#flow-general・🌊
1 messages · Page 52 of 1
@signal hare
"Anyone getting 'failed' error on video generation?
Been failing for 2 days. Korea region."
I just upscaled 200 images from my free Flow account and then the upscaler stopped working, I guess that's around the daily limit for upscaling for free.
Not really using Flow's Nano Banana 2 anymore since whoever's in charge destroyed it last week, but I still need to download my old images from when it was somewhat starting to work decently.
Or it was more like 199 images for some reason.
I imagine it's good to know the upscaling daily limit, according to tier plan you'd probably get double that for each tier upgrade or more for Ultra. The guessing game is real.
Don't wanna pay even 3 bucks offer as it is now, just gonna wait till tomorrow to download the rest or make an upscaler with Gemini's canvas, but not that much in a hurry really.
Hi Mods! I'm on the Free Tier of Flow (labs.google/fx). Can I use the videos I generate in a monetized YouTube channel if I edit and narrate them? Or is it strictly personal use only? Thanks!
Have been running into "prominent people" warning more often (there are no prominent people)
me too. I'm getting frustrated. It's hard to get any work done
@little fulcrum @grand isle sorry to hear this, would you mind starting a thread in #1238621511401734254 so we can look into this further? the more details you can share on the prompt, model used, etc will be helpful for us to debug
not what we're going for, and agree that doesn't look great. would you mind filing a bug at #1238621511401734254 while I follow up with our intern?
@little fulcrum i saw that you posted in an existing thread, no need to start another one!
anyone still getting unusual error ?
mine still happening. Even in ultra account, so much frustrating... im canceling my sub this is too much.
will cancel this month as well
yeah they messed it up
i cannot generate different angles, or the same character in a different position
Hey. Is there a Chinese-language channel available? I’d like to inquire about the rules and restrictions regarding FLOW. 1: I used a custom-designed image of a minor, but the system flagged it as a violation of child protection laws.
2: When I used FLOW to generate an image of a minor and then added further prompts, it again triggered a child protection restriction warning.
3: When I used an adult character to interact with props—such as kicking a table or a chair—the system warned me that the generated content was harmful.
I didnt remove the VEO watermark
its still way better than kling 3 omni
How should I tackle these creative ideas?
same problems
I believe the threshold for deeming character-prop interactions harmful is extremely high.
This restricts a lot of the action and plot.
For instance, if a character accidentally pushes over a minor, this involves minors and potentially harmful content.
这里有华裔吗?
Hello, does anyone have tips in order to rotate, to move just a little bit, a character ? I just want the character in the picture I send you to look a litle bit more toward the camera, but NOT facing the camera. I never succed to do these litlle changes. Even when I give examples, that does not work. There is a lack o f precision. Thank you in advance if you have some tips.
Possibly add a flag or report function when this happens, so your end can collect and work out the false negative filtering... 🤔
I have a recommendation for the Google Flow development team: give Gemini chat good knowledge of how your model works,
so that it gives you accurate answers to your questions and effectively solves the tasks.
It seems to me illogical and strange that sometimes the Gemini chat does not know how a certain functionality works or cannot give the correct prompt based on guesses but not on their product knowledge. This is very easy to
which audio model u r using
its the native VEO and Kling audio right now
I will change it using 11labs voice changer after I am done with the visuals.
okayy
do you guys have any youtube channel recomandations... started with flow few days back
I been running into that bug all week, it's super annoying
why do you need youtube bruh
just start making content
@pulsar coral you mind me asking what you use AI generated content for? your social media or something?
im creating videos..
& its best to llok for other content
sure bro whatever you do you
A question about Google Flow using NanoBanana. The images you generate in the same project, using the same And on the other hand, images that initially don't detect copyright or intellectual property rights, but then do later.
yeah its weird guardrails and im gonna start crying because of this now
Out on the street kids appear in all kinds of signs including those for schools/education-related ads, also products they would actually use or even exactly to advertise protection laws, societal norms and conscience towards them.
It makes sense to block their presence alongside certain strong trigger words or strong explicit terms, but saying they should never appear in any image at all is unreasonable.
That kind of view feels like it comes from someone who has never seen how the real world actually works, like he lives in basement or alone in island or something, but more than all lacks the capability to tweak the image models correctly so just blocked everything he could and called it a day not caring much.
Yeah even after all that's been going on there's something about VEO and close ups to real moving people that somehow come up real good for some reason... they'll find a way to undo it too don't worry. idk if this is using a reference image.
yeahh.. so itwould be nice if anyone share some youtube wokflow
i saw videos of henry .. & my mind went crazyyyyyyy
I figured. Its really crazy, I've never seen it this bad. I'm using the same character references I've always used too.
yeah i used reference image
dude I loved NBP but they really messed it up, unusable at this point
i want this person in a different position. I have tried everything and cant get it to work, request gets flagged
i never watched videos on AI content generation
I did show it the other day around here, I used Nano Banana 2 to generate art of anime characters standing at the beach near the ocean, and similar stuff.... it was working fine two weeks ago now fish float out of the water near the surface like flying fish.
Funny enough the censorship did nothing to it.
original Nano Banana could always do it better anyway
oh cool
Or as if the fish were plastic bath toys, prompting better helps sometimes but NB2 does whatever it wants now
did google staff say anything about this?
Yeah but I posted at the Gemini help area not here, I don't think the Discord is being taken seriously. tbh no one takes them seriously... really Discord started as a way to connect for gaming, not as a real form of communication for everyone.
I do have like 5 threads at this Discord they've done nothing about.
space big milker killer by charles curren is one of my top 3 AI content project
dripwarts is also top 5
if anyone has some cool tasteful high effort stuff they saw or made, please do share it with me
damn that's annoying
but who cares im waiting for gpt image v2
or whatever is the next best thing
tried seedance 2 on 3 websites and its heavily nerfed unusable trash
i want to see some of your work
Also sometimes the main character in the image comes out so perfect but then the weird ocean ruins it, it's a waste and everyone knows constantly fixing things might cause the image to lose quality or the context of the image, so it's probably better to start again.
Prompting to edit the image not to lose quality after each edit is better done only if you REALLY need that particular image, and sort of works.
Has anyone found a fix for this issue? I’m assuming it’s not permanent. My team and I bought a second Ultra account because we thought the problem might be tied to one profile, but we’re seeing the same issue on both accounts.
does Google AI Pro give you 50 daily credits on top of the 1000 monthly?
i am confused do you get less credits on the Google AI Pro plan than the free?
Might wanna try Wan 2.7 too.There's also some other new video model thing since like a month ago called "LTX-2-3" both are very uncensored too. I've seen LTX making realism well enough similar to VEO with even more people interacting and arguing lyp-syncing. Then again seeing and completely testing are two different things.
1,600 credits for free and 1,000 for the Pro ?
Yeah the math doesn't add up there, you get more credits if you come back daily for the 50 credits.
But if you pay for a plan you don't get any more daily credits.
thanks, do you know what the logic is for this?
so Pro is subsidizing free?
Yeah when the second moon of saturn is aligned with earth and planet venus a vortex opens right into... actually no they just messed up for sure.
To fix the unsual activity error and similar ones try:
-Switching image models, from Nano Banana 2 to Pro and viceversa, you might still get some more images today.
-Clearing cache/cookies and refreshing.
-Change Browser
If none of those work your account is dry for the day, you gotta come back again tomorrow around the last time you successfully generated an image, a little later even, it isn't exact 24 hours sometimes. All plans have a daily limit.
Also if you're sharing your account with someone or using a VPN you could also be getting that error since there could be two people at Flow in the same network.
Also once it works again don't generate more than 10-12 images at once wait till all of them generate then wait like 60 seconds to try again, if you go faster than that you'll also get that error and will have to start all over again.
It's much worse for free plans, can be as low as 6 images at once according to daily demand and region, but some other days it works just fine.
This goes for the amount of video generation attempts too, you could still have credits left but unable to spend them due to a daily limit according to your plan, it'll be fine too around the next day
Also if none of that works for some reason you might wanna open a thread at the labs-bugs-feedback area to inform staff. Gotta say it because there's already threads about it.
Saturday night Flow seems to work real nice with less errors.
On other notes: According to tier you get a set amount of image/video generation attempts, and the failed attempts do count.
With free tier accounts you for example get around 20-30 free attempts to generate Nano Banana Pro images and 100-200 Nano Banana 2 images, it doesn't matter if you succeeded to generate an image or not. And around double that amount for each tier upgrade. This may change according to daily resource availability.
The amount of missed image/video generating attempts has currently increased a lot because of all the new censorship and copyright security. So to avoid being flagged by google security might wanna try to rephrase your prompts making them shorter so you can revise them more easily for problematic terms instead of playing gacha trying to get an image.
As well remember if your Flow account is dry for the day already, your same Google Flow account free or payed plan links to Gemini, you can get some more images and videos there or play with the canvas mode to try to generate yourself a site to also make more videos and images.
Gemini's Nano Banana 2 can be told to generate like 4 separate images at once or similar to Flow, but yeah needs to be told to do it by prompt. Same way it can generate the image with requested aspect ratio by prompt, end your prompt in something like "Aspect ratio is 4:3" and it will obey you.
And yes the images have a Gemini spark watermark there, so be sure you use the download button on the image not right click save as, that will give the image 2K resolution automatically, and it will also allow you to use this watermark remover on it, it can do it in batch and the images won't lose quality: https://geminiwatermark.org/batch-watermark-remover
Then there's this new "Prominent people" error, that's the type of error people keep pushing regenerate to try to get lucky about and avoid to get an image anyway.
I've seen explaining to the Nano 2/Pro model your image has nothing to do with anyone else increases the success rate tho it can still be stubborn about it and fail a lot. Gotta start the prompt with something alike.
[Identity or character: This is my personal character not any existing character so never use any external sources to create it and do not mix/confuse it's appearance it with any existing person.]
But more often to try to explain it to it in some way does influence things.
it makes sense, these modern video platforms are always running an LLM in the stack that's trying to parse user intent instead of just running strict keyword filters
Yeah but even if you stop getting that error you might still get different ones. It just keeps finding ways atm. I miss how it worked two weeks ago and even back then it was problematic, it was censored/copyrighted enough already too.
yeah honestly I felt like they already had the perfect balance back in like.. november. google's prohibited genai policy was last updated in december 2024, but recently each platform (flow, flow music, etc) seems to be beginning to enforce their own policies with really messy classifiers, flow music gets stuck on simplistic heartbreak songs and throws errors about "not meeting our content standards" until the assistant rewrites 2/3 of your lyrics lol
Wish I have an undo feature so that way I can restore the deleted image.
