#flow-general・🌊
1 messages · Page 36 of 1
so if you make your prompt clearly english it will help!
Again this is combined with another issue we are working to fix in I2V
I would try it out, but i am not paid to betatest and instead pay a huge amount of money and have no credits left, cause the included 0 credit veo 3.1 was changed midsub
?is flow down
Hi Yall, pleas try I2V again!
Flow does not handle non-english prompts well, and we somtimes reject a request if its not english
@slender haven @cinder sail @spring igloo please try I2V again, and thank you for waiting!
Working 👌
Is the Nano Banana within Google Flow also outputting 4k? I don't think i see that when i download.
Please check my problem
not yet, sorry!
looks like you've reported an issue with Jules, please repost in #1374125660859142254 so our Jules crew can take a look. btw, [a fix is rolling out for Jules](#jules-general・🦑 message) atm.
my god gemini 3 is horrible. how was this slop factory greenlit
memory of a goldfish
chatgpt's free version is even better
Yes sorry my mistake
anyone notice this flaw in nano banana and even NBP that has this strong tendency to crop subjects at the top of their head?
you're good! just want to make sure the correct team sees it 🙂
lol ok that was a little bit of hyperbole but I had just gotten out of a war with it. It claims to integrate well with GitHub and it can’t access anything from my repos beyond info on their landing page and directories. I know. It’s outside the scope of this channel and I’ll talk about this where relevant. But 😡
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. we have #1319424187730755636 for feedback specific questions/ queries
i just quit using capcut. (even though i got nothing against the software itself)
But when using capcut you grant them your entire usage permissions of your content and even re-licensing (selling) your content.
Not with me boys.
Oh, the google staff @main saddle @iron matrix has visited the channel.
Any news on how long you let your customers down with their ultra subscriptions? Google is violating their own TOS and disappointing a lot of their customers.
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. we have #1319424187730755636 for feedback specific questions/ queries
I’ve been waiting for a long time for it to continue generating. How can I proceed when this error occurs?
Hello from the Flow team! We just reintroduced a zero credit Veo 3.1 Fast option for Ultra users. Ultra users will see two Veo 3.1 Fast models in the model dropdown:
“Veo 3.1 Fast”, costing 10 credits.
“Veo 3.1 Fast [Lower Priority]”, costing 0 credits; these generations get deprioritized during heavy traffic and have lower limits on concurrent generations (for now). We’ll be experimenting with different settings over the coming days.
This gives Ultra users a way to generate for zero credits, while prioritizing paid generations under heavy traffic. As a thank-you for bearing with us, today all Google AI Ultra subscribers received an additional 2,500 AI credits added to their account! Request: can Ultra users test the [Lower Priority] model and let me know if it's working as expected?
Is the lower priority model diluted at all? Ie less compute cycles?
No, it's the identical model.
Thank you!
I only created two videos today and it’s already telling me I’m generating too fast. So what’s the point of the 0-credit option? It doesn’t seem to have any effect.
It seems like you've been having that issue today even before we relaunched the zero credit model, is that right?
That's right
Can you DM me the account you're signing in with so we can look into this?
I slept for 8 hours and just woke up, but I’m still seeing the ‘creating too fast’ error.
no more gens in this channel after this but...it's nice, will be runnig gens tonight
Assuming that was from the zero credit model, it worked as expected right (didn't deduct any credits, allowed you to submit a batch of up to 4)?
I just tested again and created 4 videos at the same time.
While those were generating, I tried to start another one and it showed the ‘creating too fast’ error. After the 4 videos finished, I was able to generate again after waiting a bit.
However, sometimes a video gets stuck at 99% and fails. In those cases, when am I allowed to start a new generation without getting the ‘creating too fast’ error again?
i only submitted one, let me try again with 4,
You should be able to make 4 requests each minute with that model
yippee 😄
cool change, dunno how a low priority service is gonna feel tho when basically already quit using flow because constant failures 8/10 generations
Want to ask in Flow is there any way you can lock seed, if so how do you do that, if not is there another tool I can use, and also how do I do videos less than 8 seconds
How slow is it? asking before I shell out another $250
Generation speed is unchanged; same as Veo 3.1 fast speed has been (my last one took almost exactly one minute)
Can you explano more detailed how works this new way?
works very well, thank you for listening, I'm so excited to get into my project now, huge opportunity for me and glad I can do it using VEO
yay 1x concurrency
In the next update to veo they need to make the video extend this better it looks like a cut every 10 seconds
man what a convoluted solution. instead of just implementing one model with credits for priority gen which defaults to a relaxed queue, you have to select it, and it cuts your concurrency to 1x. incredible fix, lmao
In scenebuilder, use the "save frame as asset" - then use that with frames to video, works much smoother for me
it works but its a bit like "what cs failure is writing the frontend"
Could you please confirm whether the limit is exactly 4 requests per minute?
Because in my tests, sometimes I have to wait 3–4 minutes before I can generate again, even if I only created 4 videos. This inconsistency makes it quite difficult to work with.
If possible, could you update the cooldown so it resets consistently every 1 minute?
I will see but sadly I dont get to play with it for too much longer this is my last month
Figured this would be the case
It's exactly 4 requests per minute right now; we'll adjust it as we see what load looks like.
Feel free to DM me your account so I can see what's happening there
Has anyone here tried sora 2 pro i looked at it but it seemed like a bit of a hassle
why can't we simply drag and drop images into ingredients / frames
Working on that!
Maybe it takes more time to generate?
Does lower priority means less quality too?
failed gen failed gen failed gen failed gen failed gen
this platform is legit garbage now
I hope lower priority is just a speed thing and not a quality thing. I cancelled my ultra membership earlier today anyway over this bait and switch
first I thought it was content in ingredients like a policy trigger so I swapped to totally mundane stuff like a vehicle and I'm still failing every generation, it's ridiculous lol
this just appears over and over again til all 4 videos have failed
No…..it clearly means it’ll take its time being queued nothing to do with quality
A 90-year-old elderly man wearing a cap and holding a walking stick. A 12-year-old girl stands besid
same here, 70% failed
A 90-year-old elderly man wearing a cap and holding a walking stick. A 12-year-old girl stands beside him in a winter morning village scene. She asks him a funny question in Bangla, and the old man also replies in Bangla.
Am i the only who has nano banana generations constantly failing?
Did you see that notification has been removed?
wich notification
Free veo3
I saw that there is 0 credit lower priority, i just ask myself if lower priority means less quality too?
That notification is gone, go look
Haven't worked for a few hours for me. I try every little while.
flow error ?
Ok, but it's not listed anymore
But at least it's good news!
yes, thanks google for listeting
Yes
^^^^ all night bro😂
Sorry my bad again there's several announcements links. 😆
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.
getting this almost every image,
Couldn't generate image. Try again later.
anyone else having nano banana pro issues in google flow?
text to video seems to be working smoothly
nano banana pro have this a lot error "There are a lot of people I can help with, but I can't edit some public figures. Do you have anyone else in mind?" you have solution tu counter this ? plz
everyones trying the banana
Has anyone else experienced that they could generate Nano Banana Pro on Flow until yesterday, but can't do it today?
I can still generate Imagen and Nano Banana. If it's not on Flow but on the Gemini side, I can generate Nano Banana Pro.
