#flow-general・🌊
1 messages · Page 41 of 1
why am i getting like this whats the fix @patent stone or any mods pls helppp
🥺
@fervent bane
Help usss...i even got the google pro...which we students got for free
On Gemini, we can create a video from multiple images in a prompt. Is this possible either in Flow or Whisk? I was only able to generate a starting image from images in a prompt, but then when I am to generate the video, I couldn't include more images as a reference for characters that will show up in future frames.
Please do not spam/ping moderators and google staff. Us moderators cannot help with any technical difficulties with google labs etc
lmao its so crazy that they ask/feel entitled
sed
how to fix this, guys?
help
how can I download 4K ,it say I have to upgrade. but went I click they have only option to buy more credit but now I have 3,000 credit , by the way I gain credit from google drive 4tb and ai package
hey, I was using frames to video for a while but since yesterday it doesnt work. I tried 10+ prompts and even the prompts that was working before didnt work. Always getting this "Couldn't generate video. Try again later.". Is there anyone experiencing the same problem, what is the solution?
👋 I noticed Flow becoming available for workspace as additional service; do we know when / if it will also be covered by enterprise protections ? I can see some folks finding it useful enough to even consider going to Ultra subs (they use Pro w the enterprise standard) license rn.
anyone know how to fix the changing voice when generatinga new vid
i want the voice to be the same but veo 3 kept changing it
Can someone more educated than me explain to me why a Pro user (Google AI subscription) must have the Veo logo and 9:16 video with black bars. Thanks and have a good day (note: I'm being sarcastic).
this crap again doesnt work
Hey! Which "enterprise protections" in particular are you looking for?
Franco, I know it was a sarcastic question, but just making sure you know that Ultra subscription has no Veo logo on generations. For the black bars, is that happening with you start with a 9:16 frame or a text-to-video prompt?
Hey Zarn, the two things I do are 1) Use "Extend" (helps if the voice is in the last 1 second of the video you are extending and 2) use the most detailed prompt you can to describe the voice, and keep that voice description consistent from clip to clip
4K video downloads are available to Ultra subscribers; if you're having trouble getting to the upgrade page for Ultra, do you see it here: https://gemini.google/subscriptions/ ?
Is this still happening? If so, is this on videos you created recently, or videos you created a while ago?
Hey! Which "enterprise protections" in
So the pro users, who pay... are (fewer users than the ultra ones)... honestly I only use Flow because I can use Nanobanana without logos, for the video I set the format to 9:16 in text to video, and out of 10 videos, 8 have black bars. If I may say so, I find it quite comical. Anyway, have a good day, I have plenty of alternatives.
skill issues
Aaand here we go again with the "Something went wrong" responses from Flow., literally no reason whatsoever.
hello..
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. we have #1238621511401734254 for feedback specific questions/ queries
I have fortunately found a Fix, I had to reset my Browser and it works now. So basically after every couple of prompts it glitches and then that error
Voice consistency is an uphill battle with Veo. Replace the lipsync audio track with Elevenlabs voice clones.
how to fix ?
also the voice generations from the Fast model can sound "bubbly" and distorted. Use Elevenlabs voice clone, it will usually work with lipsync
Hi guys, sorry to bother! Does anyone know if we can generate videos without the watermark? My images come out clean, but the videos have a 'Veo' logo. Any help would be appreciated
Ultra subscription
Got it, thanks for the info
Does anyone know why i got charged for Veo 3 Audio Generation (output) when i used flow and used 3.1 fast and 3.1 quality to generate video and extend video? I thought it used the credits within my google pro subscription
is there any tool of AI which can Inpaint a video to add an element or animal into the video
veo 3
Hi @muted olive videos aren't generated in the discord server. In order to make a video you'll need to go to https://labs.google/ pick the tool you want to work with then go from there. If you need help with prompts please ask in the appropriate channels for advice. Thank you!
Hi @hearty jetty videos aren't generated in the discord server. In order to make a video you'll need to go to https://labs.google/ pick the tool you want to work with then go from there. If you need help with prompts please ask in the appropriate channels for advice. Thank you!
Guys, help...my videos can't be loaded in flow. I have some 20 videos below and suddenly only top new ones appear. This isn't first time but this time it wont load the vids below
Again 1080p upscale is not working. When clicking that option, no response, nothing is happening. It's really very diisgusting that this same problem is occuring again and again often.
Credits are how you're charged.
Veo can do that if you generate it in Flow.
is there any documentation how to do that as I have tried but did not worked might me doing something wrong
Hey guys, hope everyone is doing alright, has anyone done any parkour videos using VEO 3.1 recently? My prompts are failing somehow to generate accurate outputs.
Yeah I’ve done it just now and it looks like a
It’s bad , I’m not sure if the issue was with my prompt.
Every time I want to generate an extreme amateur fast-paced video like parkour , veo 3.1 decides to make it a slow-motion pacing video for no reason. At this point, I’m not even sure what word in my prompt triggers it.
And it tends to make the subject doing parkour doing a Mario double jump too lol
Not that I'm aware of.
its the watermark for me
Are you the master of KINE?
(also... what's KINE?)
same issue i have
Every time they release something new, it leads to bottlenecking and spikes in fails. Might be even worse than normal because Genie is insanely resource hungry.
Mine are horrible seriously, I have the ultra package and literally with the quality generation it generates some random video its so bad
Not sure if its the prompt or whatever it is but its just too AI slop
If you use text to video, you need to really put work into your prompts. Long, in-depth prompts if you want high quality generations with any regularity.
If you're using images in frames to video or ingredients, then you don't need nearly as much detail in your prompts, but still need to know how to use the tools to get good results.
There are no shortcuts. You can get good results from several high end models now, but they all have a learning curve, and there are no shortcuts.
^
Thank you, KINE MASTER
I literally just googled it.
They must have a bunch of that stuff on YouTube in the training data.
Apparently kine is a plural of cows... the heck. I thought it was just high quality green.
I wanna know how they were training the VEO model because VEO 3.1 just… it’s hard to identify what word/keyword we need to put in the prompt to get specific kinds of video. it’s like RnG at this point
For eexamplea few days ago I was trying to get Veo to generate the legitimate, genuine inside-vehicle dashboard cam footage style
I don't know what RnG is, but prompting really is a moving target, and camera controls went back to being terrible.
Out of all 12 generations I did, only one of them actually managed to look like those actual real amateur footage of dash cam vehicles
I don't even waste my time straight text to video anymore. Even if it was free it's just too much time-wasteful. I go straight to Nano Banana or Leonardo.
For real, I see 3.1 straight up being stubborn when it comes to camera controls it’s really bad
I specifically prompted digital zoom in n out , but it gave dolly in and out instead
With these motion control models coming out, I'm going to reconfigure my living room/office so the area behind my computer desk chair is empty, and I have a blank white wall to film stuff on. I have blackscreen and green screen sheets and rig boxed up from a project somewhere, but I'm not sure that's necessary anymore.
It was getting better, but really tanked in 3.1
They added those camera edits to already-generated clips, but they don't work well. Worse than the old ones they got rid of for some reason.
I still can’t seem to find actual use cases for the camera edit feature/reshoot tbh
The best camera control function from any major model is still Kling 1.5 Pro's ball gimble thing. It's a shame they didn't keep that in their new models.
And it doesn’t even give wide variety option to choose
Speaking of Kling ,I heard 3.0 Kling is out?
Yeah, there are fewer options than before.
Just an announcement that it's coming. Sounds like their version of trusted testers are playing with it.
Welp, my expectations ain’t that high now these days when it comes to video AI generation.
Ikr
I've been more focused on my music. Finished the clips for a music video, but need to finish the album first, then edit the music video together.
Answers to those questions are all pretty predictable. What matters more than everything else is how good is it. Does it make what you ask for at a higher level of quality, with fewer fails, etc?
I NEED THIS FOR FLOW VEO
I didn’t know Kling o1 is capable of doing key frame generation like that
Kling's might be better, haven't used it, but that's basically what ingredients does.
This isn't a bot server.
I have a flow pro account, I have been using it for image generation. I have not seen an official image generation limit mentioned anywhere, on googling it says I can generate 100 nano banana pro images with pro and 1000 with ultra. But I just generated 128 images and still have quota left, is there a specific image generation limit using nano banana pro in flow?
Those are daily limits. Either you generated some of them before your daily reset, or it's not a hard limit, and maybe they're flexible when there is extra capacity.
so officially the daily limits are 100 and 1000? (for pro and ultra in google flow)
Hey Any tips on using flow veo 3 with reference cuz I have tried both ingredients to video and start and end frame but not getting quite the kind of results I want, I have even tried using veo 3 gpt but no success with it 😔😔
same question
Yeah
Bots are pretty bad at prompting.
<@&1173646484311384114> so frustrated at the amount of failures today. I can generate anything. Failure after failure despite being on the Ultra plan. What is going on?
Don't know if you are calling me bot or the gpt but it's ok, can you tell the correct way ? I'd like to learn
It's the only part of what you said I could respond to. The rest is too vague to answer.
And I didn't call you anything. Come on now.
I'll try to be more specific
I'm trying to make a video using veo 3 using reference images but the things are getting messed up like change in the product form, different face, etc etc
Can you help me with that? Like there a specific way of prompting that can help

Find tutorials on what you want to do published recently (in AI video ~3 months or less is a good rule of thumb). More in depth the better.
Buy as many credits as you can afford. Test. See what works. Take detailed notes on what works and what doesn't (and never stop). Practice. Figure out how to use all the input options and settings, and other tools in conjunction (image, audio, etc. - both AI and not).
Repeat.
Ok thanks I'll try
Hi all! I'm new in video generation.
Is it possible to add an external video that I've created also with Veo3 to the Flow scene?
sadly nope
oh i know this onee
You know the solution mate?
kinda ! hold on im creatig a infographic style guide for that
FLOW IS NOT WORKING> WHOEVER IS IN CHARGE... >> WHY DO WE GET IGNORED WHEN WE PAY FOR A SERVICE 250$ PER MONTH AND ADITIONAL CREDITS >>> SINCE VEO3.1 - VEO IS UNUSABLE!!!