Yeah that's one I asked for too.
Because I accidently deleted a character that is a strawberry cake.
lol yeah they dont even ask if you're sure just yoink
strawberry cake is a character?
cool
This is what I'm working on for ZVL characters. The one with glasses.
weri guardrials? I don´t unederstan
I still want to know what veo 3.1 pro is lol they added it to gemini app limits docs when they released 3.1 lite but they haven't announced it yet
Bring back the seed.
Yes I’m on ultra
that's just generation with the pro gemini model, the two lower tiers will get veo 3.1 lite with pro selected and ultra presumably gets veo 3.1 quality, but it used to list 2/3 "veo 3.1 fast" for the two lower tiers and 5 "veo 3.1 (preview)" for ultra. they updated it to say veo 3.1 pro when they launched 3.1 lite like 2 weeks ago
Veo 3.1 pro is 3.2
who knows, generations with gemini don't seem like they're running some radically better model on ai ultra for me anyway. just find it interesting that it lists a distinct pro model variant on that page and nowhere else in anything google publishes
lol ok
I lied. Fast still uses pro model
Then why aren't your pants on fire
Maybe it's still the same model but "Pro" just gives more juice to it. Cause by checking how NB2 works at Gemini just by pushing "Pro" feature into everything just steps on the pedal. Maybe does the same to VEO.
NB2 has been stated to be just NB Pro nerfed and cheaper, cause we did have NB PRO alone by separate like 3 months ago.
Hello
That is very excessive all that happen back to back?
I'm now out of generations for the time being because of this error. I feel like I wasted generations just to get this error over and over and over, and now I'm locked for who knows how long on cooldown because they will NOT fix this error.
I'd appreciate if the google team could address this, too.
I have tried ltx
It isn't that good
I am paying a dollar per clip damn
idk gonna have to sit down and really try to prompt in it. But when it works seems it really does.
I got 940 credits left and going to expire tonight. I don't use VEO much. I am still new.
Even the only guy who was making the cute safe chibi character scenarios is now complaining that it can't be done anymore with Nano Banana 2/Pro. That should tell you how unusable the model is atm. The chinese better catch up faster... and they will.
NB2 really needs a rollback to two-three weeks ago, gotta let the money talk. I haven't seen a single person praising what was done with it all over social media, let alone a paying customer. It'll be so hard to bring them back later.
hi anyone know can we upload video as a reference in flow ?
anyone an idea how to create videos like these? Where realtor walks/speaks while room behind him changes. Feel free to DM me. https://www.instagram.com/p/DTOSdTFkb-W/
@sick sparrow its green screen ig
anyone having issues with veo?
Hello everyone, do you feel like AI filmmaking today is more about visuals than storytelling, or is it just still early?
what issues u havin
It's less about storytelling skills and more about the models' ability to deliver.
Today might be the turning point where real productions start coming to life in a meaningful way 70–80% there in my opinion, and with agents now driving the wheel, progress from here will be fast.
Yes think so too, but at some points you even see the shadows in the rooms of where he is walking
depends of autor
why can't gemini make transparent fotos
hey guys, can someone help me?
I'm using a template as a reference and I simply ask it to change a specific word without altering any other details. However, sometimes the words end up with some strange lines or symbols that weren't in the original template. I can't figure out how to stop this from happening. Does anyone know what the problem might be?
Original
Next version
are This artifacts?
Do tell it to
[Rule: Use the exact same font in the same type and color from refereference image without any creativity freedom]
Also the other stuff could simply be loss of quality so you could end it in a prompt that at least tries to preserve it
I often end my stuff like
[Complete image's quality: Keep everything true to reference image especially the facial structure, but all Professionally remastered in best known quality, HD ultra high resolution, sharpest definition, denoised, unblurry, clean lines, High contrast, fix any jpg artifacts, 4k detail, fine textures, clean edges. ultra detailed, sharp focus, high frequency detail, 8k detail. Vivid vibrant solid colors accurate to reference image]
If that changes something you could try removing "Professionally remastered"
Then again all that is just text against a black background maybe can just be done literally writing with fonts in whatever image editor.
Cause if nothing gets it done after that, gotta mention Nano Banana 2/Pro have lost a lot of coherency and quality since a couple weeks ago, which seems to come as a side effect from when someone very irresponsible tweaked with it's copyright and censorship levels, a lot of simple words alike "Pose" seem to trigger all sorts of filters along other words.
At this point I just started writing in my images instead too, not because it can't do it but because of the constant copyright error messages that have literally nothing copyrighted in it.
Added
"Vivid vibrant solid colors accurate to reference image."
to the prompt because looks like it could use it.
Now that I remember I did face a similar problem when trying to place a similar bar with info below my character images with the character's bio alike name and sorts. It had no illegal ages or anything alike but the bar prompt literally triggered all sorts of probs for no reason while the image itself didn't, wasted my work night on that.
Had to just get an image editor and write, much simpler.
Incredible this now needs to be done when the NB2 model was working perfectly fine for it two weeks ago and it's main feature and supposed advantage over Nano Banana (Gemini 2.5 flash image) is that it's able to write in signs and logos, it can't even do that properly now.
I tried changing "bar" for "stripe" or "12 pixels line" and all sorts of things just seemed to make it worse.
Good thing I cancelled already.
I accidentally deleted a video I made using flow, how can I get it back?
it's not possible
Even if I contact support?
it doesn't matter, once deleted, it's gone. atleast on your end. i'm sure some data remains on their end, but you can't recover it
and contacting support is pointless
save your work, or be more careful
Me too since yesterday while making a cute chibi-style girl that is cake.
Hello, World!
hi
I agree. today hasn't changed
Like said before I did ask for a recycle bin feature, I tested creating one in Gemini's canvas at a website and it can do it, so I'm pretty sure it can be done at Flow eventually too.
Idk why'd it take so long tho, it's basically a matter of asking AI to do it and it would, like 30 mins of testing at worse.
A lot of people do delete things by accident because Flow takes too long to generate the image thumbnail and that makes it seem like the project is empty and wont matter. I'm at a 128 DDR ram comp with decent connection and it still takes very long to load, so I have also deleted a thing or two by accident.
Also I got this mail from Flow earlier
Haven't seen such "Help improve Flow" option yet except maybe for this one, and considering data seems to have been used against us instead of being used to improve things, would really like to opt out, well it does say it's gonna be available later.
Okay so how do I change the thumbnail for a project? Each one has the initial image I uploaded for each project, and several of then use the same starting image.
Hi, I looking to generate a 4k image with Google AI Pro, but the button is still disabled with the "upgrade" label pointing to "Google AI Ultra". Isn't Pro sufficient enough ? Thanks.
facing this issue
im getting that too
@solar idol can you help?
yep, same thing with nano banana pro
nano banana 2 works fine, but obviously it is a worse model
Hello! Please submit any bugs or feedback in the #1238621511401734254 channel. The team monitors those threads. You can also add your issue to existing threads by searching for the same issue. Bugs and feedback tend to get lost in this chat so please submit them above :)
Thank you!
Is anyone else still stuck with this? Because this is honestly super frustrating. I’m on Ultra, and Flow keeps failing with the same “unusual activity” message. Sometimes refreshing gives me one successful generation, then it breaks again right after. This really shouldn’t be happening on a paid plan.
Is anyone else still stuck with this? Because this is honestly super frustrating. I’m on Ultra, and Flow keeps failing with the same “unusual activity” message. Sometimes refreshing gives me one successful generation, then it breaks again right after. This really shouldn’t be happening on a paid plan.
Not even joking anymore — GPT Image-2 needs to come out fast. At this point, Nano Banana Pro is the best marketing campaign for GPT Image-2.
is flow down rn ?
New here. Been a subscriber on Flow for about a month.
Getting the following error: Something went wrong (super useful error).
Happens when opening one specific project. Happens on any computer I use. I've clearec cache. Seems like the project got corrupted on Googles end? Any way to get the images in it? Or do I just have to wait for the app support message I sent to be looked at?
same, I'm on Ultra on two accounts and still I'm facing the same issue on both accounts. Now I'm just thinking of canceling it since it's not helpful anymore. I have been suffering from this for two freking weeks!!
Flow's NB2 did seriously give me this image by itself. Doesn't that beat all.
For the "Unusual activity" error clear cookies/change browser and when it works again don't generate more than 12 images at once, aka you can ask flow to generate 4 images at once and press the generate button 3 times, that gives 12, wait for like 2 minutes or till of them are generated or fail to generate then you can try to generate more again. If you go faster than that you risk getting the error again and you'll have to start all over.
Plus plans and above can generate like 20 images at once aka 5 generation button presses of 4 images at once, but wouldn't advise to go faster than that.
Also if you're using a VPN or someone else is using your same account you could also be getting the error due to all the samey activity, so yeah check for that. Someone could be on the same VPN network using Flow too.
is there any update about the servers right now? im getting this error
flow still not workinggg
works for me
i am using different non celebrity character but still flow giving me prominent people error, this type of error should be fixed
what plans are you?
free plan. but this is irrelevant
i got somethign went wrong loading your media to all my past generation and new upload
relevant i think
my friend using pro is able to generate and use flow
believe me, it's not. all plans are subject to the same platform bro
i use ultra cant do anything now
saw this too
it has nothing to do with plans, it's something else
can be regional too
or some account related bugs
im in the same office with my friend, using different account with different plans, but he can use it..
maybe im unlucky, but cant finish my work and its late huuu
are you from philippines, by any chance? just wondering
no, singapore
It will work if you clear cookies and refresh tab, you can get 2-3 successful generations (NB2, Pro) until the problem shows again.
I will still prefer the old Whisk for work. It was super helpful, Idk why google is doing this. Flow is good for filmmaking but we all are not film makers.
the problem is not error generation, im literally losing all my history, all images, and cant upload any image reference..
hi
Please guide on how to use flow to generate scenes just as we do on whisk?
@sick trench thank you for all your answers! gotta experiment
Megabag sounds about right
Whoever's getting the "Failed something went wrong loading your media" error, or the problem about Flow being offline will just have to report it at labs-bugs-feedback, I don't think anyone around here can solve those.
Remember to read the guidelines on how to post there and tag your thread well.
Flow is working for me atm, so if it's not working for you might wanna add your location there
Some Flow accounts can just be drastically different from one another according to location where they were made.
Or the time they were made. Even if it was not too far off.
Don't know if it's the same thing, but in Lisbon Portugal I'm getting this.
Is that for Flow or Whisk? Either way you gotta report those at labs-bugs-feedback.
It's flowmusic and sorry, will post there.
Or not, don't have access to that channel I think, but got it.