This screen popped up, and after I clicked OK, it seems like I can't generate anymore. How can I revert it or get it back to normal?
yes thank you it work,Please excuse my inexperience. My concern at the moment is that I want to maintain consistency between my different scenes. Same character (provided as input in image form), same voice, same language spoken. Currently, with each generation, I have a different voice and sometimes even the language switches to one other than the one I provide in the script input. How can I fix this, please?
Yes - NB Pro is not working in Flow for me - 99% then failed gen notice. [EDIT: finally got a generation to work after a bunch of tries - but they are unusable due to unprompted text being added on top, LOL!]
I am facing the same issue. Nano banana is not working on flow
Can videos be made here?
Did the 2500 credits get sent out already? Because I didn't get mine :/
It works just keep using it
Hello plz someone tell me free videos Can be made here?
I think the throttle they set is too aggressive some of my Nano Banana Pro image gens aren’t going through
Ultra
Pretty sure it's cooked on flow rn since I cant get anything to run
I checked this with ChatGPT and this is the result:
Capcut cannot “relicense” your content
To re-license (sell) your content, CapCut would need:
Exclusive rights
A signed transfer
Direct permission
Legal ownership
They have NONE of that.
If they even attempted it, they would be destroyed legally.
“worldwide, royalty-free, sublicensable license”
This scares people, but it only means:
They can show/process your content globally
They can use subcontractors (cloud servers)
They don’t pay you for using their editing service
It does NOT mean:
They own it
They can sell it
They can claim copyright
They can redistribute it outside the platform
It’s identical to terms used by every major platform.
CapCut’s terms are the standard “license to display your video” that EVERY editor and cloud app uses, including:
YouTube
TikTok
Instagram
Adobe Premiere cloud services
DaVinci Resolve cloud collaboration
Filmora
Canva
VN
Clipchamp
Even Discord uploads
There is a bug with making videos favorite .. it doesnt work .. when i mark something as favorite it doesnt actually make it favorite .. revert back
Yes I have noticed that now for a month or so
Is it nearly impossible to generate Veo 3.1 Fast today using the free plan?
Just tried Flow today and did a 30 seconds short that looks amazing. However, it was very frustrating that sometimes it would change the language or the look of the character so I had to re-do some scenes while you're still charged credits for every scene you create, which drained all my credits and the result was only 30 seconds. Anyone else facing these types of issues or have you been able to maintain consistency somehow?
There is no diffrnce in speed of fast creation and slow priotity . They both slow af .
Hello, "This model is temporarily unavailable due to high demand. Please use a different model or try again later." what is this error?
the announcement is at least more than I suspected it will be, but will have to see in use how usable or unusable deprioritized will be
how long does it usually take the generate ?
How long has it been since Flow was so saturated? I'm just losing credits and not being able to generate anything. Is there any specific timeframe where there are less users so we can generate?
For me both video and nanobanana pro won't generate anything always show high demand massage since this morning
Yea nano banana is not working in flow
I subscribed to Gemini Ultra in order to use Nano Banana, but this doesn't make sense and I'm really in trouble.
You mean pro version .. right ?
Yes on Ultra sub !!
Im working for marketing agency .. we have deadlines .. what is going on . Nano banana pro is not working on any app .. whisk , flow not even gemini app .. the quality is bad !! Its worst than normal banana
im not being able to generate any images of nano banana pro on flow. is it a common issue right now?
Nano Banana Pro is only available in 4k through Google Studio or other 3rd party vendors. Why can't we have access as ultra subscribers to use this functionallity? The output we get right now, if we access it via Flow, is significantly worse compared to regular Nano Banana
@iron matrixIs it possible to use Flow SceneBuilder via an API to use it through N8N automation? I would like to speed up the generation process.
It is impossible to generate anything with Veo, it has said "This model is temporarily unavailable due to high traffic. Please try a different model or try again later." for the last 24+ hours... what a waste of money!
Same, I'm just losing credits
Oh and it won't give me access to any other model.
yeah its big issue
is flow in error????/
yeah contact in private
It's spam / scam. Just report it. Completely irrelevant to this server too.
@modern sedgeHello, 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.
today the same problem, still not working
Hello! No lower priority doesn't mean lower quality. It's the identical Veo 3.1 fast model.
They did get sent out for Google AI Ultra users, yes. Did you sign up through Workspace Ultra? If so, those haven't gone out yet!
Does anyone know the limits on generating images and videos per hour and per minute? Also, how much harder is it for Nanobananapro to work with reference images?
After 2 hours, Nanobananapro completely refuses to use any references. It only allows 1 image per generation WITHOUT a reference to be generated.
for Ultra users,
Been creating videos since morning, everything is working smooth and perfect. Also thanks for getting back the 0 credits ❤️
I'm trying to right now, still not working
Hello just awnted to ask somethign very weird, i tried using THE SAME prompt on gemini for nano banan pro and on flow ( same prompt 100% ) and flow is giving me much more better results then on gemini, how is that possible when they are the same? flow is giving much better visuals and etc rather then on gemini
I've found exactly the same thing happening for me.
Glad i am not the only one, however photo resolution is like 1.5k or less on flow and on gemini its 2k+ resolution, is it the same for you perhaps?
Anybody know about this update..?
its same, cannot generate 😂😂😂😂
Google granted us 2,500 credits to use, considering the very poor quality that VEO 3.1 has been delivering in recent weeks, and nothing is REALLY being done. We often need to generate 10 sets of 4 requests in order to obtain a single usable 8-second scene.
There are several issues: ingredients that are not respected, audio that does not play, audio that is simply terrible, scenes that stray completely from the prompt—no matter how good the prompt is.
As a result, 10×4 = 40 credits for a single 8-second scene.
Considering a video of only 5 minutes, an average of 1,500 credits is spent for a 5-minute video.
One must have a great deal of patience lately to deal with Google.
anyobody get this error when trying nanobanan in flw?
no matter what i type it fails always, anyone else..?
flow working again, so we back on creating ads
Nice and clean work !
how to fix this, can someone here help from moderators or what?
you give us credits back, but how our time back ? time to try another engine
But I am not generating quickly… All of them failed, which is why I requested it again. However, apparently, even when they fail, we still need to wait before requesting the generation again.
why flow didnt working??/
Hello, as a friendly reminder we cannot generate images or videos in the server
Why am i not able to generate images in flow? It keeps giving me errors
What errors are you exactly getting?
Hallo
constantly.
exactly i can make videos through veo 3.1 but no images using nanobanana pro
is there any other place i can use my ultra account credits other than the flow webs and gemini for nanobanana pro?
I love how they are quite happy to take our money and provide nothing for it.

i cant create video too , why ?
////////////
teasing the users
What are your experiences with the low prio flow generations? just tried it the first time... got 2 done and since then 4 failed / blocked attempts
we can cry together
😂
Couldn't generate the image
what about video?
Hey guys! I've made my first Spec AI AD for Google Veo - generated completely at Flow. Pls have a look! #FromAWriter
Presenting completely AI -generated spec ad for #googleveo3
Video Clips 🎥: Google Veo (F2V, I2V) @Google
Voiceover 🎤: Eleven Labs (Speech to Speech Generation) @elevenlabsio
Images 📷: #googlenanobanana 3
BGM 🎹: @suno
Workflow:
Concept + Script ➡️ Me + #gemini
Shot List ➡️ #gemini + Cu...
Another day, another flow nightmare not generating anything for hours straight.
