry or something that doesn't need to trip on itself due to being used somewhere else. And they could even ask for age confirmation to allow to use it if they want.
... which I think pretty much exists already, it's called Nano Banana 1, and it was removed from those platforms already.
Not gonna bicker on it too much anymore tho, since it can be used with Gemini canvas sites too, too bad for everyone else tho.
Nano Banana 1 can't even really do exact copyrighted stuff either, not because of censorship but because it doesn't try to compare reference images to things existing in it's database or the internet.
It just knows what most things or characters are tho
Actually I think they did finally update this, I might be off 100 or so for Pro, but yeah it does tend to change https://support.google.com/gemini/answer/16275805?hl=en
You can upgrade to a Google AI plan for expanded access to features and models in Gemini Apps. Gemini Apps upgrades are part of select Google One paid plans for personal accounts. Important: This art
Guess I'll be cancelling my account in a month then
Just asking - who said they like Veo 3.1 - lite? Anyone -- please curious to hear.
Also, 3.1 lite - lower priority getting ingredients on May 10th?
Veo 3.1 Lite (Lower Priority extension model) appears to generate extended clips, but they’re handled as separate segments instead of being stitched into a single continuous video.
Thanks for the info. Super bummed about this for the Ultra tier.
is there that much of a difference between lite and fast, i haven't noticed it really but that might be because i use it for very simple animation?
The big difference is you can use ingredients, so the lower-priority won't do that with Lite.
speed
tricky words for me on Flow meanwhile:
#smoking ( it adds 2 cigarettes at the same time
#rightandleft ( confused on characters body parts)
#2Don3D
( it has a hard time accepting the "roger rabbit effect)
still,
i'm obsessed about Flow right now 🙂
We are really excited about Veo Lite, as many of you have pointed out, it is a very strong model. We're excited to see what you create with it!
re: Ingredients, this is a great call out and something we're looking into. Will follow up on this after the weekend, but we are committed to a seamless transition to Lite.
thank you! can confirm we are looking into this
Thank you, please consider it for ultra tier. A lot of creators depend on ingredients for workflow. Many thanks.
we hear you loud and clear and we are committed to ensuring creators can continue their workflows. happy creating!
When will flow fix there censorship stuff be fix? Cause it censored out a few image of hula image.
Is anybody else getting the "Failed: Too Quickly" message when generating images?
I've been getting it for the past couple hours. I got it yesterday as well and it feels like it lasts until the next day.
It does that when you have used up all your generations for the day
Good morning Tanong ko lang PO first time lang po Ako may Bayad PO ba Ang google lab "flow" ?hind mo Kasi Ako makapag create Ng videos
Oh, interesting. I feel like the language of the message could be adjusted a bit. Thanks for the info!
Oh definitely! I was in the same boat the first time it happened to me, wondering what was going on
Yes
so now I've found after lite has been added the "low priority" function, for fast we can only have 10 low priority??
What
@crisp basalt I mean what's this mean? 5/10....
Yes it means the fast model will no longer be available to use on 10 of may
judging by what someone said earlier here, they weren’t confident either about if the VEO 3.1 ingredients lite version is going to be available soon.
Interestingly, Gemini is already utilizing the Veo 3.2.1 model as we are speaking lmao
Yes
As an ultra plan user, I'd rather not have a lite (low priority) function because the quality is not as good as fast/quality. And when you extend the video, you have to combine the extension by yourself. This is not good for an ultra user.
Ultra users rise up 😎
I can't generate images nor videos, only this message showingup and I don't even understand why. After some hours I can generate up to 2-3 images or videos then this message starts to keep showing. I am a pro user, I have 2 more pro accounts and still can't use them... this same message kept showing in all accounts...
is this server problem or what?
Yes
You’re not the only one facing this issue
If the new function is provided to replace a previous better option....That's 
tbh the new censorship from Nano Banana 2 ain't even that much of a prob by rewording, but the way it lost logic and coherency really is and it was already bad before as it was. I hope the guy who was allowed to tweak that as it is now ends up working as mall santa in california.
temporally fix that work for me - open incognito, login
it will allow you generate few video\images and after some generation you'll need to repeat it again and again and again
Yes.
I mean, told ya.
The genius who thought he did something by "Fixing" the censorship from Nano Banana 2 really did nothing and instead pretty much destroyed the model for most other people who don't know how to do that yet
humans can go to the moon now but we are still dealing with that dumb error everyday!?
wild
inb4 we never went to the moon
please tell us any temp fix for this freking issue
unusual activity
Well it hasn't happened to me yet, so all I got for that is let it rest for some hours and next time don't generate more than 12 images at once, wait till even the last of them is generated or failed, then try again the same way and so on.