Weird, it should be around there somewhere.
Seems a lot of people are having issues to access all different lab features according to location, specially asia and europe... I've haven't had any trouble to access from South America even once in the past months.
It could also be related to connections, Flow visibly eats up way more ram and bandwidth than Whisk now for some reason.
And I don't think it started like that.
Yeah in fact I just tried hard to open a lot of Flow windows and I did get the "Error failed to load your media" errors, so that one that seems to happen when the connection can't handle all at once or is low.
My connection is good enough and has never done that to me before, that should say how much bandwidth a Flow window is eating now.
Just noticed I was also downloading steam games Faceplam, turned it off and now Flow loads the images just fine, the "failed to load your media" error is indeed just a connection issue.
Yeah fine on my end, I'm in south america. Might wanna restart your connection somehow. It might load after that.
But if not do go report the issue at labs-bugs-feedback
yeah im in south america too, never had this before
Hi everyone, for the past few days I've been testing the latest update to the Gemini TTS audio model, version 3.1, on the aistudio.google.com website. I wanted to add emotional tags to make my content more efficient, but I've been testing narrations between 10 and 15 minutes long. I've submitted them to the model on the Google Studio website, and it takes a while to process, ultimately only delivering a single line of 15 seconds out of the entire extended narrative. I've tried it several times, also without the tags, or with an average of 2300 words. It processes fine, but in the end, it only delivers a few lines. So I'd like to know what the reason might be, perhaps what adjustments are needed to make things more efficient and avoid these errors, because they waste time and effort. Thanks!
Guys, it's already the 21th and my projects still haven't been transferred to Flow. Is there any estimate of when this will happen? Do i need to do anything to change the status of pending to done? It will be transfered until the whisk goes down? any mod can awnser me plz!!!!
oh dude im cancelling my subscription this is so trash haha
both image and video generation arent working
"prominent figure" literally a person i generated in that same chat
everything is getting denied
idk what they did but they really messed it up in the past week or so
unusable garbage
It's not even good for filmmaking anymore after recent censorships
The Veo 3.1 model is lagging far behind other models. Please release a new model as soon as possible.
What is others using ?
If it was working fine some hours ago maybe your Flow account 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.
Sometimes clearing cookies fixes it too but more often if it keeps happening after like 5 mins your account is just dry for the day.
On other notes: According to tier you get a set amount of image/video generation attempts, and the failed attempts do count.
With free tier accounts you for example get around 20-30 free attempts to generate Nano Banana Pro images and 100-200 Nano Banana 2 images, it doesn't matter if you succeeded to generate an image or not. And around double that amount for each tier upgrade. This may change according to daily resource availability.
The amount of missed image/video generating attempts has currently increased a lot because of all the new censorship and copyright security. So to avoid being flagged by google security might wanna try to rephrase your prompts making them shorter so you can revise them more easily for problematic terms instead of playing gacha trying to get an image.
As well remember if your Flow account is dry for the day already, your same Google Flow account free or payed plan links to Gemini, you can get some more images and videos there or play with the canvas mode to try to generate yourself a site to also make more videos and images.
Gemini's Nano Banana 2 can be told to generate like 4 separate images at once or similar to Flow, but yeah needs to be told to do it by prompt. Same way it can generate the image with requested aspect ratio by prompt, end your prompt in something like "Aspect ratio is 4:3" and it will obey you.
And yes the images there have a Gemini spark watermark there, so be sure you use the download button on the image not right click save as, that will give the image 2K resolution automatically, and it will also allow you to use this watermark remover on it, it can do it in batch and the images won't lose quality: https://geminiwatermark.org/batch-watermark-remover
This is not related to the "Unusual activity" error message tho, I did already post about that one up there. But yeah clearing cookies and trying not to generate more 12 images at once deals with it, Pro plans and above can generate like 20 at once without getting the error.
And yeah I get the funny irony.
Sometimes it don’t even be 4K generated from Gemini
Gemini's images default at 2K only if you use the image's download button, also it's possible to create a 4K upscaler with Gemini's canvas, just gotta ask it to make you one as if you were generating an image.
Is the delete option gone in flow? I see Archive is new?
This is nice but it's just gonna be more reasons for copyright and censorship, we really need the commercial Nano Banana models separate from the artistry models. And yeah does seems there's some updates atm better lurk
I just got this too, the end is near. Well at least things seeem to be moving now.
doing stuff in google ai studio and i know now its connected to you google plan, i have the ultra plan and it seems like my qouta runs out so quickly
Flow and Gemini's accounts link with the same Google account plan, but I don't think Google Ai studio's account links to those, that one works separately by buying an api key and only charges per image generated, not by failed attempts.
i thought a couple days ago they annouced that they now connect to each other
Excited to share that Google AI subscriptions (Pro and Ultra) now work with @GoogleAIStudio!!
Come vibe code and use the playground with higher rate limits. Available right now!
Yeah just noticed they updated how it works, gonna go check how it's working now.
yeah now it seems like i meet the qouta so quickly vs before they made this change i could do more
Gonna try my limits, but at least we can use the original Nano Banana in the same account so that's something to have when Whisk is gone.
The billing for API key is also much easier and intuitive to use now, it's like a miracle they finally did this.
Compared to Whisk and Flow's plans one can also pay for each image instead
Like they suddenly noticed "hey people want to pay for this let's make it easier for them"
maybe what im doing is just that intensive, idk, just know that when i was doing free daily qouta it felt like i could do alot more
Doesn't seem this takes plus plan, only takes Pro and above. Well it's still something to be able to use NB1.
yeah, i have ultra, thats why im confused
wheres the delete option?
gone, its archive now
Does google PRO have extra usage with google ai studio? or those are 2 different things?
why not
Also I'm pretty sure NanoBanana2/Pro's censorship got reduced by some good %10-%15 at worse.
Some old prompts are working just that much better more or less
Which is still meh for what it was 3 weeks ago.
But yeah no doubt some nuthead did mess it up badly by then, gotta try consistency of prompts again to really be sure.
Yeah not gonna call this progress yet but I do notice they're trying atm.
Yeah this isn't %10 after all, might be %5 at most
i heard it slightly increase limits in playground, but im not using it, wouldnt call that usefull, ultra would give access to early models probably
still its only the playground, the api thing is the same for everyone i think
Right now Google Ai studio's Pro plans do link to Flow's and Gemini's same account plans, but Plus plans don't, I know that much. But some users complain the image generation limits are very low at Google Ai studio.
But at least it's easy to buy an API key there now like to link a credit card and it'll only charge you for successfully generating images, not by failed attempts.
Yeah I think in no way Nano Banana 2/Pro can be declared in any way useable again yet. But surely it's getting revised somehow.
The problem about the black sign I needed below the images is indeed working somewhat better now, so that's something.
for image creation which yall recommend ? Pro or nano 2
Anyone use this yet?
So, Flow now has a 'share' button... can we get a 'remix' button also? Being able to remix images and videos that others have shared and tagged as being available for further use would be a lot of fun. As it is, it looks like we can share with the prompt, so that kinda would let someone make a remix manually anyway - so why not just bake it in?
On the other hand, because I'm a contradictory kind of person, I do have to say Sora used up a lot of gens on random remixes of dog-on-skateboard type videos, so maybe forcing someone to have to actually copy and paste the prompt if they want to see the dog with a different hat would be wise, from a resource-use perspective.
For my random $.02, I feel like Google is in a strong position to let all these smaller upstart AI companies blaze the path, before, like a slow glacier, it then crushes them all. I like the interface and layout on Flow, and the Veo output seems quite nice so far. But the price and the "describe 5s, then check the result, then describe the next 5s" interaction style seems like it has a little ways to go before it's ready for primetime.
You know what this opens up a new workflow automatically for the #perpetual-stew・🫕 channel & others of course @plucky vector.
Thank you
I was wondering, now that my Whisk creations moved to Flow, I tried to reuse my prompts from my creations, but it gave me an error about AI credits needed.
Happened to me too
flow has been tripping lately, blockign everything, talking about promiment people
Mine's haven't moved from Whisk to Flow yet, and I'm sure I was the literal first guy to request to move them.
I'd rather have them moved to Google AI studio tho since they do have Nano Banana 1, it's useless to try to run a Whisk prompt with Nano Banana 2/Pro.
Yes im still waiting the transfer from whisk to Flow..
Also is there any news about veo 4??
Me too.
Hi. RIgby745 Here.
Thank you so much for reaching out! Does this happen to you all the time when you try to reuse your prompts?
I understand you've got problems when you reusing your prompts.
Change the model, and reelect the model you've used to create your images.
Tell me if your problem is unresovled.
Ah okay, will try your advice
OK. Let me know if it works.
Yeah good point! Lets give this a try!
https://labs.google/fx/tools/flow/shared/video/f5b943c6-4bd6-40bf-8f5b-4533ddad11d2
It said "sharing allows anyone with the link to view and remix your creation", why not give us some different fighters? Open to anyone who wants to give it a spin!
It works, thanks!
Hi, is veo 3.1 fast still zero credit for Ultra Plan subscribers?
You're welcome! Thank you for your help.
Hi guys, been trying to create a video using flow and I encounter this kind of error. The video basically just people walking with soft footstep as the audio. Anyone here been into this kind of error?
I noticed ChatGPT’s image model has this feature, and I was wondering whether something similar could be integrated into Flow’s image generation
Those are the same aspect ratios Google Ai studio has for images, I've also asked for them but I guess since they wouldn't work with videos only with images it's not a priority.
Hi, I would like to now how can I delete any generated image...?
the option to delete images is not visible
You now first have to archive an image to be able to delete it, it'll have the option to delete at the archive. I imagine it changed like that because so many people kept deleting things by mistake. Archive sorta works like a recycle bin now.
I heve the same question
Same answer
Every image has the archive option.... with the way Whisk works it's safer tbh.
I see, thanks a lot
I'd rather be able to control aspect ratio by prompt like Gemini and I think Google Ai studio has too, Nano Banana 2/Pro can absolutely do it and it's a waste of technology not to add it to Flow.
NB1 doesn't seem to be able to do it tho, or I've just never seen it doing it unless you already cropped the image to the size you want and the site has "Auto" aspect ratio option... which is another thing missing at Flow.
Hi all! I'm curious about the licensing terms for videos generated with Google's AI tools. Specifically, does the AI Premium/Pro subscription grant commercial usage rights for the output, or is it restricted to personal use only?
There are some things that VEO gets right, i cant tell what exactly that is but, my best guess is expressions and movements? kling looks robotic in comparison to VEO
but kling is miles better with texture and keeping lighting consistent
what's up with this error while generating pictures? someone encoutered it also?