I just came in here to say that it's not generating any Nano Banana Pro images in Flow. Yesterday was hit and miss (mostly miss) but today it's just not giving me anything. Frustrating. I'm only using it in Flow to get 4 images at a time. Is there somewhere else with my paid subscription where I can get 4 at a time? The cost per generation on like Leonardo - I don't want to have to burn credits on Nano Banana Pro when I should be able to use it in Flow.
Is your Flow service disabled for free users as well?
Using Flow gets worse every day.
Wait 2 more weeks, and suno's a goner too..
He keeps saying that to me.
why?
because they caved in to warner. the new model won't be trained on the full catalog. and downloads will be limited.
Did anyone here ever try creating conversations with an Australian accent with VEO? It works for me for a few scenes, but then it starts to just generate the voices with a US accent instead no matter what I do. Does anyone know how I could fix this?
Thank you @main saddle , I knew you were hearing us. (bow)
have you tried giving Veo 3 a reference voice audio clip, and Veo will try to use it as the narrator voice. Upload a short clean voice clip (5–15 sec). No background noise. No music. No reverb. then give it the prompt "Use the attached audio as the narrator voice. Match the tone, accent, pacing, and gender. Use the same voice for the entire video."
Heard you can also feed the audio you know you are going to use and tell it to lipsync to it. "use this audio as the narration track. Keep it exactly as-is. Sync the visuals and pacing to match the audio without altering the voice." But have not tried that.
That's a cool idea, but unfortunately I don't have a female Australian voice at hand right now.😂
Anyone has issues with nano banana pro?
The weird part is that it works perfectly for the first few scenes, but then stops working all of a sudden (happened with 2 projects already).
I keep getting British accents, even if I specify I want New York accents. Does this have anything to do with my character being named Harry? It just gives me Harry Potter lines.
elven labs has some ok aussie female voices. or can always try getting veo to do the voice if it works after an extension might be enough to save the audio and import that into the new generationsyou want it to follow.
Good advice. Thanks!
for me i found a lot of trial and error and watching a lot of yt videos in my spare time lol. Good luck!
I can't seem to upload mp4 or mp3 files as ingredients to Flow. Do you know a way to do that to follow your strat you suggested?
It's been having this problem since yesterday, and besides having paid R$129.00, it's using up all my credits without generating any videos! 😡
download each clip and resync them i your favorit editor, you have to back up about a second on the end of the first clip, then they matchup perfectly no seems
@analog radish Admin, I need help! I subscribed the day before yesterday and haven't been able to generate any videos since then! Besides the app not working, it's using up all my credits! I need to generate a video tonight, I urgently need help!
wasn't there a way to add audio through google ai studio or something? cant seem to find it, swear there was a way to do that before. wondering if it mightve been one of the 3rd party platforms?
No clue, tbh.
Google movies is an editor, you can even generate Veo clips there
sorry cant seem to figure it out, my bad for taking you down a rabbit hole lol.
No worries! I appreciate you trying to help!🫶
a browser editor? if so, could you post a link here please?
Google Vids is an AI-powered video creation and editing tool. Use it to generate Veo clips, record your screen, convert your presentations to videos, and more. Access Vids with your Workspace account or your personal Google account.
thank you. i expected you to post a link to google docs. the think is. somehow not everybody is able to use veo 3 there. for me its not working. for some people it seems to be.
WHY THIS ALWAYS HAPPENED!!
The videos page is definitely a browser only thing, you can only view created media on mobile
not sure what you are trying to tell me. you said you are using it for veo 3 creations but you watch it on mobile?
and as i said, its not possible (for me and some others) to use veo3 there.
did it get completely outsourced to the app?
You can use Veo3 there, but to create stuff it has to be done on desktop.
why google AI Ultra for bussiness not available lol
This what you will see if you load the page on mobile.
im on the pc. no veo3 available. thats why im interested in this topic. a few months ago i was creating my videos there. nowadays its not available anymore.
Hmm. Well there is still quite a bit of demand and I haven't tried making anything for months on end via Veo. Even if I did I would've done it through the Videos page and/or Gemini first to test out how my prompting is structured.
Can someone help me? My Flow app is saying it's deactivated for free plan users due to high demand, but my friend is using Flow without any problems.
I don't believe so? I just signed up for Ultra through my regular gmail directly from the Flow site. It's the "Google AI Ultra for Business".
why is this??? happened to me 20 times
thats the workspace ultra is it
How? Are we pro users on the low priority? Doesn't make sense, I'm paying for this
Also, is there any time where it is less saturated from users generating and it actually works? It's been 2 days since I could make a video
How do you do it? It only lets me create 4 videos, and I have to wait a long time for the next 4.
ive burned through a few thousand credits today and having the same issue. doing the typical incognito/browser/cache ect doesnt seem to do anything. going from low priority to fast or even quality doesn't seem to make a diff. I was able to get it to work for about 10 min earlier with sketchy hallucination audio for voice before or after the line they should say.
Hi @frank stratus , yeah that's why it's labeled [Lower Priority]. Since we have high demand, we need to offer the free model at a lower rate (to Ultra users). We really want to provide users with a 0-credit model, but need to do it in a measured way that doesn't disrupt service to everyone. Hopefully that helps
What about us Pro Users? We are consuming credits and it's eating them, I haven't been able to create a video since yesterday morning
There is absolutely no difference 🤣
Except the fact that we can create a tenth of the videos we previously did.
Nothing changed the the AI itself
it's a strange solution when you could've fixed the demand by adding a queue system to generations and forcing all ultra users to share the same permanently-spun-up TPU pool. y'know, like how sora turbo handled it. the video would queue when the system was actually ready. no failed gens, no confused users pissed over being charged... just "you might wait 5 minutes before your videos start generating"
What kind of errors? I rarely had errors before. And now it's just ignoring my request. Is not generating an error?
It's not even a queue. It's instantly rejecting, not putting you in a queue.
ultra user here and 130 generations today and only 12 had audio. just keep getting "Audio generation failed. Please try a new generation." over and over, it's extremely frustrating .
aye, I've been experiencing it too, this kind of thing only happens if their backend is legit OOMing, their architecture is not prepared for low tpu availability
Still Ultra are getting unlimited creations while we are getting 0
So then why am I paying if I can't even use the product I paid for
Also, it's been 24h and the credits I used yesterday are not refunded yet
Apparently free tier user aren’t able to generate anything?
Credit instant refund if you get failed generation
Try check here @worthy sparrow https://one.google.com/ai/activity
I got in contact with customer support and they told me it would take 24h. I used like 200 credits yesterday, all of them were failed creations and I still have the same credits as yesterday,
Was the video shows failed generation in the project page
Yes, it gave the error and the prompt was under it
Huh you should be able to get instant refund
As you can see they were not refunded, nothing changed
I don't have any new creation since the 24th
Uhhh I wanted to ping one of the flow staff here but idk who to ping about this
Maybe @inland pier ? He was talking here before
Think big text problem for the moment
OMG - the failure rate! its so annoying and breaks the productivity flow like crazy, this is unusable, and I'm spending 100 credits for every generation and it still gives me a timeout saying I'm generating too fast, when I do generate, it fails with a failure rate of 90%
any news on Nanobanana pro 2k/4k for ultra users? will it get implemented soon?
Fun new trick with Nano Banana Pro and Ingredients to Video
working on it
Perfect, thank you
They keeps saying that to me. 😭 Any idea when this will be fixed?
so now fast is not showing up
No video I'm making is generating any audio. Is it my computer or the program?
well the fast low priority one i mean
is flow working?