SUPPORT IS USELESS THIS CHANNEL IS USELLES
google actually posted their official veo 3.1 prompt guide https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1
Ohhhh they also have some things on their Google cloud page I just saw
For the image-to-video section, I believe I came across it on a different website. The guide was also created by Google staff
https://medium.com/google-cloud/veo-3-a-detailed-prompting-guide-867985b46018#:~:text=Image to video)%20Use%20a%20high%2Dquality%20source%20image @burnt shard
Damn you found this, thank you very much man
The author also created her own ‘prompt assistant’ tool, and the app itself is based on the guide she wrote. 😄
To use the app, grab a free API key from Google AI Studio, paste it into the box, and it’ll validate it for you. Once that’s done, you’re good to go
CAN WE GET SOME INFO ON THE FAILED GENERATIONS? <@&1366831844502274180>
It's getting annoying to pay a lot for a service and you cant use because u are limited because you failed us for the cheap users who don't pay but overuse ur service with different workarounds
mods are only here for discord support. please only ping mods if you see someone breaking our #rules-and-guidelines・📋 . There is #1238621511401734254 to submit the bugs you find or you can DM @real wyvern .
please also be reminded of our first rule to treat everyone with kindness and respect no matter who that person/ team are
support and bugs feedback are useless
U can check and see how many ppl have the same issue and no answer
I understand your frustration. the team actively monitors the #1238621511401734254 or you can DM @real wyvern to speak to the team.
however response times may be slower
i was gonna say something shady regarding about this but ill stay quiet for now
good mornig people of america
Thank you very much mate means a lot and helps a lot
At a glance, that's a good exaple of the sort of tutorial I mentioned above. The important part is to go through it, put it to use, practice, take notes and keep learning.
In all the graphics apps I work with,
we can ADD files to your Work space by simply Dragging them into the work window.
Why isn't this possible in Google Flow?
Furthermore, not only can't you drag files into the work window,
but you can't even upload more than One at a time.
......... Furthermore, the way it's set up, it takes at least 15 seconds to upload a single file, and there's no way to speed it up.
How can this be explained ? what to do ? Did i missed something ? ....
Same with downloading - only One by One 🥴
The taking note in question
The drag n drop thingy? I think they’re working on it 👀
Do you guys have any solution to voice difference in clips? mainly in other languages than english, also lipsync?
I hear voice consistency can work if the firsf clip has that vooce on the last second of the clip and you use extend. That being said, I would use other audio. I cut the audi gen audio out and add my own, and use other tools for lip sync.
heys guys
is flow down?
all the images and videos are just not loading for me
been like this since yesterday
working for me, often says error tho..
error " Something went wrong."
Goes up and down, depending on peak use hours. Someone worked it out a few days ago. Probably overlap of hours where North America is coming online as Europe is still up, then later when we're still active and asia is coming online.
That's not a thing.
your reply is moot to my ask
You can't add people to a Pro or Ultra plan. That's not a thing.
Actually
I think the failed generations in VEO 3.1 are because of the upscaling for some reason, after upscaling some videos I can't generate anything else and have to refresh the browser and then it generates anything else, this is what I've noticed.
Has anyone else had a weird problem where when making edited photos in Nano banana pro (like giving it 3 images and telling it to make a pic of the 3 people together) it makes an upside down photograph?? I cannot get it to make a pic that's right side up, it keeps making upside down images lol
i guess just rotate the image manually hahaha
how to join disco
I wanted to post an issue about disappearing videos to the #1319424187730755636 but I don't have permissions to do so.
So, all my videos have mysteriously disappeared. They are still available in the scene and some of them available in the Library, some of them appear as blank boxes. I generated new video, and all removed videos retured back.
Another issue - I saved many clips, in 1080 resolution, all 8 sec. But, their sizes appeared different, from 8mb to 24mb. Is it okey?
Thats a new one for me!
@nova knot Hello, as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now. Thank you!
Hi @neon bramble,
Flow's bug reporting was moved to this channel: https://discord.com/channels/1172568727942860810/1238621511401734254. You can now use tags to filter for the specific product you're discussing. As for the vanishing videos, other creators are tracking that same issue in this thread: https://discord.com/channels/1172568727942860810/1466934039876337806.
@blissful cobalt Hello, as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now. Thank you!
is it down?
Thanks for reporting this issue. We are investigating the issue where content fails to render in the UI sometimes.
Regarding the generated file sizes, the numbers you mentioned are plausible. I just tried a few downloads and landed on sizes in the range of 15 to 17MB, It is content dependent.
That's for family members, not randoms
nano banana generation always failed, I'm ultra plan user. google flow sucks.
I have a problem if I send a reference in flow ai nano banana pro from Russian model like a Dasha taran he send to me my policies don't make a photos from famous people
"Russian model like a Dasha taran" = "policies don't make a photos from famous people"
But I am want to make a selife with her in park with not any anything vs policys
Is there a big difference in actual video generation speed between Veo 3.1 – Fast [Lower Priority] and Veo 3.1 – Fast?

"don't make a photos from famous people""
Ok but if want to make a photo with my favorite singer what I doing
I want to be king of the world for a day, but alas... we do not get to make rules for anyone other than ourselves.
You have to invent a new character.
That or use image editors that are run locally
If I get a reference photo from my Google photos
Like Photoshop
I dont have a laptop only phone 🙂
There's image editing apps. If you're on iOS then there's procreate
You already know the rules forbit making images of famous people. What you want doesn't change the rules.
I want the lottery to only ever make me the winner, and everyone has to play even though they know only I'll win.
Ok what a useful from Ai if don't make any photo I want in our world we have a respectful people make this for himself only don't to any one and people have and people have this things for destroy a fame of people I from first type
I've made plenty from the Google family of AI products. It just pays better to be creative and also follow the rules.
Ok what a useful of image Editing if you don't make you have a selfie with any famous people in nano banana pro
For fun I you want to make a photo with you with game of thrones actors for example
I want this coffee mug to shoot rockets at my annoying neighbor. What a useless coffee mug...
Stop. I'm going to pass out from cackling.

Do you understand my idea if genie 3 is realased for example and one of people make a world of his favourite cartoon like a toy story and he want to walk in andy room he doesn't do this because the company of Ai don't allow to this and for all people want to make a game
For them famuos cartoon
What a useful of this
So what? You wanting something doesn't make it legal. Making a game using the intellectual property of others isn't legal for Google to do. If you want to do it, then you have to pay the owner of that world to license it.
No for his dream google is make a Ai for make a people dreams to real thing like a video or world
Like a YouTube it make a people have a money like a Mrbeast and can you watch a old videos from anything
it's normal ?
All of my prompt are not working, "failed generation" did you notice? Or only mine are not working
The occurrence of generation failures has become increasingly frequent recently
widespread
video gets more blurred after upscaling, anybody encountered this?
This is a huge problem since the realease of 3.1 but apparently nobody gives a fun about it
Just keep paying 250$ for a unusable service
This mont I could not use all my credits because of failed generations, apparently this is normal for google
Also it's normal to completly ignore all the complaints
SUPPORT IS USELESS
WHAT IS THE POINT TO GIVE FREE CREDITS IF I CANNOT USE MY CREDITS THAT I PAID FOR??
Gemini isn't working for me; it keeps giving errors. I can't generate videos, images, or text.
mb u should spend another 250$ on another account
?
are u sure
It's a joke mate
They dont care just keep paying, I literally lost 1/3 of the credits from previous month because of failed generations, flow is UNUSABLE
error after error, some times it generates one of four videos
always gives error any support ?
No. It's a widespread issue.
Is flow working or down??
I cant create anything
It saying prompt is violating their policy when its not
I guess its high time to switch to some other model
This became every day problem now days
It's not a very good filter. To troubleshoot, you can reword different parts of the prompt to identify which part is catching it up.
what is this ?
This error started today. I can't use something I've been using for 3 months.
My prompt was “create a text image written “hello”
What should i troubleshoot in this?
@teal tapir i think he is trolling us
Every time there's a big new release, there's a commensurate big spike in how often generations fail.
Well nevermind, then.
unbelievable they started to give free daily credits when they cant even handle current demand
all generations are failing past few days/weeks... why even paying for this
I have the ULTRA package and I can't send screenshots or chat; it keeps giving me error number 13.
@sudden coyoteHello, as a friendly reminder we cannot generate images or videos in the server! Please do not post your prompts in the channels as it will not generate. If asking for advice or tips on your prompts do so in the appropriate channels.
Terrible quality
Hi @hollow slate videos aren't created in the discord server. In order to make a video you'll need to go to https://labs.google/ pick the tool you want to work with then go from there. If you need help with prompts please ask in the appropriate channels such as #flow-tips-n-tricks・✏️ for advice. Thank you!
Most people are scamming customers by selling and booking slots for Veo 4 even though there is no such news about Veo 4 and this kind of scam is spreading in asian countries 🥴
@edgy holly needs your opinion 🤔
Is this why so many people keep coming in here asking about Veo 4? There's no news on that at all. It's all pure speculation, and Google is unresponsive/secretive, so there is an information vacuum that scammers can exploit.
Yeah that's right it gives some kind of space to rob people 😕
with kling 3.0 releasing i think veo 4 will be launching quite fast
they did the same with google genie when lingbot world was launched
Not yet.
its already available on higgsfield
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. please use #1238621511401734254 to report your bugs! these threads are actively monitored by the team!
Is image generation down in Flow currently? Getting this constantly but video generation seems fine.'
Anyone else disappointed by the plastic game-like look of Kling 3?
I always ignore the first week or so of what we see when models first come out. Too many are cherry picked promotional content, much of it paid for without saying so. After initial promotional spam and buzz dies down, regular people who know how to use the tools well give it an honest go.
Unless what they share from their official main channels is really bad, in which case the model is even worse, since they cherry pick the best to share there.
Just started sharing there recently.
To me. So far, Veo is still the absolute king, when used with care.
Runway is finally starting to catch up, and xAI is improving, but Veo and Kling are usually the best, give or take depending on which has updated most recently.
what is going on with ultra my credits reset to 30 out of nowhere with weeks left on my sub
3.5 coming soon.
It’s actually not bad at all when giving it a reference image
Runway isn’t going anywhere their best model in the world sucks
Veo still wins in realism I2V tho
I canceled my Ultra plan today - too many other models now greatly surpass VEO 3.1 for so much less money, it's just not worth the $250 a month. Will re-evaluate when VEO 4 gets released, but if it's not absolutely amazing, I'm done with VEO.
I can’t create videos right now. Is anyone else having this issue?
🙁
It’s still native 720p tho
Their only using an upscaler
When using image to video details start very blurry
Hi, given that veo 3.1 was a downgrade from veo 3 and veo 3 was in some ways a downgrade from veo 2, and kling and even grok imagine and sora 2 have all wildly surpassed veo by now, is there an update imminent because I'm not sure what i'm paying 250 bucks a month for....
its so much better than veo its not even funny at this point
this is veos response to kling 3 release lol
is Kling 3 better than veo 3.1?
I can't find Kling 3 in arena.ai's leaderboard
this is really good looking at it from my computer but still mouth movement lol
Uhhh lol yes
Google need to respond fast
I don't know which one should I rely on
Both is text to video
Grok imagine sucks why do people even like it? lmao
its way, way, way, way better than veo right now
wait whats Vidu Q3 Pro?