I have a lot of credits
i hate this
me too
why i cant generate images, its not normal to generate 1 image in 5 minutes lol
Seems flow is broken again been stuck with this error for a week +-
Same
I've been getting that since last night as well
Hello Google Flow Team, Can I use the images that are generated by Google Flow - Imagen 4 model for commercial purposes?
yes
they dont claim ownership for the content u generate and u can use them to advertise, make money thru monetization and stuff
What About Nano Banana 2
same thing
Are you from Google Labs support Team?
i came for this, Yes wassap with that
Any Mods here?
Hello, is there anyone from Google Flow support team?
For faster fix , just toggle between NBP and NB2
after a while its not working again
@warped wind Can I use the images that are generated by Google Flow - Imagen 4 model for commercial purposes?
what is wrong with this stupid flow site. its telling me i need more credits. it says I have daily credits (that never go down) god flow is meh compared to img fx.
yes just keep doing this till they fix it
Switching back and forth between NB2 and NBPRO helps you keep working for the time being
I have problems with the damn 99% things for image (im on mobile(
It would be great if the generated videos no longer had watermarks; it would help with competitiveness.
There's no watermarks on Flow unless you prompted it
When I create videos in Veo, a watermark appears, even though I have the PRO plan
Ah I see, sorry I misunderstood. Yeah agree it would be better without it
That's okay—it would also be great if they included a RAW version and native 1080p
Hi all! I'm curious about the licensing terms for videos generated with Google's AI tools. Specifically, does the AI Premium/Pro subscription grant commercial usage rights for the output, or is it restricted to personal use only?
Flow and Google AI studio aren't comfy at all, paying for them doesn't feel smart at all when Whisk was a better tool since months ago.
Google AI studio has Whisk's original Nano Banana but it's little AI assistant is insufferable, Whisk didn't have it.
More than all Whisk is a much faster area to test prompts with the original Nano Banana, this transfer locked my Whisk screen and I now notice how impossible it is to work without it.
Flow does have a similar multi image creation screen, but without the original Nano Banana model it's useless.
And no there's no way to get used to this, no one's liking Flow over Whisk, everyone's saying it and no one's gonna get over it by using NB2/Pro as buggy and full of errors as it is now.
Now that Google AI studio accounts link to Flow and Gemini it's much more obvious Flow is lacking Nano Banana (Gemini 2.5 flash image) since Google AI studio does have it. There's no reason to have it disabled at Flow then.
bleh. ty for the tip. sadface they somehow glitched it out like that tho.
We need this model back as it was when it was there like this, 3 weeks ago it was still viable to remove it while Nano Banana 2/Pro was still useable but not how it is now.
That or Whisk could've been left as it was.
Saying Whisk is migrating to Flow is deceptive without it when it doesn't offer the same type of image generation at all.
why does generating images say that cant do that because of insufficient credits but i have over 22k left?
Images aren't supposed to take credits AFAIK
Hoping this is just a bug
yea thats crazy
why am i receiving this error thing in flow?
im on pro plan
oh wait everyone else is getting it too
hey guys! please use #1238621511401734254 to submit these bugs! these threads are actively monitored and will be seen quicker by the team there!
I did it already
Has anyone noticed the generations lately seem to be pretty glitchy? Way more so than usual.
What are the main differences and use cases for switching between Nano Banana 2 and Nano Banana Pro? To me they seem pretty similar, am I missing something?
Even this
I get that when i have used up my daily nano banana pro generations. gotta wait for the 24 hours cycle to reset
That’s too long
use nano banana 2 until it resets
What about this too
Guess that is what we can expect of an early alpha product 😅
It says this prompt might violate our policies
ye it "might" which means it doesn't but for some strange reason the system think it does... which often makes no sense at all
They need to fix it now
eh been an issue since i started using their image gen models 1.5 years ago 🤷
Why
their LLM gets confused and their safety filters are too strict
I’m been trying to make Gwen from SpiderVerse in royal pain
It sure doesn't like copyrigted material
They gotta fix this now
Would be nice but highly unlikely imo
Why unlikely
It seems they are running a rather safe than sorry tactic. unlike image generators like grok which is very easily abusable
I hate grok
me too. Elon got enough money. i only support small startups like google
Flow still not loading past projects media. Anyone else having issues?
All media in every project are not there anymore. In both safari and chrome.
Coming back in here to check .... has anybody yet figured out how to tell the model NOT to add music to the generated video?! I'm so sick and tired of this music being added ... and crap music at that!!!
Try using images of the character in Nano Banana 1 (Nano Banana Gemini flash 2.5 image), Whisk uses it for it's precision mode, or Google AI studio also has it. Strangely sometimes some Flow accounts have it too but it's very few and very rare seems.
Idk if that’s gonna work
it would at least stop giving copyright errors but it can be a little more difficult to use yes.
Mine doesnt have that
Yes this is how my Flow account and most everyone's else's accounts looks like without it, I've already made a thread at support/bugs feedback and it receives no answer.
That's not gonna do it either, that's imagen 4 not Nano Banana. You need a website with the original Nano Banana (Gemini 2.5 flash image) like Google Ai studio has it.
Show me where it is
let me show you. First you need an image of Gwen, preferably character sheet with poses specially back and front, but other images can do too.
@pulsar coral
No copyrights no bs no problems yes it can do unmasked too but you mighy have to find an accurate face of her to use as base, NB1 doesn't search for accurate faces on the internet
Can you do Gwen from SpiderVerse but without her spider woman outfit on
Just her casual clothes
You have to either use a face or prompt the face manually on it's own slot.
Can have short hair blue eyes etc. You can be very detailed about everything.
Notice she has ballet shoes? Yeah I need to prompt that out too.
I meant this design
Wdym
Make a longer prompt the type whisk users have practice with. Flow's NB2 has been ruining our timing.
tbh NB2 also makes those mistakes about the shoes and others, it's cause I asked the character to pirhoutte
For that image bug just click video then back to image and it bypasses the bug. The red box to enter goes then
Behold the power of a Whisk user
... I forgot about the shoes again
But yeah
Also this is actually from Google AI studio's original Nano Banana. See the spark watermark? That proves it. If you want to remove the spark watermark make sure you download the image with it's actual download button, then you can remove it with this. https://geminiwatermark.org/batch-watermark-remover
But it damn allows you to work without constant copyright and censorship
You might also be able to uplod it to flow as an image asset and ask nanobanana to rmove the little corner bug watermark, though I guess that website does it in large numbers (batches)
yeah there's tons of ways, that can also do in bulk
really cool though
I was missing to prompt like this, left it as solid background to be able to do other work on the image
Then the same image can be used to work on other subsequent images
Also I got 0 warnings nor even once
blimey! look at that prompt 🤣 I use gemini to help me craft the prompt then I tweak it over a few generations.
I dont have the patience to go into that level of verbocity 🤣
Also that comes from constantly fixing what I want from other images, still forgot the shoes
It's just trial and error
Also the aspect ratio prompt, it doesn't obey it because it's NB1. But it's there in case... for when NB2 still worked
Its a learning curve, for all the rendering, re-rendering, reprompting, raging, re-rendering I am getting better results now. Just really need to be ruthless and patient
Tho that only works in Gemini, or worked, no way Gemini's NB2 is gonna accept those images and prompt either
Sadly NB1's result is surely better too once you're done giving instructions.
... that's funny now that I notice that spider gwen image does have ballet shoes
It was never wrong then, go figure
..ok then, knowing that
But why is her face different
If you want the exact same face you'd have to use a facial structure and art style prompt, I'm sure it can do it.
Idk her nose and face seems the same to me
But you surely can go completely precise
The eyebrows are thick
I know what it is, it's the chipped tooth
The facial structure prompt needs to state it should be chipped in the same part
Anyway I adde background and selected a different aspect ratio from the dropdown list, can also do that... Flow's only added like 3:4 4:3 and wide since months ago
Eitherway this is the type of thing I usually aim for, if you'd want something in particular you like only you'd know. The model needs to be explained about it
have you made any video's with these image assets?
No but I do still need to upload a lot to a Deviant
As most people I had thought it was safe at Whisk so gotta be on it.
Eitherway ain't anything most other Whisk users hadn't been achieving. There's this guy who creates way better superhero images and showed them at the Whisk gallery.
i tried wisk early on but didnt pursue it. Guess my requirements were for video so not as precise as what you're having to do.
Yea, there are times I think what i've done is alright, then there are a few creators out there that blow me away haha
Whisk's Veo was unusable whatever it was using. Way worse than how VEO 3 is atm. Everyone was there just for the images but it somehow went over Google's heads, despite all the complaints about the takedown.
I did also age that Gwen up, that's the purpose of that prompt to avoid probs at art sites, so yeah she's gonna look different.
It’s very hard
yeah it requires practice
I'm gonna do it for science.
it's really hard to get the chipped tooth thing, I don't think I have the prompt from it
Probably has to be prompted from scratch based on the face's image
That did bulk her face a little
Would have to remove the aesthetic prompt too so it doesn't conflict
But yeah that totally is NB1, and I even got it at first try without errors.
Woe is life
Skin and body type: Green ogre skin alike princess Fiona from Shrek.
it's that simple
... Flow needs original Nano Banana, now. Google Ai studio's works but it's annoying slow to use. Then there's the Gemini canvas pages but those don't obey 21:9 aspect ratios for whatever reason so it's impossible to make wallpapers there.
That's another example of Whisk type prompt but used at Flow's Nano Banana 2
It's not mine I forgot who's is it, but yeah gives an idea on how it needs to be
idk how well Flow's NB2/Pro would accept it now tho, but it was working fine like 3 weeks ago.
Just use flow
That'd be nice yeah
idk I just got practice with this, and my prompts do look similar to that one up there. Many people seem to have developed an eye for it.
After using NB2 for a while it destroys the timing practice you coulda gotten from Whisk
Also about the videos, when Veo 3 doesn't cut it besides other platforms there's also the option of 3D animation and MMD
I mean it's like vids under 1 minute too. At this point is much easier.
Can you do Gwen in the royal pain scene in Shrek the third but it’s Gwen from SpiderVerse in her makeover montage
I think I'm gonna have to watch Shrek movies again. And that's pretty innocent despite what it sounds. She just wears that kind of french medieval wig and sorts doesn't she
If NB2 refuses by a simple image outfit swap cause of copyright yeah probably NB1 would get close enough.
Taking a screenshot from the movie's scene and upscaling it could work as base.
It's kind of like doing this.
Man netflix is gonna lose the Shrek movies soon. Why is everyone like this now.