It generates the video I requested, but it doesn't add the voiceover.
Do you have a paid version or free version?
Ultra version
Sometimes it doesn't need to make about 5 Sometimes before voice
😩
Make sure you have quotes " for narration "
I´ll try. thx
It works
It works!! Thanksss! xo
Biggest thing im hyped for next flow feature update
I was in the middle of generating things and my free tier for fast generations just vanished...anyone else just get this too?
ultra account btw
Uh oh - looking into this right now!
just kill free users. restore 0 credits for ultra.
I'm on nearly 3 days since I could last generate a video. I'm paying to not be able to use this.
It was strange, I've been doing a handful of the free generations this morning/afternoon and then got repeated fails so I switched to the 10 credits for one generation and when I tried going back to low priority the option vanished
same here
Thank you for the prompt response
yeah, they use up speed, never subscribe and cost the paid users performance and stuff like free uses
basically the paid users finance the leeches
was working earlier with the fast (low priority)
I feel like he's obsessed with the instructions I gave yesterday, which have nothing to do with the new project.
Of course. The fix should be live for everyone in about 30 mins. I'm so sorry again for this hassle.
Many of you should have the zero credit model back now
yep thanks!
I can finally generate another video... only been down for close to 3 days for me.
even veo 2 uses up server performance
Agreed
Does anyone know if there are any plans to support “ingredients-to-video” in portrait mode?
anyone know whats the daily cap for ultra plans in using nanobanana pro? is it normal for it to stop working after ≈30 minutes?
Man these servers are failing left and right, fails to generate like 50% of the time, and fails to generate audio like 50% of the time. So hard to get a useable result.
are you using T2V or I2V, for me i could not get audio with T2V today, should be the opposite i know. so taking a still from the clip then doing ingredients to video with the still uploaded seems to get me audio more times then not.
I'm doing I2V with ingredients.
still on the 10 wt ultra, will that be rolled out to more users in the next week or so before i run out of credits lol
Just did like 20 generations, and not a single one of the generated the audio.
try t2v see if that works. ive been uber frustrated last day or so with the audio. about to go back to how things were done before beta audio myself.
ha was just using this
ok question settting with 9:16... my images are coming out at 40-44 PPI...is that bad?
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. we have #1319424187730755636 for feedback specific questions/ queries
with pro subscription can you generate video for free in low priority mode?
Can we have some transparency on what the cap is for Nanobanana Pro image generation on Ultra plans?
USA Police Woman Starts Dancing in Public — Crowd Reaction Goes Wild! (2025 Viral Video)
https://youtu.be/FPbD1j-JT6s
USA Police Woman Starts Dancing in Public — Crowd Reaction Goes Wild! (2025 Viral Video)
USA Police Viral Video 2025!
A woman police officer starts dancing in public and the entire crowd gives unbelievable reactions.
This wholesome and funny USA police moment took over the internet and surprised everyone watching.
If you love public reaction v...
Flow is strait up broken right now out of the last 50 generations, only 3 has audio.
feels like they reconsidered giving free fast generations + 25000 credits each month to ultra subscribers was not the optimal strategy, out of nowhere (regardless on NB2) flow is under "heavy usage". When this happened the last couple of months, the team fixed it quickly, but now it seems as if they are trying to lead us towards a different solution, instead of fixing the real problem
does anyone facing this ? Before I could easily generate videos, now i have to wait to generate next round. this is a bummer. other AI tool I paid won't be like this
especially you paid for Ultra for Veo
Yes i got it yesterday, changed to paid fast generations, its gone
How is it that as an Ultra subscriber, my generations fail because there are too many generations and other errors? Isn't Ultra the top tier? Why couldn't you take care of your subscribers?
I'm just exhausted, frustrated, and angry. How much longer is this going to go on?
Yes, there are - landing very soon.
Only Ultra plans have the zero credit lower priority model.
It's about 500 per day right now, costing zero credits.
100% of Ultra users should see the zero credit, lower priority model available in the model drop down. Please DM me if you aren't seeing this!
I'm so sorry for the hassle. Are there any patterns on when this is happening to you?
so i am in austria. i understand that american afternoons and evenings are probably high demand time. now it's 1 pm here and i guess around 5~6 am in the US, where i assume the majority of users are? which means it somehow was high demand time in both time windows, because right now i get the "Requesting too quickly" when trying 2 gens at the same time
or is it like i suspect, that the free ultra gens are just set to ALWAYS count as high demand time
Hello, has anyone figured out how to download multiple clips from a project? At the moment I download them individually one by one
MFW i get blocked results for probably wanting a corpse (not even any graphic or nude parts included) on a morgue table, but then the AI just decides it generates a scalpel going into the skin, which i did not prompt for
which is exactly the craziness of censorship...i just need basically an actor to lie still, but ai thinks i want sexy stuff, but then does something totally unhinged by itself
Create flow
I just prompted "A mortician prepares tools to the side"
Guys, is anyone of you experiencing serious audio generation failures? We've tried at least 20 generations and it simply doesn't work. Tried adjusting the prompt an infinite amount of times but it does not work, the audio generation keeps failing, do you have this same problem?
audio no, general failures, yes
i know the generations without audio, get them from time to time, but right now i either get video with sound or no video at all
You should be able to request 4 gens per minute right now on the zero credit model - is that not working for you?
You can download multiple clips if they're in a scene on Scenebuilder
Cool thanks. Mine are individual clips not in scene builder, so I guess I have to download them one by one? Just asking in case I'm missing something
Is this the same for VEO 3.1 Fast on 10 credit model?
No, you should be seeing higher requests per minute going through on the 10 credit model.
Yes, that's right. We're working on the ability to take action on multiple items at once though!
so recently google nano banana pro was realeased, and i have the ultra plan, so i used both flow and gemeni to generate 4k images as it was annouced about the model, but all i get in flow is 1k, and gemeni gives sometimes 2k ( on thinking mode not fast) so how am i supposed to generate 4k image and i already have the highest plan around
can i know when the request per minute count? after click generate or after generated done ?
@tough yacht are you sure per minute ? i already try every minutes with fast or lower priority ? still same issue
lol, its already 10 minutes,
still can't generate
hi every1 can any1 tell me if movie flow is safe to use my pc sais malware
could you hard refresh the page as well?
and then try generating. Elias's #'s are correct!
nope, still can't
i have created some videos and they have disspaeared...i created like 12 generations and can only find around 7
where are the rest? Is tis a bug?
after 15 minutes, now can. 
! that is not great... is it happening repeatedly?
let us know if you see the hard rate limiting occuring again!
first time today...looking to drop sme on the scenebuilder and they are not there
some where good too
scenebuilder can be a bit delicate... you might need to re-create things there. We are working hard to make it a better experience though!!