Some might just be clicking a bit too fast, don't refresh too often either.
Because that's how I've been going and went out of images today without that issue.
Flow can't generate more than 12 images at once anyway, so it's the same.
Also instead of doing all the gacha clicking you might just want to go back and restart your prompt from 0, reword the first lines or the ones you think might be problematic, once they work proceed with the next ones and so on until you have a functional prompt again.
They really can't do anything against that, short reworded prompts can't totally be censored because that'd surely finally be the end of that model and project because no one would get anything even with safe prompts
It really works that strangely, because it has also been stated to have been a problem with the original Nano Banana, there was nothing to do and was left as it was, and then they started all over with Imagen4 and NB2.
Hello. I tried use Flow, because I had yesterday problem with generation image. It was working 4-5 generations aand after I again get this information and I again can´t generate images. 🙁
clear your cookies & force refresh
or run it in private browsing
When I use Private browsing, 100% I don´t get this error?
works for me, yeah
until i send feedback about the recent censorship for the 23459th time, which reenables cookies. but if you do that close the browser and reopen again (in private browsing)
YOu lied me. 😄 I tried it and when I generated image secnd time, I get this error too. In Private window. 🙁
strange. works fine for me. without private browsing i get that error after like every 2nd or 3rd generation
I get this error, after second generation. 🙁 In private window in browser.
Can it be by which browser I used?
i use firefox
I used Brave/Edge
please fix this bug fk!
i cannot use this flow app properly.
wasting my time for failures.
you guys must vibe-coded this service. cannot believe a web service ran by top-tier company is like this.
Nano Banana 2 doesn't work well at all anymore anyway, unless you wanna generate cute cootie filled stuff with small sloppy details here and there that you supposedly have to ignore.
Whatever that could be useful for
It's the end really.
Hello can anyone help me how to use excalidraw mcp server on custom mcp client
Honest question. I know that at the end of the month, Wisk will be getting turned off. Can Flow work in the same general way that Wisk can, such as combining specific features of images, such as a character, an art style, and an environment? Specifically for still images, not videos.
I'm sorry if this is a dumb question, it's just most of the questions and answers I see about flow are focused on video generation. Not many are about image generation.
question:
i have no ultra subscription right now and in the latest news it says, "3.1 fast low priority" will be deleted soon. is 3.1 light a downgrade in terms of quality? i wasnt able to try it yet but actually planning to get a new subscription at some point
It’s a definite downgrade in quality.
sounds like ultra subscriptions are dead then? what should it be good for if 3.1 fast - low prio is gone and the alternative is a massive downgrade
Yup
The "Veo 3.1 - Lite [Lower Priority]" model cannot use the video generation feature from "Ingredients" 🙁
I get that "Unusual activity error" too
Not on every gen tho so still usable, just got to try and generate a few more than if it worked correctly from the start 🤷♂️
Hopefully this means a upgraded model for Ultra Users “lite” isn’t usable for Ingredients which is the reason I’m still using Flow
Can you guys tell me what is wrong with my AI video?
Why Every Rider Needs This | Cinematic Motorcycle Healing. #motorcycle #inspiration
https://youtu.be/M6lQf3ijNTg
#Motorcycle #Riding # healingSometimes, the best way to find yourself is to get lost on the road riding motorcycle . In this 8K cinematic sequence, we capture a quiet moment of healing on motorcycle. As the ocean breeze picks up, a motorcyclist pauses to take in the vastness of the coast. Watch as the world slows down, reflecting a journey that ...
Veo 3.1 Lite is a high-volume, cost-effective model (~$0.05/sec) designed for prototyping and rapid iteration, offering faster generation but at lower resolution (720p) and lacking native audio. Veo 3.1 Fast prioritizes speed while maintaining higher quality, better motion coherence, and audio for production. Yes, Lite generally means lower quality renders
what the hell is google doing now! NO ONE WANTS A LOWER QUALITY ANYTHING UNLESS ITS GONNA BE TOTALLY FREE AND NOT COST ME $300 amonth. Why do they keep temping people to jump ship. 😡😡😡😡😡
Prototyping is a stage where you want to test many things as quickly as possible so Lite works just great.
so was veo 3.1 fast + was better renders. This only works great for people who pay per render not ultra users. it is a downgrade with a kick in the nutts. this is a step backwards.
Not everyone is using ultra tier, like myself on pro and the many free users which now get extra generations; all of which help increase their skill and knowledgebase on how to improve as creators
I also don't see the point. I understand that some people want to quickly see results but I prefer quality, even when I have to wait a bit longer.