Hello, please <@&1173646484311384114> team flow can you back the delete button in the sub menu of thumbail, i don't want to archive all images i don't like 🥲
This is is ChatGPT's image model isn't it.
Yep
It's working similar to NB1 yes. Not like NB2... which is hard to poinpoint now if that's a good thing.
It has upped it's game with the writing too.
Being fair.
Might wanna report it at labs-bugs-feedback. yes people have been complaining about it all day.
Ok now, supposedly this is just the updated DALL·E from Chatgpt
It gives me a somewhat enough adult woman, I tried like 4 times and always did it.
And this is what Flow's NB2 gives me with the same prompt, no wonder it keeps tripping on itself about minors.
And that's not even the character anyway so both got it wrong, she looks like this. Guess both can fail at things sometimes.
However GPT's got much closer.
Which pretty much means the end is neigh.
NB2 got so messed up by whoever recently, the other image models are starting to work better. That's where it got.
Not even because the other image models are better than Nano Banana 2 cause they surely weren't, but because of whoever's been messing with it in the past few days.
it’s Getting so much better
Round 2
Using reference image now to try consistency and adherence to prompts and faces. Yeah NB2 still gets it right enough, it aged the character up kind of ugly tho.
or not at all.
There's GPT's DALL·E or what I think it's DALL·E. It's very similar tbh, but there's just something a little more convincing.
Sadly this doesn't give as many free attempts as Flow, so Flow's still got that going.
Also GPT does obey aspect ratio by prompt. Yeah we're getting there.
i'm the only one who get 3 errors each time i generate?
For the "Unusual activity" error clear cookies/change browser and when it works again don't generate more than 12 images at once, aka you can ask flow to generate 4 images at once and press the generate button 3 times, that gives 12, wait for like 2 minutes or till of them are generated or fail to generate then you can try to generate more again.
If you go faster than that you risk getting the error again and you'll have to start all over.
Plus plans and above can generate like 20 images at once aka 5 generation button presses of 4 images at once, but wouldn't advise to go faster than that even if you could.
i guess i'm gonna try clear cookies, but i'm generating videos
That also has the same limits, its more related to Flow than the video and image models.
Back to the images, I gotta try the more complicated NB2 prompts with this.
clearing cookies helped, thanks
I think you do have Pro plan or above, it should be allowing you to try to generate like 20 vids at once without getting that again.
can we please get a app going for the mobile
Oh yeah just reminded me I need to get the pro plan again bc
I’m back limited on generations on flow
hello i would like to say i build an saas erp cloud system and accpeted now from google for startup all i need is review about this system and i want see if there any recommnded
oh hey that works I was also experiencing this
but it started again
Just go slower, like ask for it generate 4 images at once then press the regeneration button 3-4 or times so you don't generate more than 16 images once.
Wait till even the last one is generated or fails to generate, or around 2 minutes more or less, then you can try to generate more.
If that doesn't work make it only 12 at once, it varies according to daily demand.
Sometimes it can go as low as only allow you to generate 6 images at once before getting that error again.
Also once you go slower it remembers you're going more paced and it's less likely it will flag your activity anymore if you keep that rhythm.
So basically Whisk, 6 images generation active at once only
Normal is around 12 or 16 at most. Plus plans and above do seem to be able to generate 20 at once without probs. But yeah if you get that error again just make less at once.
What's the main reason of this? Active generation limitation or?
The limit is actually 25 images at once but it's not good to hit that even in paid plans, it will flag you for sure.
If everyone clicks too fast or refreshes pages it will put strain on the website and cause it to go slow or crash, so they need to put that limit, even tho yeah it's not set perfectly.
That did happen some weeks ago before they put those limits.
Also the limits could also be based on region, for me it works fine at 16-20 at once cause it seems to work better from south america too.
But could be different in europe, asia and sorts.
I think they should set the image/video generation limit to 16 at once at most, that'd keep it stable.
Because everyone thinks they're safe generating 25 at once since Flow itself says that's the limit, but it's visibly less around 16-20.
Like getting this little red message here constantly, that's asking for it.
same here too. You're not the only one
darn gpt, so about to switch teams
I did just get the Flow error about credits too, but you can just push the red button and it will keep generating images like nothing, don't let it scare you.
... it's just one more bug to the pile. But seems harmless.
okay, cool. I'll try it out. Thanks for telling me on that 👍🏾❤️💯
I'll be sure to screen it and report it at labs-bugs-feedback if I see it again tho, everyone should probably do the same before it gets worse... and lately it always does
hmm, it seems like when I click on the red button. It doesn't do nothing. It just gives me this
idk what else I did but did start working for me again after that, maybe try refreshing.
the only way I can do it switch to nano banana pro, and switch back to nano banana 2. That's the only work around that I got for now. Until whenever google labs fixes the bug issue
Yeah maybe that's what I did too. But both models did start working.
I told you
Yeah. And also, the people on reddit had kinda did the same workaround too for now as well
https://www.reddit.com/r/GoogleFlow/comments/1ssc6cz/weird_thing_where_google_flow_tells_me_i_am_out/
Ah good ol' reddit. Well let's give NB2 another try with the same prompt anyway
Eh one or two are ok, but mostly all look samey
Could blame my prompt on that but GPT did get it
Also who the hell is that one in the middle
Being fair left most upper one does look ok.
Good thing for Google that OpenAI is still so greedy
wish there was a all in one with all models
for image creation
I can’t keep up with these subscriptions
So far I've seen going to the source of the models it's still much cheaper than paying around for sites that have them.
It can go ridiculously high about 150% more the actual cost
Ok facing GPT's DALL·E against Flow's NB2 with that spider Gwen aging up prompt from the morning now.
oh by the way, here's another thing that I don't know if this is a bug glitch issue too in google flow. But there's no delete button in the option list when you right click on an image that you want to delete. That's really weird. But if anyone is having this type of issue too. The one workaround for now for temporary time is use the delete button that you have on your windows/mac keyboard on the numpad
You now first have to archive your images, once in the archive you can choose to delete them. Needed to be like this because absolutely everyone was deleting things by mistake. Archive works like a recycle bin now... tho I dunno why not just call it recycle bin. It confuses everybody.
Ok NOW facing GPT's DALL·E against Flow's NB2 with that spider Gwen aging up prompt from the morning. First GPT... yeah this is as big as I'm making this preview, being fair it refused 4 times to make this image with this prompt, even tho it's not terrible, it's probably just the backside pose anatomy.
It's censorship is similar to how NB2's was two weeks ago, or like NB1's is now, a little more than that. But eventually delivers. Now the same prompt on Flow's NB2.
Being fair this is nice and it's more accurate to reference image... but GPT didn't shy on the anatomy kinda in the same way NB1 won't shy either.
And NB2 failed only once so it's visibly still less censored.
idk guess it depends more on what you're aiming for, NB2 prefers huge heads and small bodies, DALL·E still tends to tall small head and curvy like NB1.
Gpt's/DALL·E's does undeniably look like an adult tho, and that was the point of the prompt. One of NB2' does seem more like an adult but doesn't seem it can get the point really.
I’m honestly fed up with Flow at this point. It keeps failing for all kinds of reasons — either the content gets rejected for no clear reason, or I get the “unusual activity” / “activity abnormal” error again and again. This has been going on for a while, and nothing seems to get properly fixed. It’s seriously affecting normal use and wasting a lot of time. I’ve already canceled my subscription because the experience has become way too frustrating.
.. tbh there's too much math involved in surviving all those Flow bugs and error messages while Whisk was working perfectly.
Like Whisk didn't have even ONE of those.
Anyone was able to use a voice ingredient in Flow just yet?
is NB2 getting more censored? Becz some prompts and reference images I used few months ago worked good but not they are not working, it is now saying prominent people or violent their policies...
I did notice it was working better some hours ago, but yeah coulda gotten worse now, that's all the "updates" it ever gets now anyway
Problem is Google is attaching NB2/Pro to every other service they have so obviously they need to censor it, that doesn't work for this site where it's needed for artistry, they really need to put up a more permissive model just to be used for art.
That if they even want money from that, because obviously everyone who paid for a plan for that purpose is cancelling now.
... google photos and google ads seriously, that might be the prob. Who in their right mind
and tbh the censorship isn't really stopping it as much as they think
All that was done was ruin the image models
It was literally working fine a few hours ago, and now it’s broken again. I’m using the exact same prompts, yet now everything suddenly gets blocked or fails for no reason. I honestly have no idea what they’re doing anymore. It’s getting more and more ridiculous. I’m genuinely tired of this at this point. Nothing works properly, nothing gets fixed, and the whole experience is just a mess.
At this point it feels like every update just makes things worse instead of better. The moderation is all over the place, the errors keep coming back, and nothing ever feels stable for more than a few hours. It’s honestly hard to take this platform seriously anymore when the exact same workflow can go from perfectly fine to completely unusable in the same day.
...I'm just gonna keep practicing my prompts at Google AI studio's NB1, I wish they never released NB2, I was doing fine with NB1 and Whisk.
All these 3 months learning to prompt NB2 were just a waste of time since they suddenly changed everything, and I was starting to get the hang of it.
It's impossible to support this,
WHY oh WHY can we not righ tclick open image in new tab? Ffs just give us a way to view the image in full screen without having to download
Can we please get a real update from the team?
Flow is becoming harder and harder to rely on. This is no longer just a minor bug here and there. We’re dealing with “unusual activity” errors, broken upscaling, person-related issues, and a general lack of stability across the platform. Problems keep piling up, but there’s still no clear communication.
What makes it worse is that users keep being redirected to so-called support channels that don’t actually solve anything. At this point, people need answers, not more deflection.
I’m on Ultimate, and I used to genuinely love using Flow. But now it’s actively hurting my business. I have commercial work that needs to be delivered on deadline, and this kind of instability makes the platform impossible to trust.
Yes, project folders and subfolders are a good addition. Credit where it’s due. But honestly, I’d rather go back to the old version with fewer features if it meant the platform actually worked.
Right now, Flow feels less like an upgrade and more like a broken rollout. The team needs to address this properly, because users are losing confidence fast.
... and they did just roll a bunch of updates yesterday that had nothing to do with addressing user complaints, except maybe for the recycle bin which they named in such confusing way that people thought the 'Delete' button was gone.
Exactly. They keep pushing updates, but nothing actually gets fixed.
No real bug-fix notes, no proper update explanations, no documentation, and now more restrictions on top of everything else. Users are just supposed to guess what changed and deal with the fallout.
It’s honestly disgusting how badly this is being handled.
I feel you man. I bought 3 accounts. Feel money wasted.
It’s nowhere near as good as it was in 2025.
At this point, everything is apparently a policy violation.