I've came here precisely to ask around about this. I've been trying to generate videos in portuguese language specifically from portugal and it had worked fine until yesterday around 3pm. After that, every single prompt generates videos in brazilian portuguese. Considering that during the last 3 days i've consistently generated videos with audio in portuguese from portugal, to then stop abruptly and out of nowhere, without any explanation... And to waste so many credits trying to make it work... Really frustrating
I wasted more than 100 credits. No luck. No portuguese audio at all for around 24 hours now
are you getting videos without audio, or videos with audio that is just incorrect?
we have been messing with the audio side of things, Please flag the videos in Flow. That will help us know if a change we made was bad!
so, for context, im trying to make a bigger video, with character consistency, and audio in pt-pt. But since like 24h ago, the audio has come ALWAYS in pt-br, despite using the same prompt, and trying different prompts like telling it specifically to NOT use brazilian portuguese. Every time, always in brazilian. It looks like something specifically connected to portuguese from portugal
I flagged some of them, and have been flagging them since it started
still same, need to wait above 10 minutes to create 1x4 generate videoes
are you sure you are using the 10 credit model?
yeah its same, fast 10 credit or low priority still above 10 minutes.
fast 10 credit should not be limited in the same way...
we might have a bug on our hands... I'll try to repro in a bit
But its thankgiving so a fix might have to wait until monday (as we are not allowed to make any changes until then) I'm very sorry for the issues Bang!
no worries, thank you, i appreciate it
@iron matrix I wasted 500 credits trying to make it work. Do you know if I could get at least some of them back? Also, when do you suggest I try again? At this point, I'm always afraid of trying and waste more money
it’s 3 days flow have too much problem, some fix when?
Hey guys, do you have an ETA for Nano Banana Pro 4k images?
is it me or everything okay with flow today?
Anyone have a trial code / invite they can share?
no, the singing is back, 🤣 😭
please fix that, I think that was the most annoying bug
thats not the worse one lol
why does this happen? i was able to do 5 generations now i can do only one with ultra plan
i have this since 2 days now
Yes , it been like that for 2 day, only can create clip every 5-10 minutes
Fast generations return the "couldn't generate video. Try again later" bug constantly, like 80-90% of the time.
Hi all! Does as anyone know if it’s possible to select resolution when generating images with flow ? I’d love to generate in 4K but don’t see the option.
This chat has become seriously depressing - can we add a Moan Zone channel and keep this one general chat?
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. we have #1319424187730755636 for feedback specific questions/ queries
Agree, it’s been a wild ride these days, but it’s so nice to be able to chat directly with the developers and know they care and are working on fixing the issues. Thanks a lot, and please don’t remove the 0 credits again 🙂
Can anyone check if swedish language work?
Not trying to be a jerk but why do you need to spam 20 concurrent gens
why not ? problem ?
It used to be normal to make longer videos to speed up our workflow, but now it’s just frustrating to wait so long. And in the end, the videos don’t always turn out how you want anyway.
Well in my opinion iterative prompt optimization vs roulette is faster
veo 3.1 has been making much lower quality videos for me for almost 2 weeks now. It doesnt follow the prompts correctly and also makes worse videos with way more mistakes. Is this just me or is everyone else dealing with this? Does anyone have a solution? does using quality help?
All days only can create 3 videos
There been a hit to quality yes
is it confirmed or like have google acknowledged it in anyway?
I’m assuming because they’re testing trying load balance and they confirmed they’re working on how much compute allocation each gen gets
And yes I’ve seen a hit to quality I use this for work too and we have an API to veo and the gens are different quality where two weeks go fast was really close to the API which is almost as good as quality
oh so they actually said something like this "working on how much compute allocation each gen gets" ...man that disappointing to here. The lack of computing power is really showing.
Yeah something to that effect
hopefully they got those new data centers built soon 😅
@timid hollow Hello, we wanted to remind you of our server guidelines, stated in #start-here・📍 rt-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.
OMG the quality of veo fast has dropped so much lately
😩
Is it possible to pay for more when we run out? It asks me to upgrade but Ultra was the highest tier.
Anyone else having this issue?
Nano Banana Pro won't use my credits even though I still have almost 10k credits. It keeps showing an error and I have to wait a few minutes before I can continue.
why can't generete videos more than 2 and also have to wait 20-40 seconds to upscale them in 1080p if we do constently it gives error, then what is the benefit of having a ultra account
Is the limited free ultra veo3.1 fast in any other way debilitated besides from generation limit?
Either that, or the content filter is a lot worse these last 2-3 days. Prompts that worked before with 4 succesful generations are now failing or giving me 1 or 2 results only
Is there any admin here I need help?
Why do images created in the untral account not have high resolution? Previously, I created images with size > 1mb, now they are only over 100kb.
if this was not backed by google, i would expect an announcement that they overestimated server cost and demand and that they go bancrupt and have to stop their service, cause it looks like everything is bursting at the seams and constantly worsening
Hey guys, is there a problem right now with Flow? i can't play any generated videos
I dont think anything looks like its getting addressed, still waiting for them to fix the ingredients to video with a quality model after they took it away from veo 2 quality
veo 3.1 fast seems to be the only one we can use and the output its so sub par
I've been trying to play my generated video, but still not playing. What about you?
The two Veo 3.1 fast models in Ultra are both the same underlying models. Can you DM me examples of prompts that are failing for you? So sorry for the hassle.
sorry no. before you know it, my prompts are getting used to tighten the already overdone content moderation even more, so i will not share any of my prompts
I wouldn't have used them that way, but no worries! Sorry again for the issue you're seeing.
I make tiktok videos with ai, I have 13 scenes and I make 3 generations in one time so it’s faster
Hey! I can't reproduce this problem. Is it happening to you 100% of the time, or only on certain videos?
You should be able to request 4 videos per minute with the zero credit model and more with the 10 credit model. Is that not what you're getting?
We are working on that (enabling users to pay for more), but it's not ready yet, apologies!
@tough yacht any idea why we have no ingredients to video for veo 2 quality? would love to get an update on why that was taken away?
I believe it was to free up capacity for Veo 3, including Veo 3 ingredients to video.
what do you mean by capacity? I'm totally happy paying the 100 credits it cost per gen, it was the main model i used and there seems to be no indication of when ingredients to video for 3.1 quality is coming to veo. so i've ended up with half finished projects because the other models dont gen anything remotely close
Has anyone got a good prompt i can add to my prompt to keep character facial consistency as soon as my character moves or the camera moves the face changes any help would be appreciated 😀
I thought the fast model that costs 10 credits was supposed to work during high traffic, but the generations still fail 🫠
Is there a package where video generation is unlimited?
Hello Nico, I've sent you a direct message to follow up on this opportunity. Looking forward to hearing from you!
to me it feels like that since the upgrade from 3.0 to 3.1
Im having so much trouble generating the simplest of scenes. Im providing a first frame of a male in a car. All he has to do is turn the key and drive away, but he keeps doing all kinds of random crap like opening dorrs and driving backwards
any help?
Every image I upload gives me an error "against community guidelines", anyone knows a fix?
Are any of you experiencing increased censorship on image generation? Im getting too many errors on Nano Banana Pro while yesterday I didnt
yeah same same
Is there a daily limit to banana pro generations on flow for ultra users?
I'm hoping that's a bug
Im getting the stupidest erros...a car that is supposed to be driving forward drives backwards with the driver in the passnger seta, a ficed camera turnes into a crane....
The most simple of scenes has taken me around 1k credits and I sill dont have it
meh...
I'm getting SD resolution 720p on image output using nano banana pro, I am pro subscriber. Anyone getting better resolution in flow -> create image?
I believe that's the standard resolution for everyone. They're planning on adding higher resolutions eventually, whenever that will be.
Is there a way to get 1:1 ratio on flow banana pro? currently only 16:9 and 9:16 visible for me in settings
actually, what is the different of flow and vertex? which one is better and what is for what is
Ultra tier has the highest daily limit, but yes there is one, based on traffic/load. We're hoping to raise it in the next few days.