The few coins you have on an ultra subscription don't get you anywhere, when each generation costs you coins. And prototyping doesn't prevent you from getting different results / random outputs on each generation. Also there's differences between the outputs between the models. Even if the prototype was decent, it doesn't mean that a single 3.1 fast generation will deliver instantly a perfect result.
So basically 300$ to create 3 short videos a month? 🤣
please fix this issue!
Yay for you! but sucks for people paying full price. At that rate they should leave Ultra users accounts alone and add this lower quality bs for people that cant afford Ultra. Everyone is not in the beginning play around or prototype stages and are paying full price for a serious production tool. not a playground to waste money in. Like i said a kick in the balls for ultra users.
i totally agree with you. im here for quality and not spamming low quality videos. and for 300$ ultra gets very unattractive now. you dont get far with that few coins
how to i remove the watermark on the videos?
all the time happening We noticed some unusual activity. Please visit the Help Center for more information.
even use just 2 images on ultra tier
frustrating really all time i work really slow not share my account with anyone
Is it harder to work when attempting to get an image ready to generate for a video ,well the more characters the harder it is right. Does anyone with more experience have a maximum amount of characters to work with. Any advice is helpful because I have played with it a lot
its intended on free tier
Why does Google keep shooting themselves on the foot daily with everything... something's really fishy here. Should probably start taking screenshots of all this and report to the actual customer support. They might wanna know some mongoose is sabotaging them.
None of this has any logic
ingrediants was the go to fix for multiple characters. but since that's going away to some extent wont be so easy anymore. Klings version of unlimited ingredeints on the other hand lets you put in a ton of refernce images for each character and then render a perfect group shot. Hopefully google will catch up but as it looks they are going the opposite direction now.
After hundreds of generations, I found out I could not reliably get more than 2 characters reliably if I had anything else complicated in the scene (like an accent for somebody talking, or one of the characters is a dog, etc)... less characters, more reliable, less mistakes.
same
It's much easier for VEO to understand how 2 characters should interact without areference images and just allowing it to generate the characters by itself, sadly that's not always useful, like this one's done without reference image and seems ok recognizing who should talk and when
Seems to have heavy problems to recognize who is in a reference image, even if using sides, hair colors, clothes, genders etc as pointers to help it recognize the characters
This is the kind of stuff they should be updating instead of tampering with the qualities
Nano Banana 2 has no problem to recognize who is right or left who is a man or a woman etc... or it used to work that well last week before some mongrel started "updating" it
Veo 3.1 lite is horrible 💀
Yup, I am not going to be renewing my Ultra sub. Seems like Seedance and/or Kling is now way better than VEO anyway
What the Hell, google takes one step forward a two steps back. I´m always struggling to justify my Ultra suscription, and right now I can´t.
Is there a subscription based service similar to google ultra?
Unfortunately Kling's main site doesn't have unlimited, but I believe there are unlimited options on Runway and Higgsfield for Seedream at the moment
Flow question:
I have NB set to 2 generations/prompt
if I select "add to prompt" on a pictue and tthe type "use different colors in each card"
does it actually understand that it's supposed to generate red in the left picture and blue in right or is this just random variation I'm seeing?
It can difference left from right for sure, specially if you give it enough cues of what should be to what side. But it takes it as viewer's left and right.
so basically, even tho the prompt is "use blue in left and red in right" - the prompt is describing ONE input - just taking 2 random seeds, correct?
It's all in the wording really, be precise about what should be blue and red like if it's for background
[Background: Half red to the left, half blue to the right]
Or if it's a character wearing something do state it and give the model more clues about who it is, like if it has black hair etc
I'll rephrase: in the prompt settings you can select how many pictures will be generated (for example 2)
If i tell the model, generate blue background in the left picture and red background in the right picture -will it understand it?
or
is the model generating 2 variations of the same prompt and cannot reason/understand that it's supposed to be 2 completely different generations?
Ah you mean you uploaded more reference images that's not the same really, not it's not left and right in that case, you also need to give more clues about each picture and works better if you say the order in which they were uploaded
Lemme do something visual with it brb
last one: do the failed or rejected generations count toward limit?
Yeah it does count even if you payed, pro mode attempts dissolve so quickly with that. Ok now behold!
See it understood me perfectly by telling it the order in which the images were uploaded, not to the project but to the current image edition down there.
So if you have like 5 images and you want to transfer or explain something from the 5th uploaded image, it's the same
last one: do the failed or rejected generations count toward limit?
Yes it does absolutely count. They're heartless.