Not better than Kling tho
no
altho testing both at the moment and for some things grok is giving better results
one good thing about grok is there is no garbling of details on fast motions, that scrambled mangled look you get in a lot of AI video tools, grok doesnt do that at all (talking about 720p gens via API, they are very high quality)
I tried grok i wasnt impressed tbh
well i guess it depends what you tried it for
Maybe i need to try it again but damn
Chinese video generation services are already surpassing Veo 3
who was it that decided that veo 3.1 t2v should have a super high contrast oversaturated oversharpened look baked in no matter what you ask it to do? Like were they listening to one guy that really, really likes the look of bleach bypass and convinced them that everything should have a bleach bypass look?
grok one is low res
Hi guys, how are yall doing long AI videos, like 8-10 minutes for example finance animation?
pls i need help
752x416, it can do 1280x720
but even the low res looks way better than the gargled incoherent nonsense veo spits out
also again the veo output is over contrasty, over sharp, over saturated. Why?????
who decided that this is how everything should look?
8-second animated video, Pixar-style 3D animation, high-quality cinematic lighting, vertical 9:16. An angry anthropomorphic green mucus character sitting on a human finger. The mucus has big expressive eyes, furrowed eyebrows, annoyed and impatient facial expression. A human mouth slowly approaches and starts eating the mucus, exaggerated cartoon action, funny and uncomfortable but NOT disgusting, family-friendly comedy. The mucus shakes, resists slightly, looks directly at the camera while angrily complaining. Audio dialogue (Spanish, fast, scolding, comedic tone): ““Etiqueta a tu amigo que se come los mocos y sígueme… o tú te comes los mocos.”” The action lasts the full 8 seconds, slow comedic timing, exaggerated Pixar-style motion. Bright colors, soft global illumination, shallow depth of field, expressive animation. No subtitles, no on-screen text, no logos.
ok nobody gave me a good argument to continue paying 250 bucks a month so i downgraded to the 24.99 a month plan. if veo 4 comes out and is good I'll reconsider
veo3.1 is not even mid
very funny, always error 403 after generating 10-12 video/image. gonna ask for refund
That ranking is super interesting: https://arena.ai/leaderboard/image-to-video
The one I sent is from https://arena.ai/leaderboard (29 Jan)
Oh wait you sent the image to video one
Yes. There's image to video or text to video: https://arena.ai/leaderboard/text-to-video
Isn't everyone resting so late?
Silicon Valley remains the center and frontier leader of AI
is there a bug on google veo 3 at the moment?
hi ,
so I cant generate images and videos here ? like midJourney?
What was the prompt?
flow is completely broken. error party all the time. senseless errors. not downloading pics or videos. need to refresh the page so errors doesnt appear. after every generation page needs to be refreshed. so laggy and just terrible tbh
i pay 250€ for this?
Maybe you can open a new project than start from scratch
did several times
Let's have a checkup:
- Have you mentioned shot type ?
- Have you set Background
- Have you imposed Boundaries, Specifications and Frame or Image ?
?
?
Let's have a checkup:
- Have you mentioned shot type ?
- Have you set Background
- Have you imposed Boundaries, Specifications and Frame or Image ?
what r u talking about man? wrong guy
Then try to restart the PC
Uh, You should check the prompts too
you should stop giving advices. feels like troubleshooting of microsoft. never helps.
Brother
Okay. I've also works to do. But Flow is nice for me. Atleast better than Sora and Firefly
perfect
Hey @crystal talon What is the prompt [inputs]? I can test it for you.
it has nothing to do with prompts. they work. but i have to reload the page after every single genration, otherwise i get unknown error. this goes since a week.
despite everyone saying every other model is better then veo, thats probably true when make videos in english, nothing else but veo supports my language, so thank u google
am i right that a "pro account" doesnt really make sense nowadays? at least not in terms of the usage of veo? a pro user gets 1000 coins a month. a free user gets 30x 50 = 1500 coins. 😄
this is you rn
isnt it 50 a day x 30 days?
haved u tried asking gemini
free credits doesnt roll over tho If you don't use your 50 credits today, you don't get 100 tomorrow. You have to be consistently active every single day to hit that 1,500 total.
i never said they roll over. but i realized a flaw in the calculation: you can only use 2x20 a day and 10 get lost.
im getting many generations that are very low fps. anyone else experiencing this issue?
yes. the math looks good on paper but the daily restrictions make it a pain to actually use. pro is just less headache
Actually I have no idea what I said just now
Not really , i only get low fps when i prompted it
Strange.. It used to never happen but now it happens 1 out of every 4. I generate 20 videos at once usually. 5 sets of 4..could the mass generating be causing issues?
Can I see the video that u generated that has low frame rate per second
And the prompt just in case
Or it could be flow pages lags when there’s too many vid on same project page
Cuz I got it too.
good point. ill try downloading and accessing
In Flow? 24fps is automatic, I thought. Can you check to see if it's actually an FPS issue, or is it using some kind of stutter effect?
^ + Always share the prompt and inputs. Otherwise its usually too vague to help.
You can describe it as a stutter effect yes
i just love paying $250 for something with poor performance! 🙂 3/5 generations take forever to load then it says error and entire generation prompt disappears so i gotta type it again... PERFECT! but please keep people use it for free to take more performance out from paying people.. cmon google u can do better.. i removed every single project... its so annoying
yup same frustration here, ultra subscriber for 6 months now. Flow still provides good value, but it's nothing compared the other video models out there now. I really hope the next update will smash (again) all comptetion
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect.
please use #1238621511401734254 to report your bugs
i would not say its bug anymore, but feature. its not first day discovery of this..
I cant use flow
Hey guys, what do you recommend for starting to monetize with Flow on TikTok or some other social media platform?
file managing simply doesnt exist. terrible UI guys. When will they update this? I am looking for an old prompt of mine and jesus christ, i need to scroll down forever.
is there a way to generate videos less than 8 seconds?
What I'm saying is download the file and look at the properties of the file to see how many frames per second it has.
Flow started to refuse to create video all of a sudden. Videos of me talking in front of monuments. It did so happily last week. Now I get a very vague "Failed Generation". Am I hitting some policy violation or something else is off?
Uh ok just searching I'm not the first
Just generate a video and then either shorten it in scenebuilder or download and trim elsewhere
Try the „fast - low priority“ if failure rates are high this should work more often
Sure but no way to output anything other than 8 seconds without editing?
Probably overload. If you violate a rule, you'll get a different alert.
If you pay for each generation individually through the API (in other apps), you can make clips as short as 4 seconds, but it's a static 8 in Flow.
Good to know thanks!
I don’t think they’ll be doing 3.2 maybe 3.5 or 4
my account on free plan, but can't download upscale 1080p. why this happen? anyone can tell me bout this probs?
Is anyone facing problems generating on FLOW
what is the flow?
I am new to flow any tip o video for start ?
@wild mason
How to get free flow buddies
Hello, we wanted to remind you of our server guidelines, stated in #start-here・📍 : “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now.” Please keep in mind that non-English discussions and content may be removed. Thank you!
Hello, as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now. Thank you!
anybody else having all generations failing?
why is flow too stupid to generate punches or kicks? and most of the time it doesnt listen to camera movements prompts. i say move the camera or orbit the camera but instead the actors turns, so annoying.
yes, you have to reload the page after each generation.
flow feels so abondoned. the ui and the background tech of this site is just so behind compared to others. i mean waaaaaaay behind.
Has anyone used the "Character controls" feature in Flow (with video input)? Described on the deepmind website as "Bring characters to life, using your body, face and voice to animate them." but can't figure out how to do it
clear cache and cookies. fixes it for me for a while and then rinse and repeat
does anyone know how to change camera position in flow generating images? say like move camera across the room to another corner, capturing same chair for example
prompts like change, move camera angle, dolly, rack, arc, doesnt work for me.
Why is it that in some prompts, it allows me to make what I want, but when I add a reference, it doesn't and says it's against the rules, right after it made the prompt with no reference?
Cinematic historical battle shot, ultra-realistic live-action look.
SHOT 1 (3–4s) — Tracking wide:
A mounted archer gallops fast across an open field. The camera tracks alongside the rider at a steady speed (smooth gimbal), keeping the archer centered. Dust kicks up under the horse hooves, cape and cloth flutter with believable physics. In the far distance, a tight Roman infantry formation holds shields in a disciplined line.
SHOT 2 (2–3s) — Medium close tracking:
The archer rises in the saddle, draws the bow in one clean motion. Muscles tense, bow bends naturally. Brief slow-motion at the moment of release for clarity (very subtle).
SHOT 3 (2–3s) — Arrow flight + impact:
The arrow launches and slices through the air with a sharp wind streak, maintaining a crisp silhouette (no warping). It flies toward the Roman formation and slams into a shield with a solid thud; the shield jolts, the soldier staggers half a step, nearby soldiers flinch but formation holds.
Lighting: natural daylight, cinematic filmic grade, soft highlight roll-off, neutral saturation.
Camera: stable tracking, no fast whip pans, no jitter.
Negative: no oversharpening, no oversaturation, no “crispy” HDR look, no deformed limbs, no melted faces, no morphing weapons, no scrambled motion.
Yesterday, the Chinese rolled out an updated KLING 3.0. Will there be any progress in VEO 3.1 or will we be stuck with the same inconvenient interface?
please update veo so that it can not play music unless you specifically ask for it.
Someone at Google said this is the month of generators so never know
Hello, as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now." Thank you!
Hi guys
Anyone faced this error often? Its said "Something error", and i checked the network request, it got all 403 Forbidden?
hello!
"/generate "un perro con una corona"".
'd be very surprised if Whisk's model gets updated to NB Pro. Its just too expensive to allow almost unlimited free generations. I have a fondness for both the original NB and NB Pro, and I take on board what's been said about text rendering and consistency as well as censorship.
There has been a recent, very strict clampdown on a lot of things on Google's image models lately, and in all honesty, it has taken a lot of the fun out of it, I can't deny.
There is a lot of selfish entitlement going around.
New to using Flow. I have been using the image generation to build character reference sheets (culled the prompt from a YouTube videO0 and while it works really well (for the most part) it doesn't always fllow the prompt instractions presisley. It seems to be hit or miss. Does Flow allow/provide feedback to help improve the tool when it glitches like that?
Have yall seen seedance 2.0?
Does anybody knows why Veo 3.1 Quality gets stuck at 99%
@ornate thorn
RAO
When flow update voice ?
I am very confused. Every time I create an image, the second time it appears, this page will definitely be displayed
first kling 3.0 then seedance 2.0... its time for veo 4 to join the party
Guys we need image uploads in flow (press the chrome dino if you want this too)
salam
is this server just for tech?
or does it have science
and biology
????