Same workflow, same kind of prompts, and now it’s just wall after wall of false policy flags. This is getting ridiculous.
Your generations also get stopped? I thought it was a me problem
anyone have experience with the extend mode? I have been making videos like 5-10 minutes long and they have been downloadingup until now, I can't get this one to download it keeps saying fail. If I download the clips separately they don't flow together like they do when you download the 720p clip.
Am new to using Flow... so can anyone tell me when the image gen limit resets for Plus accounts?
I keep getting the 'Failed - You're requesting generations too quickly. Please wait a moment and try again.' I've waitied and retried every few hours for the last 11 or so hours and it keeps saying it...
Something is wrong Meaca, if you get the 'failed - requesting too quick' when you aren't requesting things over and over than you might need to contact support. That only happens when the system is bogged down and you try to keep pushing generations through.
I usually do batches of 100 scenes with 1-4 video generations so I get that message if I go too fast and im on Ultra.
ohhh okay, I see now
thanks, how do I contact support, I can't see anything for that
I would ask one of the google staff when they wake up, if you look at the list of people on the right they are all lit up in green.
OH do you know how to clear your browser cache
yeah, does that help fix it?
Try it, that has definitely helped fix a few bugs on my side.
didn't work... guess I'll have to wait
I rely on Flow for my work, and it simply hasn't loaded any images for two days. We haven't received any response from the team to fix this yet. This is absurd for a paid service.
How to deal with this recurring issue... I have removed the audio "NO AUDIO" but still the same error. Please anyone can help. Thank you
so... are they awake when online or just when their name gets the green (online) dot
it shows a half moon on all their portraits so I am just assuming they are sleeping, I live on the east coast so its 6:40 am, if they are on the west coast its only 3:40 am there.
well am UK so, it's 11:40am
Unfortunately we are like vampires to them
you could always try messaging one of them
so I just pick one of the staff and message them?
thats what I would do if I couldn't make anything, worst they will do is not respond. Best they will do is fix your problem or point you in the right direction.
well I would if the message would go through...
@tough yacht can you help direct Meaca_G
thanks, I dont know the protocol here for pinging people
does the video generations stop at 60 percent and just produce what was finished?
Most of the time the video's generation's percentage is a hoax
do we have problem with flow right now?
I can't create ne project or see my works
Didnt know about than, but why does it change its actions, doesnt do what the prompt is saying? Is it a prompting problem rathen than a generation problem
so I contact Google One support as that's where I ended up... they say they'll get back to me within 24hrs... I can use Whisk, but Flow wont let me gen images but will let me use credits to gen videos... so I'm at a loss
Veo tend to smoothen the prompt and add more stuff to it
Ahhh does that happen with nano banana 2 and pro as well?
Does anyone else have the bug where the 2K upscaler adds new stuff to their image?
AI upscaling basically regenerates the image, so that happens
i get error you need to provide promt ....
I did write promt i tried with image and without image as reference
Still same error on both Pro and nano 2
Hello guys, does anyone know why my Flow is not working when I give it a simple prompt with a base image? I get back like it's not working
I was on the non-ultra life for like 3 days. Tried to use Flow image gen for work (mainly Nano Banana Pro for editing), and yeah… people weren’t joking. I think I only got like ~15 images out before it hit me with that “generating too quickly, try again later” thing. I waited like an hour or two thinking it’d reset, but nah… still couldn’t use Nano Banana Pro at all lol
im back on ultra now today but damn… that 3-day non-ultra run humbled me
Has anyone dealt with this error before? I've emptied all my caches, but Flow fails to display anything I upload or it has generated.
Hey guys, this is actually really annoying. Why is it when I give her a prompt and I'm using either Chapter 2 BTA or, I keep getting errors? Is anyone else experiencing this? It's so annoying.
Same, I've been experiencing this and I don't know what to do. I'm using ChurchEBT to change all but I keep getting errors. It's so annoying. I don't know what to do. I've been asking about and I don't know. When you need this, this is one of the best tools.
grass.gif
I've only seen this one about "error loading your media" when there's connection problems, might wanna restart the PC and router and make sure you're getting all the bandwidth you can. If that won't work and you're sure your connection and PC are fine might wanna file a report at the lab-bugs-feedback area from this same Discord.
Flow uses a lot of ram and bandwidth to load the image and video previews.
That said that's specifically for THAT error message, it doesn't apply to the other 4 or 5 error messages that Flow's been giving.
Do also make sure your PC doesn't have viruses or anything that could be slowing it down, sometimes happened to me with like 80 chrome windows open.
... most of them just Flow tho
Oh uh , i just realized my ultra plan did not gave me 25k ai credit, i only got 12.5k
I just noticed Frontier Modal left this building, no wonder he hadn't been seen around helping lately. God have mercy on Flow's soul if it was because of how things are going lately. but idk really.
More often just come back again the next day around the same time you successfully generated an image, it doesn't reset at a set time but according to when were you making images. And yes that message is like a daily limit.
It wont matter if you come back 5 or 12 hours later, usually needs a little more than a full day to refresh.
Also try swapping from Nano Banana Pro to Nano Banana 2 and viceversa, each model might have some more images left for the day.
Or more like daily attempts, free accounts get like 100-200 free image generation attemtps for the day and the failed attempts do count. Each account tier upgrade gives around double that, I think way more for Ultra.
And the attempts do consume real fast now due to all the copyright and censorship blocks
The chokehold/limit is at 3 back-to-back-to-back now. Crazy.
This has been happening to me for the past two days; even the images I had already uploaded are no longer showing up.
POWAA
investigating
Seems like the "delete" option is no longer there? Is it supposed to be that way?
you should be able to delete your projects from the 'Projects' overview page, this is on the bottom right of a project. if you run into any trouble with this, please let us know in #1238621511401734254
I found a prompt that make your banner like a professional one and i used it on my channel. This is the correct channel to post this?
Thanks for your quick answer, appreciate.
Yeah, no problem there about what you wrote but I meant that with a right click, the option to "delete" a picture / video was available before and now I only have "archive" and needs to go to archive to delete things, was there an update for that?
#flow-videos・📹 should be the right one if u wanna share... wait no are u using gemini?
used ur prompt if u dont mind
archive them, go to archive, delete all. only way I could find
o I was scrolled way up my bad
Let me translate it in English so you can use it 😹
Yeah, I did that but it's extra steps compared to before so I wondered if there had been an update related to that and why~
It makes the use of the " @ " far more cumbersome since I make looo~ooots of images ^^"
also kinda feel like this server should split the flow gallery on discord , like flow-video and flow-image separately?
you read my mind 🧠 been exploring this, and expect to see this soon
there was recently an update! sharing your feedback with the team
yea I'm in the process of trying to organize better myself I have too much lol
"make an art, a design of this image (insert your avatar) with the name below (insert your name) with a description that it says (insert your description). Advanced design 3:1, for a Twitter banner"
It should work
Thanks a lot ; )
I don't mind the archive since it makes it possible to go back if you made a mistake but having linked to the " @ " function makes it very bothersome to find the correct ones ^^"
Yeh, me too, feeling is shared ; )
The Daily Credits you have in the Free Flow subscription is 150. It states that they are reset on a daily basis. But my subscritpion has been on 140 credits for a counple of days now. It doesn't go back to 150. Any idea how come?
This is new to me. Why I get this after generating 10 images. My daily credit is intact.
Anyone else seen it?
Request for those seeing credits errors: aka 'you need more credits' - more info needed plz
hi all - I've passed these along to the team as we've seen an uptick of these reports (thank you for continuing to report 🙏 🙏 )
for those effected by the "you need more credits" notice, could you please (1) emoji react to this post with a ✅ and (2) please send me a DM (this is my only account) of your account email so we can take a closer look? I'll be replying to posts in #1238621511401734254 asking for the same. plz. thank you
m2 🙁
seems like if you refresh, you can do it once or twice to generate and refresh again
It works like Georgia says, but you can also delete your images by first sending them to the archive, once there you'll see the delete button for them, it works like a recycle bin. Seems to be like a new security measure because users kept deleting things by accident. One of these days you'll thank it's still there.
Yeh, i'm aware of that~ It's not that part that I had a problem with.
Yeah but you just gotta imagine it doesn't say "archive" but "Recycle bin" then it makes more sense.
Projects can be placed inside one another to make more folders... so an archive was kinda redundant.
Yesterday people were fixing this "You need more credits" bug by swapping from Nano Banana Pro to Nano Banana 2 or viceversa and then generating an image by just pressing that red get more credits button. It did happen to me too and it got unstuck that way somehow.
right make sense, refreshing seems to work for me
Yeah I did that too, tho eventually it did get completely unstuck.
... which still is weird
I swear tinkering with that red button made it pop like an old bottle's cork or something and it stopped bothering me... I imagine it's not gonna work the same for everyone, because it rarely does around here.
looks great but % processing is dead, you can just delete it from video and this would look amazing
Seems the credits count has gone crazy in all sorts of ways for everyone tho
Really just when one's wondering what else could happen
how can I find the achive and why can't just deletestuff?this makes no sense. Also, why isnothing generating today?
Archive was added recently and Google Flow terms were recently updated.
@main saddle and @sullen onyx I have created a ticket as requested.
Name: Flow Credit Error / Warning
Sorry I replied to your post, as you have alrady started addressing Hcat's query about Archive.
no worries
Just imagine the archive is called "Recycle bin" instead, that's it's actual function so people stop deleting things by mistake cause that would happen a LOT.
Also your generation problems depend on the error message you get, some have explanation and you just need to wait till it again works next day, others can be fixed in some way or another alike by clearing cookies.
Maybe you haven't seen it or it hasn't happened to you, but I assure you that you'll be glad you didn't lose a thing deleted by mistake.
Thank you for transferring my libraries to FLOW.
The pending doom of manually transfering everything would of ended in tragedy for me.
I am looking forward to FLOW bearing some kind of resemblance to Whisk someday...
Really did not want to move over to gemini or dreamina.
Am very Greatful for the time I had with Whisk.
Very impressed with access to project image directory in the prompt box. Going into refine a generation and selecting another image in the prompt would be the only means I've encountered to combine 2 images. Unfortunately, all actions to converge two selections result in the system crashing.
I guess it's altervate to paste
i would like to see the ability to navigate from one generation to the other after clicking the individual videos. Also, if possible, I would like to delete them or favorite them as I examine them individually please.
Hey there
Is this new? I requested for this feature few weeks back, but it's still not available to me.
🛠️ Fix deployed: AI Credits errors
Flow team has issued a fix that is rolling out at this time, please note that it can take a minute before it reaches your account. thank you Flow team, and thank you to the community for continuing to report your experience. appreciate y'all.