Even tho somewhere they did say those don't count. But it's not like that.
I was using pro here, but NB2 does it the same way
I could have added more clues like "The girl with black hair and uniform from the first uploaded reference image"
In some cases you might indeed need to give more clues
The second image was probably the best one, it'll just try to understand you
And you can still give more orders anyway, same eyes, same hairdo, same background etc
thanks
I wish they added - "ask NB" mode like coding agents have
... man this does indeed only work for cute things now.
Like if wanted to give it a sword or a gun I could also upload an image and explain where what it is and what to do with it
It's that simple really
will it actually not spazz out and censor a gun? or does it need a Tolstoy nover explaining it's just a picture?
Nah I think it's ok with weapons, done it a lot
But the more sensible the prompt gets less chances it works now, might wanna go short with em'
really, I assumed that is anono without trying
I will not be giving this egg a gun to try
at this pace of guardrails, safety-ism and censorship people will just go for the Chineese models
Oh yeah that's indeed another thing, someone REALLY derpy has been tampering with Nano Banana 2's functionality this week and goes beyond the censorship really.
It's like a 12 year old with crayons or something was left in charge now... which would explain why mostly only cute things work
Cant' really promise all will work as intended beyond that
Hello! new to Flow. Just wondering why i cant find the Jumps option in my scenebuilder. Please help, thanks!
Its no longer available
have they said anything about the recent censorship changes? tbh the main reason i joine dd this place is to moan about it and see if anyone else has said anything, or if anyone from google has said anything
the changes make iut basically unusable for me
almost everything i want to is blocked now. have to try like 50 generations to get like 4 outputs. and then of that 4 i have to hope, firstly, they're actually what i asked for and secondly if it's a good interpretation of that. and even if it is the about of mucking around kind of takes the fun out of getting those images to begin with. previously i could just do one output of x4 and get 4 images
it's only been this way since last week. did they break something? was it intentional or should i expect a "fix" eventually?
im saying this without any animosity, but could there be a consideration to rename Veo 3.1 - Quality into something else that more accurately describes it? After using it for 6 months I am certain that it is anything but quality. it produces more slop and ignores instructions way more than fast ever does. The naming of it is entirely misleading. Unless there is some trick specific to it, in which case this must be stated, but to just call it quality gives the consumer the impression that they're getting results that arent gonna drain 400 credits per genenration that all go to trash.
An example of this is happening right now. 16 videos done with Quality Frames to video. All of which I clearly instructed for the clip to be a single take with no cuts. NONE (0/16) did this. Yet when I ran it with fast, it immediately followed that instruction. That's a lot of credits wasted.
I'm personally going to stop using it for good, but I still think its ethical to accurately tell consumers what "Quality" means.
It's also really bad at lip syncing. Just overall an expensive trap
I am really loving Veo 3.1 lite
the others are fine. they work okay. i just think quality needs to either be removed, or for them to accurately state what is quality about it.
as an Ultra user I am really not loving Veo 3.1 lite
May I ask why not
grok is pulling ahead anyways. and it doesnt run on a credit system. something thats making ultra a hard sell for me
Actual Google staff hasn't said a thing about it, they only talk around here when there's something minimally related to the video generation, seems to be mostly what they are in control about around here.
Or care about
Sometimes the mods do talk to say "this is fine" "nothing to see here" "all's better"... tho it visibly isn't
censorship? you're barking up the wrong tree. if you're aiming to create content with anything sexual, even implicitly, you're gonna give urself a headache using veo
Best thing you can do is restart your prompts from 0, reword everything that could be barely problematic and shorten your prompts, that's how I got around it. Can't do detailed pretty stuff anymore.
Veo 3.1 lite is a reduction, lower quality, has no ingredients (something you really need for some shots) offering faster generation but at lower resolution (720p) 😡
That or make yourself a canvas site with Gemini that can still use original Nano Banana, which is what Whisk uses for precise mode, that's as permissive as it's ever been, and way prettier than anything Nano Banana 2 can make, at least now,
Last week it sure was working fine enough before they gave someone with woke-itis the wheel.
its getting better at understanding context and innuendo from what i heard. i dont waste my time trying. i can imagine trying to get it to do something like that can be really frustrating. so i think you need to set your expectations rather than hope for it to change. they will never make it lax. they have advertisers they need to be friendly with. generating deepfakes and all that is way more risk then they will ever allow
brb fightcade match.
doesn't surprise me. i doubt they even know or care. i tried asking ianfromlabs about it on reddit but still awaiting a response. probabkly just how it is from now on. shame, was fun whilst it lasted. useless now