\
imean as i saw everything is about coding and IT and those stuff.... is it all like this??
We already do.
Kling is pushing now . When will Veo 4 come. Veo3.1 really sucks 😑 right now.
check seedance 2.0, veo needs upgrade really
I just realized, that Nano Banana image upscale to 2x, is counted as Image generated.
I was downloading my old images today via upscaling them to 2x, after a while it started giving me a message, "you have reached the limit of nano banana pro for today"
I'm on Ultra plan, is it just me, or others have experienced this too ?
I spent 100 credits just because some random character appears in the background..
Anyone have some tips on creating a character sheet that works to keep a consistent character and not make the character look like plastic?
Hi there, this server is dedicated to the Google Labs products and all that encompasses. You can check them out here. https://labs.google
Hello, as a friendly reminder we cannot generate images or videos in the server! Please do not post your prompts in the channels as it will not generate. If asking for advice or tips on your prompts do so in the appropriate channels.
Any one interested in a upwork account?
I am a paid PRO user. My generation in Flow has been stuck and the button is grayed out for 4 days. Please reset my session
That's more of a tech support thing, I think.
hello, does anyone has any tips/prompts for nano banana pro for pictures being less AI and more natural ?
Hey folks I’m stuck with Veo 3.1 object identity drift and could use expert help.
I’m generating a short “how-to” clip featuring a small wooden crescent/half-moon handled brush with a single dense bristle band + a row of gray silicone nubs.
Issue: as soon as animation starts (especially when a hand holds it or it touches skin), Veo morphs the brush: bristles split/splay like hair/fur, duplicate rows appear, or the brush turns into a different shape entirely. The wooden handle sometimes stays closer (but even that warps terribly into something else), but the bristle head warps badly frame-to-frame.
Tried:
- reference image / “ingredient” image
- super explicit geometry description (“rigid hard object”, “single continuous bristle band”, “not hair/fur”)
- strong negatives (no warping, no splitting, no physics, no duplication, no flicker)
- slower motion / simpler camera moves
Looking for: prompt patterns or workflows that reliably lock product geometry (e.g., turntable-first, first/last frame anchoring, compositing workflow, camera constraints, settings).
If you’re down, I can share the reference image + a warped frame + the exact prompt/negatives in the DMs
how can i download png format
Hey - important question for nano banana pro - when i'm either creating in gemini or in flow i used to upload product pictures of clients and create some cooler pictures out of it. Now, it always says that there is some privacy / data protection issue - i always get this response: I can help with editing images of people, but I can't edit some public figures. Is there anyone else you'd like to try? - and it deostn generate, however, its just a product of our client. Used to work everytime
Need to fix asap
Be specific as to what exactly you mean by "more natural". Find different ways to say it. Take notes on what works.
Hello, somebody have good presets I could use?
Dear developers, please add voice control features. Every frame has a different voice, it's so annoying. Please, or give instructions on how to get a strictly static voice.😭 😭 😭 😭 😭 😭 😭 😭 😭 😭 😭 😭 😭 😭 😭 👏
Hello, good day. Has anyone reached the daily image generation limit with Nano Banana Pro? Could you tell me how many images you can generate per day before reaching the daily limit?
I have the Plus subscription.
We currently only support JPEG. Do you feel like you are losing quality by using jpeg?
Is the issue specifically in Flow. If so, can you attach the message that you are seeing?
Yeah we hear you (no pun intended). Our team is looking into voice controls. I can't promise any dates, but we realize that this is a necessary feature for consistency.
Hi @desert matrixHappy to help!
Could you DM me the inputs so I can take a look? Thanks.
Hello do you know which is the limit generation for nano banana pro in ultra users?
Goes up and down with server load, but I think the default is 100/day.
Thank you, is for pro o for ultra users?
Yeah. Ultra gets 1,000/day I think.
How? Hes not aaking technical questions. Hes explaining why the tool isnt working as described on a fkundemtal level
Because it's not a common issue and he's asking for something that requires account access, which mods here don't have.
So jsut say, hey you need to get the required paid account
Its a general question and common
I have employees who are less tech savvy, non software e engineers or devs thag struggled with understanding the region lock logics and the lack of real info explaining why some of tools arent accessible unless you dig deep
The fact that half the lab tools are so ambiguous is an issue and the separation of tools like flow and spark etc, is jsut to get users to increase usage
The issue he described shouldn't be with a pro account.
Two quick questions! Is there incresed Nano Banana Pro usage in Flow with Ultra vs. Pro?
And, it seems that start and end frame in Frames does not honor the end frame well (vs other tools)...is there something I'm missing?
For your first question, yes Ultra has higher limits in general.
And yes, current end frames are not 100% preserved. The model team is aware of this and hopefully should be resolved in a future release.
Thanks so much! Any quick tip on getting a front-facing selfie (like one the character would take)?
"Studio Ghibli style, magical forest, flying castle, spirited away vibe, soft pastel colors, anime, hayao miyazaki"
Is it just me or is video creation broken? Every time I try, it fails on audio generation.
Is all generations using a reference image flagged currently? I'm trying to make some concept images of myself and nothing will send through. Was working decent this morning and now... no results.
Anyone else today?
It’s exactly the same for me.
It was going fine this morning 😭 sucks cause i needed to get some stuff done and simply no other image gen competes with NBP
This is absolute garbage! We spend hundreds of dollars every month, and it's constantly making mistakes.
I made a mistake investing money in this artificial intelligence research. Things are getting worse; image creation errors occur constantly, and image downloads are very poor. I feel stupid and disappointed. 😡
@violet heathHello, as a friendly reminder we cannot generate images or videos in the server! Please do not post your prompts in the channels as it will not generate. If asking for advice or tips on your prompts do so in the appropriate channels.
Anyone making ads with AI?
For ads, in my opinion, the best Kling 3.0
Hi everyone,
I'm brand new here and I don't know anything about AI video generation, but I have a project I'd like to do because I want to ask my girlfriend to marry me. We're both big anime fans, so I'd like to create an anime with my girlfriend as the main character (at the end of which I'll ask her to marry me). I don't have the story yet, but I'm looking into the feasibility of my project. I've seen some amazing clips from @dark bluff , for example, which reinforce my belief that it's doable.
But as I'm a complete beginner and would need to devote around 700 to 1,000 hours to this project (over 18 months) (all the AI work + editing in Premiere with sound design, etc.), I'm wondering if it's really feasible, given that I have to learn everything from scratch.
I would like the anime to be at least 10 minutes long, but if I can make it longer, I would like to go up to 20 minutes.
I really don't realize how difficult the AI part is, and what is feasible or not, hence my message.
Thank you in advance for your help!
Definitely doable. Veo hasn't even been out 18 months. Just start learning now and keep at it.
Key to a project like this will be the input images. To have consistent characters you need to have good reference images, so start learning how to use Nano Banana Pro for ingredients, character sheets and reference images for keyframes along with learning how to use Veo in Flow.
Thanks a lot for your answer ! Yes I saw that consistency was the key in this project, will work on that haha thanks !
How me canHow can I stabilize a character's voice tone when creating a video in Flow?
When you run into issues, go into the #flow-tips-n-tricks・✏️ room, and share the specifics (what you want, the generation, the prompt and any input images) so people can help. A lot of questions people ask in here are too vague to give any real help.
Is there a cpp program in veo?
It's noted will do that thanks a lot !
how to use for free unlimited?
Can you describe a bit more about what you are trying to do and what you are seeing? Is this using Nano Banana or Ingredients to Video?
😄
China’s Kling and Seedream make amazing videos
I can’t believe how fast they’re updating!
Essentially using my likeness as a reference image hasn't seemed to be working in NBP
Interesting. Feel free to DM me and I can help look in the logs to see what the issue is.
For Ads, we can use AI Video Marketing
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. we have #1238621511401734254 for feedback specific questions/ queries
OK its my fault 😭
No way 😱
Seedance 2.0*
I keep getting "You cannot generate a response to this prompt due to Google's guardrails related to third-party content" on all my generations for a prompt. its a simple cartoon animation that is 100% original and ive tweaked the prompt wording 5 times
basically PNG files allow transparent backgrounds which can be a real time saver and yeah png give better image quality
why thumbnails not showing up ?
Cool idea, but please ask her in real life face to face in oldschool manner. Chicks dig that.
Hey guys, I haven’t used Flow for quite a while. Is Nano Banna working normally now? I mean, is the “something went wrong” error fixed already?
Visual Prompt: “Old hotel room door creaking open, empty room, dusty furniture, flickering lamp, realistic horror cinematic style.”
Text Overlay: “Room 7… they said no one stays here anymore.”
Audio: Door creaking, wind blowing through broken window.
How too use this app
Also i can prove that im a student
Obviously, my initial plan would be to rent a movie theater, invite family and friends, show the episode, and then ask for her hand in marriage at the end with a little speech
Ok sounds better
Happy to help if you have questions
Thanks a lot ! 😄
@dreamy gyroAs stated in #start-here・📍; currently, our team can support engagement in English; the discussion in this server will largely remain in English for now. Additionally, there's no way to generate images or videos on this server.
Sure thing man, sending it your way rn also thanks for reaching out and choosing to help
seedance 2 is taking over twitter, veo 4 got to put a stop to this
That sounds like a scam. Only buy directly from Google. I don't know if that's why your account was banned, or if it's because you broke any rules, but either way if you only want to pay for Veo, the only way to do that would be using it through the API, like with Freepik, or any of the big third party apps.
I would pay to google but they are too expensivee.
What do you suggest me, Freepik or higgsfield?
I'd look around and try a few of them, and decide for yourself. Higgsfield is catching some heat right now for some... bad stuff, so watch out. Freepik usually is one of the cheapest per generation.
Thank you very much Sir!
I’ll try everything
YEah, shop around. Pick a few to try. Go with what seems best and stay flexible. Don't be 'brand loyal'. Stay open to using other tools.
Also, @inland pier Is there still a lot of server-side issues? I'm noticing a lot of "Something went wrong" and failures well after a 24 hour reset
if i can suggest something, Kling directly from their website, not by third party vendors.
Why?
I dont need it for “movies”
I need consistent character for my content.
Talking head videos (70% of the whole content)
For that Veo was doing really good job.
Is kling too expnsive with credits per video?
Feel free to DM me your prompt and I can look into it
Feel free to DM me and we can try to troubleshoot
Depends on the account level, but my biggest issue with them is it's so SLOW. At the moment their latest models are generally better than Veo (each have their strengths), but 3.0 gens can sometimes take a half hour to generate a clip.
Testing one right now (720p, 15s, image to video) and it's been... 45 minutes and still hasn't processed. (2.5/2.6 aren't nearly as bad, but still much slower than Veo).