Thanks for the quick turn around @main saddle. Would you be able to share what caused the problem?
Bro I'm not done rewatching Shrek yet. Need to get to it before it's taken from Netflix so I understand what to generate here.
So sick and tired of not being able to use images because it keeps mistaking it for a prominent person. I get why its there but you guys made it way too sensitive.
Yeah that's why I only make images of Shrek now, he's national property
Are those Disney characters? oh gee how did all of them got there at once, ain't that copyrighted, shame on you.
What you think I struggle a lot? Takes me 5 mins
The scales are the harder part
Project date's there, imagine ruining the Nano Banana 2 model for everyone thinking you're blocking copyrighted stuff when it actually can't be done because of the extense amount of variations and exisinting terms..
And no I'm not putting the prompts so they get patched.
So, she said,
"What's the problem, baby?"
What's the problem? I don't know · Well, maybe I'm in love ♫ ♬
Must've gotten scales better there but it's a puzzle. Dunno if Ralph is really that huge.
🫡
This clearly isn’t just a normal bug anymore. It looks like an overactive policy filter that keeps false-flagging perfectly normal generations, and there’s still no clear public explanation or ETA from the team. That’s the most frustrating part.
Waiting longer does nothing. Retrying later does nothing. This issue just keeps happening, nobody seems to care, and nothing is getting fixed. This is beyond frustrating.
Yeah you could say these bugs are ogre-whelming.
Wouldn't bother that much anymore, any minute there's gonna be something better, this year even. Here they're obviously not gonna fix the models anyway, all feedback bounces off of them for some reason.
After trying open ai's models I'm sure there's gonna be something better.
It's the usual thing with Google, too late again they will see it's impossible to bring the customers back.
What makes it worse is that no one from Google seems to be acknowledging this or giving any real explanation.
Everything that could be said to everyone it could be said to and everything that could be done has been done already. Now we wait for the killer beees.
Yeah, honestly this just makes me hope Sora 2 comes back sooner. Google keeps doing the same thing over and over: launch too early, ignore the obvious problems, and then act like users will just keep waiting around forever.
smh
Hi! In Whisk, we had a 'Base Character' section to keep characters consistent across all generations. How can I do this in Flow? Is there a way to 'lock' a reference image so I don't have to re-upload it for every prompt? Thanks!
Yup, same here, every time I use flow to make a prompt it gives me the same issues, and none of my prompts have anything that violate the policies so I stopped using it and went to Higgsfield
I felt like Flow team has finally getting better.
Btw the Archive function is good, you can archive images, and delete them if needed afterwards. So no trouble on accidentally deleting assets.
scratch that, here we go again
Hey anyone with a premium flow account that could help me with (probably add as a member I have a school project I need to do 😭 )
is it me or nano banana 2 looking a bit bad today
Hello. We can’t log in to our Google account. Flow will automatically renew the membership, but it does not allow us to log in. Does anyone here have an idea of what we should do?
Can I see the prompt
Yay got my 25k credit ❤️ thank you!!
Hello,. Can I ask something? I generated some image and I add this image as reference, but I want other new created images in this illustration style, but no content of image. When I write to prompt "Use same style", new created images have all same. persons, environtment. But I want only style, no characters from reference image. How can I create this correct prompt, if I want use only illustration style, no characters for new images? Thank you for help.
new error unlocked without doing anything related to it
You can try Generate an image with the same style with different <details>
Hi! Is anyone having trouble uploading images? I keep getting errors.
same
I've been trying for days to delete four empty projects in flow, but everytime i refresh the page they come back. I've tried everything i can, and nothing works. can anyone help?
this happened to me the other day, it took almost a whole day to resolve itself
Can we talk about Nano Banana only making Napoleon with Joaquin Phoenix face?
That happened to me a few times when I was trying to generate an image of my OC
Like who does he look like?
Asked for large businessman in office chair in congressional hall corridor. It keeps giving me Chris Christie..
Ofc I won't let me push it through Veo3 as it suddenly has a guilty conscience and flags it as a prominent figure..
You sent them to the archive and they can't be deleted there? Cause that's how it works now you first have to send there anything you want to delete.
no I just pressed the delete button on each file
Just try that and send all the stuff you want deleted into the archive then delete it from there.
I'll try
Yeah.
Apparently, all the errors in Nano Banana 2/Pro happen because Google has to protect us by enforcing Google’s policies on behalf of Google itself.
Kid you not I was just told something alike at the bugs support area.
Oh sorry for saying kid, Google, I mean I legally adult you not.
Im not an adult yet
I want to become a SMM using Google labs xD
Unfortantely, he lost as he was actually still a highschooler.
While TFS hasn't made an Abridged version of the Buu Saga, they made some abridged clips of it for TNM.
Original Video w/timestamp: https://www.youtube.com/watch?v=Ju2W0Snbm8Y&t=3760s
Soundtrack: https://www.youtube.com/watch?v=7yTt5jGEoaM&t=130s
The clip and soundtrack was made by...
But do you have any solution to counter the 99% problem cause it infuriated me
It depends on each error message tbh, but at this point everything's just too much math
It's not a new error. You have been rate limited.
Try again in 12-24 hours
Idk if that's the error he's having but yeah that does solve at least 3 of the errors.
But I am not
I can still generate but not edit whats generated
i can't figure out how to send my empty project files to the archive
It stuck on 99% for a long period (more than 5+ min)
And it don't have any error text
Then please raise a ticket
And it does happen to all the videos/images?
No need bruh
Yes all image (i don't use my precious credit)
it depends really, if happens to all his stuff all the time he does need to post at the labs-bugs-feedback area.
Because I don't think he didn't just refresh and try again
I refresh
Still don't work
Could be related to that new problem about the credits, Georgia did say they've been rolling fixes to it but it's not working yet for everyone yeah you might just have to open a thread at the labs-bugs-feedback area. Also you could try clearing cookies or a new browser, that sometimes fixes things.
But I haven't seen it fixing that one yet
I've tried that to solve my problem, but it sadly didn't
anyone got this?
No I got the 99% loading screen
This is the one where you need to wait. Try swapping models from Nano Banana 2 to Pro and viceversa and you might have some more images left, if not it just means you hit your daily limit, you will be able to generate more images in that account tomorrow a little later than that the same time you were still able to generate images the past day.
Free accounts give you around 100-200 free daily image/video attempts for NB2 and around 10-20 for NB Pro, the failed attempts count. You get around double that for each account tier upgrade and surely way more for Ultra. It all depends on daily demand.
can you tell me how to put them in the archive?
oh ok makes sense, i hit the limit on whisk a few times, nb2 is working, its a bit confusing, maybe it could say 'You hit the generation limit, please wait about a day'
Never seen a screen like that tbh, you should be seeing something like this.
Maybe you have a lot of projects there and they aren't loading yet, might wanna check if your connection's fine or reset it
that's because there's nothing in these projects, they are empty.
i deleted everything that was in them, because i couldn't delete the projects
my connection is also fine
this is also not a problem if i make a new project now, those i can delete without them coming back
Yeah if you can't see this little button here or open the project your account is just having some unknown bug. You'll just have to opean a thread at labs-bugs-feedback to inform it, they do sometimes shake things and they start working again.
i can see that button and use it to delete them, but they just appear again after refreshing the page
Yeah also sounds like some sort of cache issue, did you clear cookies? But do tell them that too.
I've done this now, i hope it gets fixed
I am so tired of Flow or VEO adding background music to my video's as I tired everything to remove it as traffic from outside the inside a store and it keeps adding music as I find music in video's is a distraction so can we have a option to remove or adding music to video's
so then just edit it out 
what's the deal
Tired of unusal activity error and even cooking fried rice violate policies. 🙁
fried rice... well I'm not surprised
hey people, what are the credits used for in flow image generation? i see like 150 credits in the free tier.
how can videos longer than 8 seconds be created in flow?
0
huh>
Image generation costs 0 credit
However, you will still encounter the daily quota limits
It’s only video that costs credit
i was thinking about buying the google one pro, if that helps increasing the daily quota limits or if it helps with some things in image generations (which you said are none)
If you never subbed to the Google one before, you can claim the one month trial
roughly how many does one can create in the free vs the pro one
ahh gotchu! thanks for the help! im prob gonna buy it tho
Wait let me see if I can still snag the four month trial offer
Nevermind it’s not valid
hmmm no worries!
ooh. Wrigley boys to let u in fazz
Working fine from South America atm, Flow seems to have a lot of regional problems. Might wanna inform it's down for you at the Discord's labs-bugs-feedback thread.
thanks man ! but ive already solved the problem
idk why it goes offline at random and at random places
idk thats the first time with me tho, ive waited around 20 min and the problem got fixed by itself
It's ^. North American
Yeah but USA is a big place, a lot of states might work differently here and there, hell hawaii counts as USA.
Cena 1: Close-up cinematográfico de um copo térmico moderno em tom rosé sendo colocado suavemente em um porta-copos de carro, iluminação suave refletindo no metal, câmera faz um movimento de aproximação lenta destacando textura e acabamento premium, interior do carro com detalhes elegantes, sem texto, sem logotipos, sem legendas, foco estável, movimento coerente e estética de frame de filme em movimento.
Cena 2: Plano detalhe da tampa sendo aberta lentamente, revelando o canudo de aço inox, leve vapor frio saindo do interior indicando temperatura preservada, iluminação contrastada com fundo escuro, câmera faz um travelling lateral suave, sem texto, sem logotipos, sem legendas, foco estável, movimento coerente e estética de frame de filme em movimento.
Cena 3: Mão segurando a alça ergonômica do copo enquanto o carro está em movimento, vibração mínima demonstrando estabilidade, luz natural entrando pelas janelas, câmera em plano médio com leve movimento de acompanhamento, sem texto, sem logotipos, sem legendas, foco estável, movimento coerente e estética de frame de filme em movimento.
Cena 4: Close no líquido gelado dentro do copo com gelo cristalino, condensação realista na superfície externa, iluminação dramática destacando o frescor, câmera faz uma órbita suave ao redor do copo, sem texto, sem logotipos, sem legendas, foco estável, movimento coerente e estética de frame de filme em movimento.
It doesn't reset at a set time, but a little later around the same time after you generated your last image successfully yesterday. And it's still at labs-bugs-feedback cause they could indeed make this more clear for everyone.
Also Pro sometimes resets faster than NB2
Yeah using screens is a good idea cause this Discord won't allow to link to it's own threads directly by users.