Sounds good. Just wanted to make sure it wasn't another bug or issue. Thank you for the response.
Thats simply not well structured prompt.
[Title]
Rear View – Truck Drives Off (Driver Only, Fixed Tripod)
[Start Frame]
Static rear shot of the red pickup beside a Spanish-style villa on a dirt road.
[Camera]
Fixed tripod on the ground. No pan, no tilt, no zoom, no crane. Single continuous shot.
[Driver / Occupancy]
- Driver: Bill is seated in the DRIVER seat and operates the vehicle.
- Visible cue: driver silhouette/head movement through rear window; subtle steering wheel motion.
- PASSENGERS: NONE. Front passenger seat EMPTY. No one in rear seats.
[Action]
- Engine cranks and starts.
- Brief idle beat.
- Truck pulls forward and drives straight away down the dirt road, receding from frame.
[Audio]
No music. Production sound only: starter, engine idle/rev, tire-on-dirt, ambient insects/wind.
[Look]
Natural daylight; slight dust kick-up as it departs.
[Negative Prompt]
empty cab, passengers visible, autonomous vehicle, camera movement, text overlays, watermark
thats just one of the 10-12 variations
Yep, really actually
can you help...what do I have to prompt
meaning...I clearly state a driver in the drivers seat...no passengers
here is another:
[Title]
Rear View – Truck Starts and Drives Off (Driver: Bill)
[Start Frame]
Static rear shot of the red pickup parked beside a Spanish-style villa on a dirt road (use the provided image as the exact first frame).
[Driver]
Bill is driving (inside cab, unseen from this angle).
[Camera]
Single continuous shot. Begin locked at rear-center. As the truck departs, perform a gentle crane-up (2–3 m) with a small pull-back to reveal more of the road. Keep movement smooth.
[Action]
- Engine cranks and starts.
- Brief idle with two natural revs.
- Truck pulls forward and drives away down the dirt road.
[Timing]
7–9 s total:
- 0–2 s: crank + start
- 2–4 s: light revs
- 4–9 s: pull-out and departure
[Audio]
No music. Use production sound: starter, engine, revs, tire-on-dirt, ambient insects/wind.
[Look]
Natural midday sunlight; light dust kick-up as it rolls away.
where do I need to make changes?
Im getting everything correctly excpet the driver seat empty, or 2 passengers...you really think its the prompt?
Few things i would do in this particular case: split the main scene(the 8sec one) into microscenes. First one-the care to ignition camera movement. The ignition itself. You can either use start-end frame for the first one, or use extend option(extend one requires a couple of generations till u hit the result you want. Than is the starting phase. Skipping this process gives ai too much freedom and hallucination potential. Thats why the result is far away from what you desire.
well...i simplyfied and im just starting with a first frame of the bumper and the truck driving away...i keep getting passengers
ive tried only the ignition, but he ket opening doors
Send me the starting frame, let me try to fix that for you, will provide u the prompt later
Action – Exactly This]
- Bill is already seated inside, driver door CLOSED and latched.
- He turns the key; starter cranks, engine catches.
- One beat of idle (~1 s).
- He pulls forward and drives away down the dirt road, leaving the villa behind. No reversing, no stopping, no extra gestures.
Im not asking for much
The problem is not about how much you ask, but how you ask it. You can make 10 scenes in a row with super smooth, consistent movement which replicates this action, or you can make 1 scene with 4 microscenes in it.
All i need is for this dude to start the car and drive away
My personal take-use json prompts. They highly reduce the chance of hallucinations and strictly fallow the rules you establish
ill give it a test
one sec, let me show u my take
So when do we get 2k Nano Banana Pro on Flow, like we do on Higgsfield?
Or this take...with a fixed camera...he drives off..but theres always someone in the passenger seat, or no one driving
give me 5min, creating the prompt for it rn
testing JSON...ive simplified my scenes and would just be happy with the pickup driving away on a dirt road starting with the bumper
Again! Tha mf!
is it so hard ...the prompt says no passngers..just driver
thing is ALL generations come with this man
much cleaner yes
so i want to learn this....what did i prompt wrong...did i give it too much info?
no, actually you gave too low info. sec ill share the prompt here
(the one i used)https://docs.google.com/document/d/1DFRqJ_nKkv8NSdoj3SW4ACCpSTLB0BxbanUkkyifznw/edit?usp=sharing
mmm...i see...need my Got to step up iots game then
The only useless part of this prompt(the one i provided), is the negative prompts, flow simply does not fallow them strictly.
you are welcome
mind if i feed that structure to my gPt so it can sleep on it?
this structure was designed to suit this image., however i dont mind at all, its simply called Json prompting. It basicly provide strict rules, timing and detailed explanation of every single detail which you would like to happen into a video. This prompt was like actually a demo version(a good structured one can have everything with specific SFX, music, VO lines, exact timing), basically everything you can imagine of. This way you can reduce the hallucination to minimum. Does not mean it always works, Ai sometimes still ignores some of the rules(happens to me quite often when trying to replicate super complex motions or dialogues), but for general use its quite good. For complex scenes you need to do micro cuts and adjust the sequance based on the part you like vs the part you dislike.
no problem
JSON is key, agree
yep, much slower but times better in terms of results
Absolutely
JSON is like the coding in a program. A set of instructions.
Never send your wife out to get Christmas lights! 280$ she just called me so I wouldn't be shocked when she gets home.
JSON offers no benefit over natural language prompting. Have tested it extensively.
Np, good luck
I have tested it aswell, prolly couple of thousands of times. Even if you perefctly structure a holistic prompt and include all the details as of a json, ai hallucinates much more.
quality is very bad now from flow veo .. more unreliable than ever , anyone else going to jump ship to another AI?..if so where you going?
how can you make this kind of detail? that's crazy. GG
There is literally no difference in result between a JSON and natural language prompt, other than the normal differences that you get between generations.
I feel like VEO 4 is getting ready to launch, which is one of the reasons why quality has dropped lately.
Can't just be Nano Banana Pro messing up the servers
Does it difficulty to have a selection for background music?
The background music that generated by veo3 is so suck.
so quality of veo 3.1 still bad today?
I heard mixed opinions about this, thats why I stopped using it...will have to test on my own and see
Hey folks
Fast has been better for me idk why
The main reason(atleast trough my pov) for background music to suck is that in most cases it either conflichts with VO or with SFX. Tried to do a fine polish about it, barely works.
Maybe Veo 4 is around the corner wonder what that will bring lol you know Google love to stay ahead
rly looking forward for veo4 myself aswell. Rly need that longer video versions:)
Well in film we don’t really dialogue for more than 7 seconds until it switches anyway
rly depends on the type of content you create.
Yeah, I always have to tell Veo "no background music" , but sometimes it still generates.
In most cases, if the video is with sound it does
Same. "No music." Works 90% of the time.
Or "No dialogues"
anyone has this problem where veo 3.1 persistently shows tears with blood? my prompt just said tears... but it keeps doing this. even in previous projects
Good morning to the Veo community btw. Look at us. Aren't we all beautiful 
Yes that happened to me. I usually fix it by not prompting tears explicitly but only things like: sad, sadness, emotionally hurt, etc..
weird. i tried drooling and it came out fine lol. sadly its not what im going for
thank you
with women its easier because i would write mascara washes out and stains her cheeks
so that covers it
True. I noticed sometimes I get stuck in being too explicit with the description of actions. But letting Veo interpret emotions or cues work sometimes (!) better.