Just finished as I was typing. Ended up taking.... 45:55
A 15 sec clip is on the longer end of a single segment, they got that going for them at least
Didn't hold onto the style very well, but none of them do painting styles well. Followed instructions, though.
This morning in flow trying to generate images I'm getting "Couldn't generate image. Try again later." Prompt doesn't matter. Anyone know why?
Prompt adherence for me would be the most important, so kudos there.
On that I find Kling to be best. Veo is pretty bad at camera control and directions (left/right, though up/down seems to work well usually).
But for less money, I could have used Veo fast to generate a few dozen clips in that 45 minutes and using Extend or split into two clips with a cut gotten the same or better result. That's just too slow to be viable for anything I'm working on.
Anything over probably 6-7 minutes is just too long for anything specific (which, for me, is most things). I almost always discover something better after I get into a scene, so iteration is really important for my process.
Same for me
Same
veo keeps saying "Something went wrong."
anyone else struggling with this error?
same, my ultra flow it's broken. they show "something went wrong." everytime i generate picture
yep
same
damn...perfect timing
Yeah image generation is down for me, not generating anything
Right when I started working, went rip
Doesn't work
well...there goes my clients deadline...again
I've already post the issue to support chanel. Hopefully, they'll see it and fix it.
They probably won’t fix….
If anyone knows how to do this, set up a bot that tracks when people say they're having these issues, here and on X, and we'd be able to work out when their servers are overloading. It comes in waves.
Where are you testing kling 3.0?
Leonardo. I was in a live audio chat space this morning where we talked about it and generation times from 15-30 minutes (on all the platforms, Kling and API) are common.
In your opinion…
Which model you think is the best for classic talking head videos
Something like this..
Something that simple is within reach of a bunch of them now. More a matter of different than best. Veo tends to over-act. HeyGen was built for this kind of thing. I know Hedra and ElevenLabs does that stuff well.
Is there any news about google new video model, if something like nanak banana came, it i’ll be crazy
No, and they almost never say when major updates are coming until like a day or two before they start rolling out. Pay no mind (even better: block/mute) anyone who shares that kind of clickbait stuff.
May i know which tool are you using rn.
And for what purposes?
I would love to know
A big determinate for the best tool for you specifically would be your budget, and how many of them you want to make per month. If you're really churning out tons of them, then Google Ultra is probably your best deal, because you can make tons of them in Fast mode either for 10 credits each, or zero. If you care about the voice stuff, you can replace it with various lip sync models out there.
Mostly Veo for video, an old voice changer for dialogue, Kling lipsync for that, and Suno for music. Dabble in other stuff here and there.
I tried every lip sync model and it suck really. I also tried ltx audio to video, on my opinion it is bad…
Top paid models are almost always better than open source ones.
I don't do any kind of fast action or wild camera stuff, but I find lip sync to not be very hard. The more the character is moving when you do it, or if multiple people are talking, the harder they have from my limited experience.
Seedance 2.0 might be the first exception, though it apparently is very expensive. Open source, but waaaay too big for anyone to run on a normal PC.
I also tried kling lipsync on higgsfield and it is pretty bad
That's where I've used it. Works well for me.
May i know what type of content you create
I'm working on a TV show. Noir-inspired, set in the near future.
NBP is on
did nano banan get more censored out of nowhere? like i tried to generate a photo of Lois griffin but i got the third party flag
Every update gets better at that. Especially with Disney-owned IP after they made a deal with OpenAI.
nano banana still down friends?
Can anyone tell me where I can find strong prompts for veo3 in this server?
Is anyone also encountering export failed on Scenebuilder?
Why do you think seedance is very expensive.
Model are not out yet
Why is flow not getting any updates?
File management terrible, errors non stop without telling what kind of error. Nsfw errors with no reason. It used to be the best site only 3 months ago, now it is the worst.
Can anyone help me troubleshoot why I'm not able to download my video from scenebuilder? I keep getting a "export failed" error
You right-click directly on the video, then choose the option to save a video from your browser. It might not download in high quality, but it's better than nothing.
Woot, that did the trick! Thanks Nats!
yes I shared a tip I just found at the top
please do you make UGC content?
any updates?
How did you do for keep a realistic face ? I tried podcast on flow with realistic pictures but, veo keep plastic faces...
@tiny canyon
Some people in China have it, and some people outside of China have early access.
Everyone in chian has it lol
China
just need a chinese number
Bro you make in labs google, not in here
Veo 3.1 is looking like a old model now. Ai is rapidly improving meanwhile veo 3.1. when will Veo 4 come out.
Is flow Nano Banana Pro Unlimited now ?
noo
Out of curiosity, how long is the video that you are trying to export?
I think the problem is with your image reference.
Use nao banana pro, not nano banana classic.
I dont see the problem with consistent characther rn
can anyone tell me about flow update?
I assume flow is Vortex Studio etc, the ecosystem for all the AI that came with the free trial credits?
Hi, is there any Google official MCP server for Flow? They don't have it listed inside antigravity.
would be nice if flow let us generate images in 1:1 AR too
Not realistic ? can i send you complete video in dm?
Then what is the problem with consistency
I sent you a DM. I keep the exact same character for the first and end frames; it's just that for the end frame, I always ask Nano Banana Pro to slightly modify their facial expression, and I always pay attention to the quality of their skin.
Why doesnt flow have 21:9 format? Are they antifilm?
@tiny canyon
Flow is the worst page to use veo.
Because google abandoned flow
What do you use?
Which tools, platforms etc
Openart, higgsfield, pollo ai. You can apply for CPP if you have anything to showcase, if they like it you will get monthly free credits
Is something going on with the scene builder? if I do any adjustments to the length of one of the clips I can't play it at all now, it stops the whole project.
My male character just stared doing pretty much only an Australian accent today, screwing up all my takes. Even if I tell it no to. Also getting freaking subtitles again, as well as random mumbling from the girl that is not supposed to be talking. Uggg....
Wow
We don't have an official MCP server for Flow. What are you looking to do specifically?
This isn't out of the question. Let me bring this idea back to others on the team and get their thoughts.
Hi @inland pier when does next veo model come out?
I can't share any dates unfortunately. Stay tuned!
To make sure I understand: you are trimming one of your clips in scene builder, and then it won't play back? what happens if you refresh the page?
You understood correctly, I tried refreshing a few times and it does the same. I am going to try and clear the cache and see if that helps.
OK thanks. if you want to DM me a short screen recording, I am happy to share back with the team.
building a personal pipeline that connect flow,stitch+3rd party service together to create beautiful 3d websites. I need to programatically create videos inside antigravity, i could create the mcp myself but i am not sure if that would break TOS.
Ahh that workflow makes sense. Yeah at the moment we discourage using our system programatically. Your best option is to use the Gemini API, where you can use an API key to programmatically call Veo.
Hopefully it will be available inside antigravity soon, just like stitch is.
How does Flow fit into building 3D websites?
Regardless, Veo is available in Vertex/API.
unfortunately I couldn't get it to play again so the video was just a screenshot lol... my apologies I tried to get it to do it again. I haven't cleared the cache yet but I will do that and hopefully that will fix the issue. Been hanging with my 3 year old.
it does not, but I can use it to create 'videos' of 3d models, then I animate it programmatically in the browser in a way that feels like its a real 3D.
guys i havent yet tried it.... how do i start with flow?
i want to do an assignment
You don't need Flow for that. Just use Veo via the API.
Hello everyone
hello
Hello guys
@inland pier
Hey Ronny.
I'm interesed if i can have some discount as a student on google AI ULTRA package?
Right now it is $150, which is pretty expensive.
Are any discount avaiable
Is anyone able to download the 1080 upscaled version in Flow? I'm getting issues.
This is urgent. Can somebody please help me
I need to download the video in 1080 resolution
Does anyone have information about character stability and transitions between different scenes?
Wait couple of hours and try again, it happens from time to time.
@solemn crown Hello just a friendly reminder that videos and images aren't generated in the discord server. In order to do so you'll need to go to https://labs.google/ pick the tool you want to work with then go from there. If you need help with prompts please ask in the appropriate channels for advice. Thank you!
Hi @silent hedge to report any product issues please use our #1238621511401734254 channel. When creating your thread please post as much as possible so that the team can properly investigate.
Thank you ☺️
Why is nbp so stupid and thinks everything is nsfw?
Is there an outage today? all my videos are failing and saying the reason is the audio.
I really hope nbp will be killed by some chinese ai. Who runs that place? The most sensitive karens on earth? Everything is nsfw now.
Hi @true sky just a friendly reminder that videos and images aren't generated in the discord server. In order to do so you'll need to go to https://labs.google/ pick the tool you want to work with then go from there. If you need help with prompts please ask in the appropriate channels for advice. Thank you!
Unfortunately we don't offer student discounts for Ultra
Go "https://labs.google/fx/es/tools/flow" then click "Nuevo proyecto". Finally enter your prompt in the text box. Click the arrow button to generate your video or image from text or ingredients.
when i use the generated image to modify it and again use the new generated image to modify some more the quality is dropping very rapidly, the image becomes blur and the quality is loosing, how can i prevent this?
Do they have Server Problems i cant go to my Library if i Go there the Site is crashing
Flow clips show "add to scene" button even after being added to scene. This just started (for me) yesterday 02/12/26
create a ticket in the #1238621511401734254 channel, i have the same issue
Why always failed generate ?! Im tired
Is there ever any plans to have precise mode when using images with nano banana pro in flow like there is in whisk?
hey, Flow keeps throwing the message that I've reached my daily limit of 🍌 Nano Banana Pro generations, although I've just upgrade my plan to AI PRO. Any ideas why is this happening?
Interesting. Is this Nano Banana Pro that you are using? Do you have a screenshot of your edits?
Flow TV --- How do you become a creator/film maker on Flow TV?
That stuff usually gets hand picked by the team/community managers I believe. Though I would hedge a bet more of it was done in house by Google employees
just as in today?
Pro has a limit. 100 per day i think.
It has limited generations because of the server, even though I have 1000 credits available. Right? I didn't know that, I hope they upgrade that system soon
Dieses Video bringt die rohe Kraft der Wikingerwelt in einem epischen 35‑mm‑Filmlook zurück.
Nordische Landschaften, Schlachten, Nebel, Fjorde und der Ruf der alten Götter – alles vereint in einer cineastischen Atmosphäre.
Ob Kampfszenen, mystische Runen, stürmische Fjorde oder der Klang eines alten Wikingerliedes:
Hier verschmelzen My...
Nano Banana doesn't use credits.
100 per day for only 20 bucks a month is a crazy good deal, on top of everything else we get. You'd pay a LOT more on other platforms for that many.