Hi
Is this a bug? I've been getting a lot recently even though I'm using the same prompts I did before
Depends
If you wrote a prompt which is against Flow policy, it doesn't generate videos
So, the Whisk transfer to Flow from my account is done now too, I think all galleries went through. But more than all when your images also transfer to Flow already, you all can check this icon at the images and it will tell you what model was used to generate them.
So yeah as you all can see here this was correct for sure, the model Whisk uses for precision mode is Nano Banana (Gemini 2,5 flash image) without a doubt and stated by themselves, can't be more clear than this. If you can't make the same images from Whisk at Flow it's because it only has Nano Banana2/Pro but lacks NB1.
And this is very deceiving to people, because they have been moved to a completely different website which lacks the tools from Whisk. It'd really be good to adress this.
It wasn't imagen 4 etc for sure.
If you make an image with Imagen 4, it will also distinctly tell you if it was used.
great transfer guys. really. i lost all my images. i wish there was way to download all of them
i only saved a portion of my collection on my laptop
They did gave u chance to “download all”
Instead of migrating
Also I used one of your stuff https://labs.google/fx/tools/flow/shared/video/3bbbf73e-fe55-43e2-86c2-6822c52d2623 @twin raptor
Question: If you are using an image youy created with Flow as ingredient for another image, does the generation work better with the 1k, 2k or 4k version?
mine didn't 🤨
yeah when i enter the site, i had a download button for 3 seconds then it was refreshing itself to "transfer your media will be available". then just a week ago, it was only transfer.
Well mine has transfered successfully
Clicking the reload button on each image doesn't work? I had that happen to some of my transfer stuff, but the little reload loop arrow button brought them up fine.
Hi guys, i have a bug on google flow, where the screen dies not load all my projects stay black. And when uploading a picture it gives the error: something went wrong loading your media
Anyone knows how long I need to wait when receiving this msg? u_u
How do you guys keep the voice consistency across the entire video?
I see that some ai videos doing it with one character in the video.
2.30h later and it still gives me that same message u_u
hi, I have a question so for the pro plan the videos will have a watermark
is it the same thing for ultra plan?
ultra plan wont get any watermark
At this point, I’m honestly exhausted and frustrated.
The response I received was just asking me to provide more information, but this issue has been happening repeatedly for a long time, and it’s clearly not just my account. Many users are seeing the same “unusual activity” failure in Flow.
I’ve already tried waiting, refreshing, restarting, clearing browser data, using different sessions, and reducing how often I generate. Nothing actually fixes it. Sometimes it works once, and then the next generation fails again with the exact same message.
I understand that support may need details, but at some point this should be treated as a product-side issue, not just something users are expected to troubleshoot endlessly.
Please escalate this to the engineering/product team and provide a real update or a reliable workaround. This is seriously affecting normal work, and simply asking users for more information does not solve the problem.
Hello, is it possible to adjust the default collection creation settings, which tend to behave a bit erratically when moving items to the prompt at the bottom of the flow? Thank you
Exactly. At this point, it feels like Flow has multiple ongoing bugs at the same time, but there doesn’t seem to be any clear acknowledgment, update, or timeline from the team.
Users are reporting different issues — unusual activity errors, unstable collection behavior, failed generations, inconsistent asset recognition, and broken workflows — yet the response is still mostly “please provide more information.”
We need the team to acknowledge that these are not isolated user-side problems. These issues are affecting real workflows, and many of us rely on Flow for actual work. A clear status update, bug tracker, or reliable workaround would be much more helpful than repeatedly asking users to troubleshoot the same problems on their own.
video
OK flow you now made it not worth using you . What State are you in as it's in the 1700 as when I try to creat a Simwear Shop it's SEXUAL ? Even Undergarments shop is sexual ? I suggest you never go to the mall as it may offend you seeing swimwear and a bra in a window. Last month since you killing Whisk you go all strict as soon all we can create is teddy bears while other AI out there will take a famous cleberty like Tayer Swift and have her posing nude as go to Deviant Art you will see. And I just creating retail display and that it sexual ?
look at the result despite extensive prompt by claude and 3 iterations
There is an expirimental feature called "Voice Ingredients", where you can use ingrediants to upload not just an image, but also the choice of a voice. You can read more in the announcement section.
Also I shared a couple of "how to" videos in the flow-tips-n-tricks channel
video
I kept noticing that each time you edit something with NB/NB2, the image are coming back with more JPEG compression artifacts. This is bad for cartoon/anime. :/
/whisk prompt: a traditional gaucho atelier with wooden textures
/whisk prompt: A medium shot of a man carefully crafting and holding this traditional cuia (1.jpg) in this workshop setting (2.jpg), hands in focus, cinematic natural lighting, artisan atmosphere, highly detailed wood textures, 8k --v 3.1
Glad you liked it!
It seems like you didn't provide a prompt. Please go ahead and provide one, and I'll be happy to assist you!
well but you cannot use a custom voice there
You're not gonna believe me guys...
I just found out you can use left/right arrows to move between assets now . that’s actually so much faster to review stuff
Pls I need help, how do I fix my characters saying their lines,not another character line
Here is my prompt for review
Camera:
Fix camera, no camera motion
Music:
Engaging Afrobeat piano with soft drums, low volume, fun and stylish.
Background Motion:
Mall glass doors open slowly. A few shoppers pass behind. Palm leaves move gently. Luxury car light blinks softly.
Character Motion Before Dialogue:
Bisi adjusts her sunglasses and smiles proudly.
Kemi lifts her shopping bags and laughs softly.
Tola holds her notebook calmly and smiles.
Dialogue With Accent Lock:
@ Bisi in pink— fast Nigerian female accent, playful and loud:
“Tola! Ah-ah, babe! This your simple dressing again?”
@ Kemi in white and multicolour jacket— fast lively Nigerian female accent, teasing and laughing:
“Enjoy small abeg! You no go fit get man with this your serious look ooo!”
@ Tola in black gown — fast calm Nigerian female accent, smiling gently:
“I am enjoying, Kemi, I just don’t want enjoyment to punish me later.”
Thanks in advance for positive feedback
Lad, I need to generate some images, but could these massive problems be the cause? Come on, can you sort this mess out!
Is there anyone here who’s really good at using Google Flow to create thumbnails for YouTube using Nano Banana 1 or Nano Banana Pro?
Voices
Hi there! 👋
Just a question:
Where to report a bug in Flow - in the "Google Gemini" server under "bug reports"?🤔
#1238621511401734254 here
And another question (since it's still possible to delete images):
Is "Delete" the same as "Archive"?
i was wondering this too! delete button is "deleted" lol
i didnt use archive one since idk what it does... buuuut
Thank you! I'll do so... 🙏 👍
selecting an image with your mouse (or more images) and pressing del on your keyboard is a good workaround btw
Same here ... but since the "Delete" option in the context menu was replaced(?) by "Archive", I'm assuming "Delete" was imo rather some sort of archiving files.
Done 😉 Many thanks again and enjoy your Sunday! 😎👌
?
thank you. btw! i tested the archive mode, it's basically like a recycle bin, where you discard stuff, and you can restore them back or delete it forever
idk if they get automatically deleted after like 30 days or smt
Ohh... great to hear! Haven't tested it yet... will check it out soon! 👌
Hmmm... how do you access the "Archive"-mode?
I can only access the "My Library" (https://labs.google/fx/library) which only shows videos... 🤔
Or am I blind? 🫣
Pls forget what I said.... I finally found it. My bad.... 🫣
For anyone who's also looking for the Archive folder (maybe it's only me?).... it's here:
Does anyone get this? and then you can't gen with pro for a day or longer? I generated using this once every 10 or so minutes then suddenly it says I've been doing it too quickly.
Just wondering if there is a actual limit or not instead of this message and I'd like to know thanks! ^^ have a lovely day/night
I do get this also several times... just let it cool down and you're good to give it another try.
Sometimes it even works (at least for me) if you request just 1 or 2 images and not 4 in a row.
Thanks man, took me a while to find it too.
Hi, I’m new to discord.
If have idea how to reduce ai hallucinations, where can I share it?
Well, I made these with Google flow
I need help. I've accidentally archived a couple of images from my Flow library. How can I retrieve them?
Did you see my post from today? 👆 Maybe it helps? 🤔
any ideas?
I just encountered this problem too, and it's completely unsolvable. It's incredibly frustrating and completely disrupted my train of thought and ongoing work. Google really needs to fix this immediately; it's a serious issue
I just tested a method I found on Reddit: switching between Banana2 and Banana Pro solved the problem.
👍🏼
I think I found the cause on mobile end—it seems the Flow web app loses focus and kills the image generation if you switch apps, even for a second. It looks like it doesn't have background processing capabilities, so you have to keep the tab active until it's finished.
Yeah I've also seen the device needs all it's available ram and connection power free as to get that error less, I've only had that one when my connection was busy with downloads or like 100 chrome windows open... which yeah at some point even 128 ram won't hold.
Might seem not but Flow uses an insane amount of bandwidth and ram.
It's not a warrant tho, Flow will indeed sometimes give whatever type of error message for unspecified reasons even tho the error is supposed to be fixed in whatever specific way. It's a box of chocolates you don't want.
Eitherway you could still try restarting your device and connection, might do something,
for some reason using nord vpn on usa servers fixed it
That's another yes, some countries seem to be given better connection to it than others, the website won't even load at all in certaion regions sometimes.
But more likely that one error always seems to be centered on some connection problem of some kind or the device not having enough free resources atm
Seems you're on a desktop platform 🤔?
yep
Then again Whisk virtually never has that or any of those bugs either, and its Nano Banana (Gemini 2.5 Flash Image) model, which has been demonstrated to be used for Precision Mode, is perfectly usable as it is, and people were actively maintaining subscriptions for it.
They give users no explanation for their reasons behind making seemingly irrational illogic decisions...
If, for some reason, Whisk was no longer viable for them to operate, then at this point they would be better keeping Flow and still generating a site similar to Whisk dedicated solely to image generation—preferably managed by a completely separate team that is more empathetic toward users and is not centered only in VEO, listens to feedback, and keeps the original Nano Banana model available alongside NB2/Pro just like Google AI studio does without any issues.
why are there so many google branches
there is labs, studio,
(im being stupid rn forgive me)
I just woke up
Aye is anybody else having this issue with flow like auto logging you you and having to click the create woth flow button over amd over again.
Yeah it does log me out often too, and considering how long a Flow project can take to load that's not practical to work at all
I still don't understand the purpose of Flow, and trying to use it for video creation is a mystery. It seems unnecessary.
Before it got messed up this month NB2 was giving me images with skies like this, and I sure was using it. They seriously need to roll back whatever they did at the start of the month, and even better fire whoever did it