I start with the still showing tears in nano banana, but then only use things for Veo to interpret: "man sits lonely, emotionally hurt" and not describe the tears again.
He's not having a happy time it seems. Everything okay with bro?
An experiment i ran a few days ago was the negative prompts into main json block. They are basicly useless. AI prioritizes fallowing the logic of main prompt, almost always completely ignoring negatives.
tried 2 different approaches-putting the negatives at start of the prompt and at the end. Absolutely no difference
You will laugh if you know what happens in this project. I’m setting the mood to the lowest of lows then crafting an abrupt transition abruptly to the complete opposite.
Correctly write about the tears prompt as you are giving some wrong details to VEO
good to know, gemini gave me a video prompt (yeah i know it does not REALLY know veo) with negative prompt, but i thought it was not really supported, but didnt experiment much
When you create a prompt via AI, specially when using gemini 3.0 pro, always mention the creation model you will use. I experimented on that aswell. If you use gemini for prompt creation, it provides better version for it than GPT. However, if you use GPT for sora, gpm performs better:). Nowdays, im using GPT for prompt adjustment, simply because i use templates for my creations. Really depends how you will adept your workflow and the goal of the creations.
wow! cool
was an experiment for a YT channel, did not worked, hook failed dramaticly:D
Ok but respect the car's private parts please
❤️ will do:D
Looks great man
Share the prompt 🙂 #flow-general・🌊
i made this a month ago, cba searching for it
The best click bait. People love to correct others 🤓
The exact reason:D
And find "mistakes"
Whats up with this failed generations ?? I cant even generate with veo 3.1 high quality .. keep hitting time limit .. when im paayin 100 credit for each generation i deserve to be pass through and not get walled . Omh
After how many requests are you hitting the limit? You're using the Veo 3.1 quality model, right?
Yes im using high quality or fast and keep getting error u are requesting too many times .. even when there is at least 1 or 2 min between each request i send .. i guess this error should only be sent to those using low prio free mode .. not the ones paying credit .. whats the point of credits then ?
.
This does not sound like it's working as intended. I'm going to DM you some questions to see if I can figure out what's happening
Ok
Again?
please can all be reminded of our first rule: Be compassionate and treat others with kindness and respect. we have #1319424187730755636 for feedback specific questions/ queries
I would like to flag a few bugs that I noticed on flow:
- It’s a bit missy to view the clip in the scenebuilder if I have trimmed some part. When I play the trimmed clip on the scenebuilder, the trimmed part is not shown but like 1 second from the original clip’s end is shown in the end even if the clip is trimmed. It sometimes makes it confusing while editing and knowing if the part is still there.
- The scenebuilder and a random number of past generated videos sometimes disappear after I reload or enter the project after sometimes. It happened to me twice. I had to rebuild the scenebuilder.
Feature requests:
- I wish I could paste the image directly in the prompt box by CTRL/CMD V or just pasting. Sometimes it’s really convinient if I want to put an image from google into the flow prompt box for ingredients/starting-end frames rather than saving it on the computer and then import it using the plus icon.
- Generating images in 1:1 ratio in the image generation section.
- Accepting ingredients in any aspect ratio. As the ingredient directly plays no role in the video so maybe it could just be uploaded directly in any aspect ratio as the model is only gonna put the ingredient in a scene rather than generating the whole video from that scene. Or maybe a wrapper which would automatically crop the image in the required ratio by a filler transparent background in the absent part.
HI there, please make sure to put this in #1319424187730755636. Things like this can get lost in a general chat channel, Thanks!
ohh sure thing!
Hey
I have problem with subscription I want to upgrade my flow
How to upgrade my subscription ?
we need 2k 4k quality。😋 😋 😋 😋
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!
These are all great suggestions. There are definitely some paper cuts with Scenebuilder, which we are working on. Copying/dragging images in is a frequent ask, so we can prioritize that. Not sure if 1:1 aspect ratio is on our roadmap at the moment. Our video models only accept references of specific aspect ratios, but we can consider deferring the crop region to the time of generation (apply a non-destructive crop on an image during generation rather than doing a destructive crop at image import)
Are you referring to image, video, or both? 🙂
thanks! :))
"To clarify, what I urgently need is 2K and 4K support for Banana Pro images. I suggest requiring extra points/credits for high-res generation to manage server load/resources. Video isn't as urgent right now, but it would be a nice bonus if added."
Ok great to hear. We are working on high res NBP so stay tuned.
That's awesome news! Really looking forward to it. Thanks for the update
Anyone else getting failed generations...my last 7 attempts have been rejected
Change your prompt, sometimes something trigger it
been rying different...also switched from ingredients to frames
any ETA on this? So i stop my subscription on third party providers?
May i ask if you could add more aspect ratios than 9:16 and 16:9 specifically for image generation?
NBP high resolution
So need 2/4 K and more aspects for NBP in Flow...
Can't promise exact dates, but we should have something in <2 weeks
What aspect ratio(s) would you all like (besides 9:16 and 16:9)? Also, what are your typical use cases?
Social media so i use, 1:1, 3:4, 4:5 and sometimes 4:3
Upgrade to what? Ultra plus? for 5k usd/month?
Good to hear. I'll bring it up with the team and see what it would take to get this supported.
9:16, 16:9, 1:1. For me personally, those are all 3 im using - ad creatives, YT content(short and long). Something i experianced was when using a refference image from whisk and try to use it as first/last frame on flow, their aspect ratio are different(the one from whisk is not filling the flow aspect box properly). Thats for 9:16.
Thank you!
Hey everyone, I build AI tools for entrepreneurs & marketers.
I’m dropping a free Lite version today — it gives an instant marketing diagnosis.
Posting it just after.
100% free, 100% optional.
Hello @stray topaz 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!
Hey everyone, is anyone else struggling with VEO 3 or Nano Banana blocking completely safe prompts involving children?
I’m working on a short film about fatherly love that needs some childhood flashbacks. Since yesterday, every time I try to generate the main character as a young girl, I get hit with the "Policy of underage children" error.
It’s strictly cinematic/sentimental stuff. Has anyone found a specific keyword or prompt hack to get these safe shots through without triggering the filter?
Are you seeing a black border?
Yep, exactly, on both right and left side
Yeah this is a bug on our end. Unfortunately all the image models are at slightly different resolution, and we need to do a better job at cropping/resizing before passing to the video model (for Frames-to-Video). Vision models love dimensions that are multiples of 16, which causes conflict with 9:16 and 16:9 aspect ratios.
Got it, thanks. Its not biggie, just when doing mass images, i prefer using whisk for faster creation. But NBP as it now properly works does the job good enough.
When I design a photo using Flow, the quality is very low. It comes in at 311x553 pixels. Is there anything I can do to make it 1080p or higher quality?
This is a known bug. I believe if you download from your asset view (Images tab), the resolution will be high. It’s only if you download from the image editor view where they come out lower res. Fix coming this week.
Let me know if you are seeing a different issue.
thank you
Hello
Is veo 3 still making lower quality videos?
kind of yeah !
Bummer 😢 The type of videos I make have been massively affected by this and recent downgrade in quality.