Oh, didn't know Nano does not use credits. But for me, some days I need more than 100 images per day, when I'm at the middle of a work project. Now there I know, I'll use it wisely 😉
Pretty much every platform (the ones with a menu of models, not home to one company's models) out there has Nano Banana Pro now, so if you need more on a given day, you can sign up for one of them (Freepick, Krea, Leonardo, etc.) and use credits there to gap fill.
Guys
I keep getting" generation failed, something went wrong." Error
I can generate nothing
No text to video or frame to video or any other service
I have Utra account and nothing works.
Hi @split sleet to report any product issues please use our #1238621511401734254 channel. When creating your thread please post as much detail as possible so that the team can properly investigate.
Thank you ☺️
Thanks
I will
also have issues with Flow right now.. your projects does not load etc
Hey, I’m new here, so don’t eat me alive.
I’m designer and working a bit in Flow, a bit through the API.
Honestly, the Flow interface, UX, video/image generation, and documentation are all at a tragic level 😅.
The quality of the software, the docs, the updates — it’s "classic Google x1000". Feels like flying Red Bull planes with duct tape, and I’m supposed to work on this.
1. Is there an official, continuously updated FAQ somewhere that actually tells me what the program supports on which day? (Formats, durations, number of images, image dimensions, number of image colors, etc.)
I spent a month going through documentation and trial-and-error. Nothing is concrete. Things change all the time. It’s infuriating.
2. Question that none of the eight support reps could answer:
Am I correct that in the ULTRA package in the EU on GOOGLE APPS, there is no Gemini API included (I already used the 90-day API trial), so I’m paying ~€200 for Ultra plus API access?
3. For some reason, after connecting the API to Studio, the Veo 3.1 model wasn’t available—and of course, none of the support reps could help or even answer whether it’s supposed to be there or not.
4. In the EU, for the ULTRA plan, I understand there’s also no access to VERTEX?
thanks
Describe what they do, not what they dont. Negtive prompting often backfires.
Hey all I have been generating images primarly by chatting to gemini. I have done hundreds of images that way. I have used flow for videos but just started using it for images. Would you all say that flow is better for image gen than chatting with gemini? I am on the pro plan. Is it better to use it during non peak hours? Should I be using whisk?
A gentle ambient meditation song designed to calm the mind, relax the body, and guide the soul into deep inner peace. Soft choir voices, slow breathing rhythms, and soothing harmonies create a serene atmosphere perfect for meditation, sleep, prayer, yoga, or quiet reflection.
Let the waves of calm wash over you as tension fades and stillness tak...
where can i check flow credits? Because ofc there is nothing like in help 😐 :
"To check your remaining AI Credits:
On your device, go to Flow.
At the top right, click your profile picture or initial Account Circle.
Under your profile information, you’ll find your remaining credits. "
ok i found it... Just account circle disapear. but link is still there
Why always failed generate zzz
anyone have greate prompt for smooth long scenes connection
anybody?
Ai came so far, yet changing angles is still a pain. How is this still a problem?
i need this shot but they must stand on the sidewalk. camera should be on the street filming the sidewalk, so the background is just building and shop in a row. no matter what i do it doesnt work. and also i think nbp doesnt understand distance prompts. i tell his the dog must be 10-20 meters away from the policeman. can anybody help please?
Hello @eternal cobalt as a friendly reminder, videos and images are not generated within the server, you'll need to head to https://labs.google/, pick a tool, and generate it there. If you are asking for prompt advice please do so in the appropriate channel. Thank you!
Hello @next kestrel , as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now. Thank you!
working on "CameraStudio" for Flow
Hi, is anyone else having trouble generating videos in Flow? Mine keep failing, and there's nothing wrong with the image or prompt.
hello
Noep
Nope
i want to chat
Dig it.
#1238621511401734254 too many black screen videos and credit losses, even with Veo3.1-quality. Refund credits if no video is generated.
Yes, at least they should refund credits but they take credits and return the videos with a single black screen..
Draw it out and give it as a scene blocking reference
It typically does.
It usually does. If it doesn't for you, then contact tech support.
I built a scene blocking tool with camera references in a 3d space
That works too but just drawing it works
Even a bad line drawing
I did, maybe my drawing sucks. Didnt work at all
Label it too
Ok
Label it as foreground, middlegroubd and background
Hi @autumn lantern you have to create a thread in the #1238621511401734254 channel if you are having issues. In the thread please leave as much details as you can so that the team can properly replicate and investigate said problem. Thank you!
how ?
Flow is terrible failing I don't know what is the problem anyone have similar issues with flow and nanobanana pro
Huh i have pro and I dont have limits at all (is it bcs i live in sweden?)
Send a link to flow I am to lazy
Anyone getting failed generations but still getting charged? I just had 5 fail in a row and charge me for them.
You do, you just havent reached rhem.
Normally you'll gte them back. If not, contact tech support.
Where do I contact support at?
Never mind, I think I found it. Thank you.
yeah there are limits, some days I just download my old previously generated images via 4k, and the limit gets exhausted after a while, then it wont let me generate a new image, till the server resets.
hello! can i use flow if i don't have google ai pro?
if so, how many videos can i generate per day?
how you are doing this ?
extension
I think there might be a bug in the code that was causing me to not get refunded for a failed generation, I figured out that I was using an image I deleted so that might be why.
I have a veo 3.1 api key and a base url server, how do i use it?
Yeah, it's easy to find if you click around.
I was looking for it in flow, found the chat option in Google One.
Yes. Go to https://flow.google
You get 50 credits per day, which is basically 2 videos per day using Fast mode.
Flow is an AI filmmaking tool that lets you seamlessly create cinematic clips and scenes using Google DeepMind's most capable generative video model, Veo
Thanks for the response. I have posted the same on labs-bug-feedback.
Does anyone else get error messages when trying to download upscaled images?
#flow-general・🌊 One Piece: Luffy fights Zoro
that's actually quite generous
is there a slow mode otherwise
If you have an Ultra account, it's free (free isn't right - it doesn't eat credits) to run low priority Fast mode generations (but during peak hours you'll get more fails).
Yeah, and you get a few gens in Gemini as well, AI Studio, too.
how much does it cost to be ultra
$250/mo
can you share it on several accounts like for a family for example or no
I thought no, but I saw someone say yes a couple weeks back, so don't know.
If I buy Google AI pro do I get 1000 for daily use?
and how many credits is a video?
That's monthly, but you only have to spend it on certain things. Other things have daily quotas. Nano Banana Pro you get 100 generations per day, for instance.
If you use Fast mode in the Flow app, it's 10 credits per. 100 credits for quality mode.
You get some video gens in Gemini per day (forget how many) that don't use credits.
thanks man ❤️
Three a day in Gemini.
There's also Google Movies where you can do ten a day and it also support image uploading, but has a 100 character limit IIRC
*Vids
Vids yes! Thank you for the correction 😄
Vids is great because it's also an editor platform and can store projects much like Flow
Biggest perk of Google Vids for me is the free stock audio (also images and video).
It's basically the same thing as Canva video editor, but just less fleshed out.
anything but fixing nonstop generation errors and falsely blocked generation of sexual content
I've not used any credit today on flow
Why am I getting this
How to remove watermark from veo generations.
Credits are monthly if you have Pro or Ultra. Daily credits on free accounts just haven't reset yet if you're on free.
Type your prompt into Gemini and tell it that it’s triggering the censorship and you need it to refine the prompt.
since when is "show these 2 men face to face from a profile side view looking like they are engaged in a serious conversation" a content warning trigger?
literally just 2 men talking.. it just wont let me generate it today
ive used it hundreds of times before
Worth trying, at least.
One guy a while back had images rejected, but couldn't see why, so he showed it to Gemini. Turned out there was a kid hiding in the background somewhere, just barely, and erasing that was all it took to get it working.
If you're here just to vent, then vent away. If you want a solution, then all you have available is... try different things until it works, or give up.
Up to you, mate.
try different things like what? it dooesnt get any more straightforward than that.
Perchance... "Type your prompt into Gemini and tell it that it’s triggering the censorship and you need it to refine the prompt." @near gate
🙂 I've done that and also reformatted my prompt into JSON. Alternatively, and even with better results and consistent memory, try to google docs refine feature.
but stilll
nothing is workig for me
Oh... i see what it is. I tried to re-upload one of the images (which was created by nano banana) and for some reason that AI generated man is a promiment figure. genius.
screw it. ill just have grok do it for me
Prominent person error from image I created with nano banana pro? Something is wrong with the filters AGAIN?
im running out of image generations for flow like theyre soft banning me for generating too much and i have to wait super long. i tried using another google acc but its the same but that other google acc doesnt haev a paid plan. if i get a paid plan on two diff google accs can i alternate between the two when one is soft banned or r both them just connected and even i have 2 paid plans on 2 acc i will still be unable to use image generation ?
hey guys is where can i check the prompts on the background videos in https://labs.google/flow/about ?
@edgy holly do you know if there's a difference in quality in terms of the video that is generated whether you use a 1k, 2k, 4k (start) reference image. Does it anyway compress the image if you upload a 4k one?
Hey everyone, is anyone else getting errors while downloading a 4k image in flow ?
I'm constantly getting "Couldn't generate image, Try again later"
No errors while generating, only via downloading through 4k
I've experienced that, but more like in 5-10% of the cases.
@minor badger aah I see, I'm not able to downlaod anything, each download attempt results in the same error today, it was working fine yesterday though.
I changed browser too, opera, chrome, edge, firefox, all to no avail 🥲
This is frustrating. Does it happen in all projects? or just the one?
I created multiple images yesterday, and ran out of the 4k download slots for the day, so I waited for a day, and today I'm not able to download any of the images, they are like may be about a 100, I tried with atleast 50 of them, same error haha
❓Good morning! ....(there have been many update to veo3 - been 2 months i'm away) What happen when we sub to the 250$ plan, is it still unlimited creating normal Veo 3 or we have to wait in Queue for fast video? thanks (edit How many Nano banana *Pro Picture can we create per day?)
there was a minor update to veo 3.1 from veo 3,
videos are still unlimited.
But nano banana pro has a limit, I'm not sure how much, I think may be 250 or something like that, I may be wrong on the limit number though, but there is a limit for the images.
I'm talking about the ultra plan the $250 one
My Google Ultra AI is set to renew on February 26th
Seedance 2 releases on February 24th.
-> 2 day test period to potentially cancel Google Ultra 😅
Honeslty I', blown away by the quality there, although there are something which it cant do, I mean it can, but veo does it better inmo
I think it depends ont he use case, like the kind of content you are generating..
my used case prefers veo for now, it may change later hahaha
If Seedance reveals up to expectations. And considering all issues with Flow, I won't hesitate.
But again.. Veo 4 might release on February 25th lol and then... dilemma
I doubt it lol, there has been absolutely no news on veo 4, is there though?
I might have missed something
I'm just dreaming
I hope they do, if anything, the UI requires an averhaul, the present UI is not very friendly and has a lot of errors.
The thing is the Veo model is completely independent to the flow UI. Can be used over an API. So Veo 4 won't necessarily equal new flow.
it would be such a dream, if they introduce a select option for videos and images, where you can just select them, and then choose to either download all at once, or upscale all at once, rather than sitting wasting time, upscaling downloading one at a time, when you have like 100-200 generations, it takes the whole day
I see what you mean, veo 4 would be good regardless, yes.
I've made a 8 min long short. It's a lot of generations 😅
Brother I feel you, I make music videos that are anywhere from 2 minutes to 6 minutes long most of them being the 4 minute sweet spot of straight lemming drop off in engagement. The amount of generations I am going to look up right now lol I think there is a tally in the main google account.
and half of the time it takes just to download from the UI,no generating its blazing fast, but the bottleneck is the downloading LOL
Furthermore, in the middle of it, UI starts giving errors on the download, then you need, to refresh a couple of times, or change the browser, it just wastes a lot of time.
I'm not complaining, but the UI bottlenecks are expected from smaller companies, NOT from Google haha
and many times, you think you have downloaded everything, but while upscaling and downloading, it automatically misses a few upscales, it shows its upscaling them, but sometimes just misses them, and you get lesser downloads.
cant find the page anymore, I used to be able to see how many transactions I was making to flow labs
which transaction page ?
i'm still new to veo 3, is that normal to get lots of Failed Generation these days? (due to sounds) for 100 video, how may video you guys get rejected or failed?
Hi team, Flow was working perfectly yesterday but is completely unusable today.
Symptoms: Stuck on "Loading..." screen or constantly redirects me to the landing page.
Troubleshooting: I’ve already cleared cookies/cache specifically for labs.google, tried Incognito mode, and tested different browsers.
Status: My subscription is PRO, is active and verified.
Is there a known server-side outage or account validation bug? It seems like an authentication loop. Thanks.
might have to be under a family plan, I got rid of mine so that might be why i don't see it. I have ultimate now couple months ago i had ltimate and had a few people under my plan and I could see all the transactions
Yes
It has to, yeah, because it generates at 720p.
OK I just did a test.
If you upload an image below FHD (1920x1080), it keeps the original resolution. For example 1K (1376x768) stays 1K.
But if you upload anything larger than FHD, such as 2K or 4K, it downscales it to FHD.
Of course it's still amazing! One of the best.
I made a microseries using almost entirely veo save for about 5 secs out of 28 mins of footage and it all looks like this
With consistent characters , sets, wardrobes and props
hello guys
Not that other models don't do better than Veo on some things, but boy... NONE of them are as fast. I have access to all the top models on a couple other platforms, and it's torture waiting for single gens that take longer than I could get 12, 16, maybe 20 in Flow, iterating each time.
Hi everyone. I'm new here, and I'm not sure if this is the right place, but what do I do when Flow generates a black screen and swallows the credits?
why cant i upload my self when i was a kid photo to edit?
wdym i dont even work for Google🤣🤣🤣😭
Seedance 2.0?
{
"scene_number": 3,
"scene_title": "Leaving Home",
"duration_seconds": 8,
"style": "ultra realistic cinematic travel vlog",
"character": {
"reference": "use uploaded image as exact face reference",
"expression": "determined excited",
"movement": "realistic walking physics pulling suitcase"
},
"environment": {
"location": "American suburban house exterior",
"time": "early morning",
"details": [
"front door",
"driveway",
"parked car",
"quiet street",
"soft sunlight"
],
"realism": "must look like real recorded footage"
},
"camera": {
"type": "cinematic tracking shot",
"motion": "smooth gimbal follow",
"angle": "front tracking then slight side angle",
"focus": "sharp subject background depth blur",
"lighting": "natural sunrise lighting"
},
"shots": {
"0-2s": "door closes behind character",
"2-5s": "walking forward pulling suitcase",
"5-8s": "character glances back at house then looks forward"
},
"dialogue": {
"tone": "natural vlog tone",
"line": "Alright… here we go.",
"lip_sync": "perfect realistic sync"
},
"audio": {
"voice": "natural human voice",
"background": [
"morning ambience",
"rolling suitcase wheels",
"light wind"
]
},
"color_grading": "cinematic travel film tone",
"realism_rules": [
"must look real",
"not animated",
"real shadows",
"natural skin texture",
"real motion blur",
"realistic lighting physics"
]
}
HI google team. When are you releasing a new version. It's strating to feel a little old. I'm an ultra subscriber and am just kind of hanging in here. thanks
It says the video has been upscaled and downloaded, but I never receive the files in my offline storage. It's a lie and a waste of my time. That's after having a 1:100 success rate with getting the image or video you are after with these tools. What a crapshoot.
Is photo ko koi sa video mein convert kar do
Some gens from people I've talked to have taken over half a day. Hopefully it'll come down, because it looks great.
They don't let logic interfere with their venting, haha
🤣🤣
Hi
anyone here had succes on youtube?
true success with AI videos, I would invite you to check gossip goblin.
Censorship with minors.
Hey guys! Im got a problem with Google Flow. Im from Russia and im trying to enter Flow with USA VPN but it still says im in not supported country. How can i fix that? (2ip is saying im in Chicago)
Google uses a series of safety features, not only IP address / location. Just the fact that you might have created the email address outside the VPN is enough for Google to trace.
dayum im just want to generate loop gif for my steam artwork :(((
what about using third party platforms like openart? you get access to all models.
You can't. They can tell, and have to follow the relevant laws. VPNs don't work.
^ this is the closest thing to a "workaround". Google sells to them, and they sell to you.
Edit: MIGHT, anyway.
Hello, I have a question, how can I submit an experiment to google experiments because the submit button is not working
Is anyone here still experiencing reCaptcha errors when producing veo3 flow videos?
I was made for minutes… not centuries.
One sip and I’m tossed like I never mattered.
I don’t disappear — I linger in landfills and oceans.
Convenience shouldn’t outlive you.
hello guys, Can someone explain why my PC always encounters an error when generating a video in Flow, while my laptop does not? This uses the same network.
I got seedance 2.0 and this is only thru on lower tier accounts
But they need to update Veo tho
Same Flow is using Veo tech right? just a better interface
Huh
I just realized that free accounts get more credits monthly in total than pro accounts!💀
If they use it everyday
In a free account you get 50 daily redits on flow+ 100 monthly.
Total 1600
while on a paid account you get 1000credits only
Is this a joke?
Paid account also get daily credit
They get daily credit with Gemini right ?
No
This is my paid account vs my free account
Hey, I got some issues with the upscaler. I've generated a couple of videos before upgrading to Ultra and wanted to re-download those in 4K now. It works fine with every generated video but one. The pop-up tells me that there was an upscaling error and that I need more AI-Credits, however I still have more thank 24k Credits in my account?
It works fine with every other video I generated before switching to Ultra. Any idea, why this could be?
@unreal wrenHello, as a friendly reminder we cannot generate images or videos in the server! Please do not post your prompts in the channels as it will not generate. If asking for advice or tips on your prompts do so in the appropriate channels.
Are you using a VPN, or some kind of security extension?
Generations in Gemini don't use credits. They have seperate daily limits.]
How do I put the same voice?
the same voice for what? its completely different people and scenes? can you give more context?
can someone solve this guys?
what kind of error?
I'm creating a documentary with the new feature. I'll extend and meter the prompt and voice like this: Visual Prompt: Extreme close-up of young Caesar's eyes, reflecting the warm flicker of a nearby torch. Intense, cold determination in his gaze. 4K detail on skin and eyelashes. Characters are silent and do not speak.
Narrator Text (Sal): "It wasn't just ambition that drove him, but the firm belief that he had a date with destiny." (10s). But the narrator's voice changes each time.
always like this
ok just skip the Narrator text. generate the Broll in veo3 and create the Voice Over with Elevenlabs or similar. two seperate processes.
yes I knew that, but I wanted to do everything in one place,
then try the "jump to" function maybe that maintains the narrator voice, but idk its infinitely times more complicated and timeconsuming to try and do it all in one place.
Yeah they confirmed this in the feedback channel the other day.
Obviously they can set these different credit limits however they like, but it seems bizarre that if you actively pay for the service then you get access to fewer credits overall. There are certainly Pro Plan benefits available beyond just the AI credits for Flow and Whisk, but if that's someone's main use case then they'd potentially be better off not paying at all
bro i have ultra plan
it is very expensive, for these things to happen is INADMISSIBLE
🎬 8. Visible Swelling (5–6 sec)
Red bump forms gradually on skin surface.
🎬 8. Visible Swelling (5–6 sec)
Red bump forms gradually on skin surface.
Is flow having issues today? I can load my libray and my projets, both on PC and telepfone.
It shouldn't, but I have a completely stock Chrome browser install I only use for Flow. No extensions. Leave settings as-is.
Shouldn't work, but it does sometimes.
india?
the same case as mine
I think the problem is just with generating too many at once, that's what's causing the error, but why only on the PC? I tried it on the laptop and it always works.
Should I reinstall my Google Chrome? or do you have another solution?
That's why I'm suggesting you try a clean browser.
You can install Chrome Canary. Just don't change any settings, or install any extensions. Seems to help, for some bizarre reason.
A lot of users are experiencing an issue where the generated video only shows a black screen.
Can someone please check what’s going on?
I need help people, I am not able to upload anything right now,
and ive tried everything I possibly can, this is google flow
I have AI ultra but it doesnt matter apparently
[I need help, 😫 ]
I am trying to create a short 3-second sequence using three images.
Image 1 is the starting frame (with HUD visible).
Image 2 represents the middle glitch effect.
Image 3 is the final frame (HUD completely removed).
The camera must remain completely static.
However, the model keeps animating or increasing the gauge bar in the HUD, even though I do not want any value changes.
How can I prevent the model from animating the gauge bar?
I think Veo 3 tries to animate everything in the frame, and that’s my main pain point. Some areas are supposed to remain completely static.
This is the prompt I used:
First-person POV. Static fixed camera using image1 as the exact starting frame. The cave environment is visible with the HUD overlay in the lower left corner. The HUD is a baked 2D overlay layer. It must remain pixel-identical to image1 during the stable portion. Do not reinterpret, rebalance, or recalculate the gauge bar. Do not animate any values. Do not adjust the bar length. Do not synchronize the bar with the number. The number "3/100" is part of the static image. It must not change. After a brief stable moment, introduce subtle horizontal line distortion based on image2. Mild pixel displacement and soft scanline interference appear. The distortion affects alignment only. No UI logic changes. As the glitch stabilizes, the HUD layer fades out unintentionally. Final frame must match image3 exactly, showing the same cave environment without HUD. No camera movement. No value animation. No automatic correction.
Guys still error using flow?
Animate the chibi miniature character from the photo so that the face, hairstyle, and clothing remain exactly the same with no identity changes. The character suddenly becomes lively and starts jumping up and down playfully, then transitions into a cute energetic dance with small rhythmic steps and arm swings. Add natural body bounce and squash-and-stretch motion to enhance the chibi style.
Expression animation: bright eyes, wide happy smile, cheerful laughing expression, subtle head tilts, blinking, and lively cheek movement. The character laughs joyfully while dancing.
Motion details: quick light hops, spinning half-turn, playful foot taps, hands moving side to side, then a final happy pose. Hair and accessories slightly bounce with each jump.
Camera: medium full-body shot, slight handheld feel, gentle cinematic micro-zoom in during the dance.
Style: cute chibi 3D animation, smooth motion, cinematic lighting, soft depth of field, high detail, vibrant colors.
Duration feel: 3–5 seconds loopable animation, energetic and joyful tone.
Dude just use capcut
drop the images and use a glitch effect it will be static and like you want
I am consistently getting “this generation might violate our policies” error when it’s literally an input like “car driving down road” anyone else facing similar issues or know a fix?
Do you know if i can download videos from google labs web.
I mean i looked up if they are in files but cant find them.
Anyone help?
I’m 90% sugar and zero chill!
You stab me with a straw and expect calm energy?
I spike, I crash — I am the drama in this lunchbox.
why the phuck did all my chats disappear
You have to share the complete prompt and input inages if you want help on things like that.
So I inadvertently ended up making a feature film (75mins approx) almost exclusively using Flow (albeit with music via Suno and some sound effects off Pixabay) and while any lessons will no doubt go out of date within a matter of weeks, my main takeaways were:
-
Don't do it, because...
-
It took me around 3 months, which meant Veo 3.1 and Nano Banana Pro came out while I was making it. The difference in quality between the first scenes and the last is significant simply because the models are improving faster than someone can turnaround a whole film while working on it in their spare time
-
Flow is horribly crude as a filmmaking tool and seems to have the most restrictive content blocks for Veo 3.1 and Nano Banana Pro (i.e. it'll block things that Gemini or the API won't); I'm sure any other platform would be a better choice
-
But the Ultra plan makes Flow a compelling option. While expensive, being able to generate a much video as you like for a fixed cost is less stressful than watching costs rack up as you try and get a workable generation
-
I'd guess 90% of the production time was spent fighting the models - trying to get a reverse angle shot of the same scene, trying to get someone to move left rather than right, trying to stop characters turning into celebrities, etc. All the stuff that's 'easy' for a real world production is stuff that's hard with AI.
So, on the one hand, it was a frequently (constantly?) frustrating experience, but on the other, it is pretty amazing what's possible now.
One final note: all the writing and editing was entirely done the old-fashioned way. I don't think AI is anywhere near good enough in those areas for the kind of filmmaking I've been trying.
Sorry for google but for experts the most recent veo model doesn't satisfy them.
Still enough for your hobby
5 is the weakest part of Veo. Has been for a while.
Only thing I disagree with is 1. Do it. Roll with the evolution.
Point 1 was a bit tongue in cheek. Better to be riding the crest of the wave than watching from shore. Though i have tried to adopt the rule of 'never go back'. No point attempting to improve the janky earlier footage because when you're done, the latest models will be better again.
Doesn’t work, I tried several times
"Better to be riding the crest of the wave than watching from shore."
100%
What is the time of reset in flow to generate images again?
You guys need to make some sort of timer showing you why you cant keep generating or better error messages, this is by far the worst software I have used in my life for locking you out every 5 minutes with no explanation. Best images out there right now is the only reason people deal with it.
Can some1 help me out literslly every time i try to generate anything pictures that cost 0 credits for example, i get this Errol " cant generate image. Try again later " on every device...
Maat tyfus hinderlijk ofniet
Vgm alleen maar mensen in NL dan
ik zou het echt niet weten mr zou best kunnen
Had dit gister ook al
Idk met vpn werkt het nog steeds niet
@topaz vector ging het gisteren wel weg
Probably servers topping out with the release of Gemini 3.1
Definitely not the prompt.
Oohh okay
Thank you bro
Had the same problem
hello, I can't create image. It's just me?
No, it seems to be a general problem. Others have already reported it here, and nothing has been working for me for two hours either.
Nope
I'm getting a 403 error and just found an extension that can fix that damn error.
Google is doing something that's causing manual video creation to result in a 403 error due to reCaptcha, but I just found a fix using this extension.
STOP RESTRICTING USERS
flow totally unusable for me today no work done and gonna give up, no NBP, Veo's not working either its adding dialog and as soon as i try adding no dialog or no audio to the prompt it just fails the gen everytime, seems like theres major issues going down right now at Google.
So many people are using Nano Banana Pro via Flow for work, please stop restricting image generation even without content violations, it makes working with SO MANY absurd limitations with completely legal content very difficult. PLEASE DEVS STOP DOING THAT.
Yeah, it ebbs and flows every day, following when the most people in US, EU and Asia normally use this stuff, but fails ALWAYS spike the few days after a big release like this.
If u get the ai subscription with free trial of a month the problem is resolved
Youll be able to use flow without any problems after
They don't go away with paid accounts, but it is less.
My problem litt went away
Usually all you have to do is wait a bit and it'll work later.
Few times a month I run into issues, but I just go do something else for a while (never short of other things to do... music in Suno, voice stuff, editing, keyframes for upcoming scenes, etc.) and it usually works when I come back.
Sometimes it even works if you jump from Flow to Gemini, or use it via the API in third party apps (they appear to prioritize API calls).
Just gotta be flexible.
Hi Hyeokjune, see atached. The root of what you are experincing is down to VEO 3.1 having cienematorgphy architecture as the back bone of the logic. Compared to older "video to image" or "image gen" models, where prompts that gave the model freedom to interpret tendeed to output better results, now the lack of those pre determined constraints its the cause of the hallucination. I have done my best to break down the new essential logic required as port of the promt + a reworked metada template that you can use as the backbone promoting to guarantee consistency. Let me know if it helps!
So many people are using Nano Banana Pro via Flow for work, please stop restricting image generation even without content violations, it makes working with SO MANY absurd limitations with completely legal content very difficult. PLEASE DEVS STOP DOING THAT.
ah-
@main saddle hello, curious if you've been able to read me Q.s thanks!
The guardrails are to strict - converting a sketch style out put of a fantasy character in to photorealism >> I can create images of people, but not ones that depict a real person like that. Do you have another idea you’d like to try?
Alwaaaays error
It's been like this for several days!
There's been no help from Google
anyone ever gone more than 10 minutes without this error?
why are you using veo for this? can't you do this in a video editing software?
Flow working for you guys? I can't upload anything
I'm in Ultra plan but why can't I create anything, images, videos, etc.? Flow keeps giving me an error: "something went wrong.". Meanwhile, on the same computer but with a different (free) account, I can create things easily?
I started to interact with "Flow" this month, and it has made my life better than before. I really enjoy that it brings my ideology closer. Lastly, I will look forward to keeping it alive within my lifetime.
Try using this.
im unable to generate images in flow anyone know the reason why? i used to use ultra now im just pro but even after waiting a whole day i cant generate images after generating a lot yesterday how long is the cooldown ??
I feel like Veo 3.1 is falling behind a bit.
😹
Brooo. You're my savior. You're just awesome. I was looking at POST GET requests, but I thought I had a problem with my VPN. I couldn't solve the problem for a long time. I went to see if everyone else had the same problem. I read that many people did and almost gave up on it all. Thank you, you're a hero!
No problem, I'd also like to help others. If you know anyone who hasn't been able to fix the "something went wrong" error when creating photo videos with Veo3, please recommend it to them so they can benefit.
Try using this tool to fix the error.
whats the live request log for? @desert geyser
It helps fix the errors caused by Google's reCaptcha.
I mean do you got ultra on your google account?
Yes, I am using the Ultra plan purchased from Google.
Dms
How to fix the "something went wrong" error when not creating videos and photos on the Google FLOW page.
@main saddle Is there a time frame for a fix for flow? 2/3 days now nothing is working just failed gens all the time, Im running way behind on my work so would appreciate a reply...
You can temporarily use my solution above.
how is this going to retify failed gens? is this app/script from Google?
greetings guys...please i'm new to google flow and i have a question: i generated a script and inputted in google flow, the video generated was only but the intro of my script....please what am i doing wrong, or what do i need to do to solve this problem
Good evening @vapid moon pls I also new here too i want to know about how to generate script and what it is meant for how they use and make video
and please how do i maintain the same character since it seems one needs to keep adding scenes
hi guys . how do i remove the veo watermark from the downloaded videos my plan is pro
Flow makes 8 second clips. Break what you want to animate into pieces.
Reference images
where veo 4?
where can i speak to support ? all my video are signaled as not following the rules
You reference via the expand feature right????
Okay thanks
No. You make images of the person you want to show consistently, and upload them as a reference (either keyframe to video or ingredient to video). Otherwise the model can't be precise enough to make them look the same just from text prompts.
Okay thanks a lot
Maybe I will would try to watch a YouTube video now on how to do this
here's a link! Google One Support
With Google One, you can claim your new benefit or trial from the Google One app or the Google One website. You can also get notifications about the latest updates and offers from Google. Tip: Googl
Just make sure to specify Veo 3 or veo 3.1 when searching. Veo in Flow isn't the same as other tools.
Hey in the country section why is there no option available for India?
hey! sorry about that - can you try again?
I already filled out one with a random country so it says you've recorded response 🥲
@main saddle hello may i know the cooldown time to make images? its been 2 whole days for me and i still cannot generate images after initially generating a lot. i used to only need to wait like 8 hrs now its 2 days
Bro bypassed the copyright law restrictions
I'm not trying to circumvent copyright, I'm just helping to resolve the "something went wrong" issue.
Because I also experienced a similar situation even though I had purchased the Ultra plan from Google.
Okay, it's fine for now, thank you bro, I really appreciate your answer.
How technology and AI has shaped my life and educate me in my dream career as a medical doctor in the future
How to purchase flow subscription??
Happy to help.
Is VEO 3 generating AI videos for you?
Was an hour ago, or so.
Some creative I made that I'll be using. It's so awesome! Can't wait to play more!
Google Vids?
Made those with Pomelli, it's an awesome asset generator and I literally made those in like 20 minutes my first time using it
But I do use Veo 3.1 and other AI tools for image and video generation. I made a Happy Birthday music video for my granddaughter when she turned 2 using Suno and another AI tool I use called freebeat.ai and it was awesome, she loved it