The service is basically unusable this past week, what a shame... I’m glad I’d subscribed to some competitors so my business isn’t completely stalled, but I definitely won’t be recommending Veo 3 to anyone anymore
I NEED google team member on this PLEASE EXPLAIN TO ME how this!! violates policy!!!!!!!! Logan is on his knees, claws clenched, peeling the black symbiote off his chest in violent, ripping motions. As the goo stretches and snaps away from his skin, Logan unleashes an extremely loud, deep, demonic, raspy scream — a roar of pain and fury that echoes through the entire forest, bouncing off the trees and shaking the silence around him.
this is beyond stupid and needs to be fixed!!
it is beyond stupid on how pathetic veo3 has become!!
veo3 has become paper thin bro!!
Off the bat I see a few blocks, on his knees, peeling back, violent ripping motion, snaps from his skin, demonic, scream, pain, and fury
Change that and it should work, use chatgbt to find solutions
Understand pixverse as well is a goto
at least kling isnt a chump
Understand, but Kling is China based and Google US
but yet i can get carnage literlly ripping mary janes guts out blood splatter all over the place in veo3
I know
Kling have safeguards too, pixverse is better for not getting flagged
I know, funny.. but what it produces you can't frame back. It made it
Lol
Nice, I get flagged in Kling more than Pixverse
crazy what do u type?
All kinds of things especially Billy the Kid.. shootings I get flagged a lot
oh u can post shooting into veo3 that dont get flagged lol
Well, it's hard to blast a person 10 feet back from a gun shot
lol
i know i have them as well but its mad annoying
same i dont
like we are just out here trying to create way better universes then hollywood isnt that the whole idea of all this ai
copyright violation maybe?
Does the "Favorite" feature work for most people? I can fave Videos but not images i.e. I can press the button but it doesn't do anything (it doesn't actually add the image to my faves) and when I navigate away and back again it has been removed from faves
try stationary locked shot of...
I refreshed the page and it is now working, emergency over, everyone stand down
Try stray jacketed insivably confined
Or severe case of rumitisum lol
Try Hedra dot com for that. It works
Up to 5 minutes a video lip sync and it's good, only problem not much else is going on in background, solution? Use Hedra for zoom in commentary A roll and veo for B roll. Action!
what is the best way to use Veo3? by Flow or api?
Flow! labs.go
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
Flow is the best aigc web app that i have ever seen.
The reason for blocking is usage of "logan"-its trademark protected. Try using a lookalike character, which you strictly describes logan. This way you will bypass the filters and trademark infridgement(in case you get strike on YT for example), while you will still recreate the idea.
best way of doing so is-go to GPT and tell him that you need a super detailed character description for a character, which looks alike logan, but you are creating a video and you want to use a lookalike figure, safe for video generation usage, without triggering safety filters and not violating any TOS or trademark infridgement.
still failing generations please fix this man \
Yep, all the time. For me, it's completely unusable this past week. Paying $250 to get less than what other services are offering with their free trials is just a disgrace. But what's even more disgraceful is the little to no communication regarding the issue.
if ur trying to generate more than 4 videos simulatiously with the free generation you get this message. on 10 credits option, you can generate up to 20 simulatiously(5x4 at a time). But the free one - 4 per minute. Works for me perfectly.
Hey gyys , " couldnt generate image try again later" all the time ! I cant generate anything since yesterday , does it show this error when it triggers a moderation bot or something ? or is it "just" a server issue ? cuz the service is barely usable
try to hard refresh the page (CTRL+F5 on windows), or try on incognito/edge/any other different browser. Seems like a caching issue for me.
i think it's 4 images per minute
its 4 videos per minute, personally im not hitting image generation cap. However, afaik there are 100 generations per day for NBP per day. so you might hitting this error
Oh sorry I thought you were generating images
now it says ive reach my limit 😅 Thank you
❤️
not 100% sure, but think the caps are: 3 free gens for free plan, 20 for pro, 100 for ultra.
thats for NBP, normal nano banana got no daily limits
im on pro but im using nano banana mostly , 20 gen ? naa I think thats more it would be really expensive
as far as i know(not 100% sure, barely generating more than 20 images a day on ultra, never hitted cap), only nano banana pro got hard caps, normal one doesnt
if you have hitted the pro cap, you can try the normal one. Or if you hitted the normal one, try with pro. Im actually curious about that:D
i generated more than 50 images with NBP (in the same day) on flow so its more than 20 for sure
(with a pro plan)
As i said, not 100% sure. MIght be 100 for pro 1k for ultra. Never hitted a hard cap, but im simply using first frake image prompt in my video prompts:D this way i dont need to generate first frame image unless i need to generate continious scenes.
im spamming NBP pro since 2 days but i feel limited unfortunatly cant afford ultra and regular nano banana dont do the trick for me.
difference between NB and NBP is massive indeed.
i wish I could knw when my cap reset lol i lost tracks haha
you can easily track it. Go to your last generations and see when u made ur first image.
my last ? but im doing images since 2 or 3 days
i cant even upscale to 1080 to download XDD what is going on? We are ULTRA USERS cmon
I get the same message even with the 10 credit options.
I got the ultra plan and it’s annoying as f that I can only get 4 gens a minute a time.
@haughty bone im also ultra and i cannot even upscale to 1080 to download, i have to wait xdd
DO NOT use the upscale in veo3.1 ! is it not good !
where then?
honestly i havent pushed the limits since the 0 credits is back. So cba rly going for it rn, but atleast according to the feedback im reading past few days, you should be able to generate more than just those 4 per min if you are using the 10 credits one. The 4 ones per min still should be enough honestly. And thats something im thinking past few days about. Im quite sure that if we find a way to teach ppl how to properly prompt their ideas, im sure that the traffic requests will drop dramaticly. And its simply the prompts you are using, not some major changes. Consider it like that. If generations and AI is fallowing the prompts properly(works for 90% of not super complicated scenes with properly structured prompts), if prompts are good enough will simply require 2-4 generations top. And honestly since i learned how to properly structure my prompts i barely need to generate 50-60 times the same scene. So id suggest considering that as implemention. A specified prompts for different usages. Similar to expand option currently but with a library for it. Idk if its possible to implement something like that. Thats just a suggestion, but quite sure it will dramaticly reduce the amount of simulatious requests per user.
its actually helps
Imho it worsen the image quality if you want to upscale even more than 1080 p
but its says 1080p tho
yep but you can see a weird filter than is added on 1080p video
Hey guys , is there a difference in the ability to understand and render the prompt correctly between veo 3.1 fast and quality ? or is it just a quality thing ?
Are you using Text to Video or Frames to Video?
Higher resolutions are being worked on
ronny, its not the case about higher resolutions. We cannot use veo3 as we used before, constantly getting errors about generations or too fast
feel free to DM me directly and we can check your specific scenario.
You shouldn't have to change your mental model for prompting with Veo 3.1 Fast and Quality. Quality is just a more expensive model, but trained in a similar fashion to Fast.
Yep, we are working on Veo upsampler upgrade, one that should play nicer with Veo 3.1, and cut down on any artifacting.
Hello @surreal moth 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.” Also 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. Thank you!
Hello @edgy coral 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!
and what about this ?? how long we should wait ..my subscription is almost over and im not paying another 250 $ for somthing like this
Which model and how often are you trying to generate?
i only use this with Ingredients to video and none of them works
and you can't get any generations to go through? it is lower priority, which means we rate limit on our end.
....
I'll DM you to help you debug
Try to change your prompt a little .... it happend some times
Yep, we debugged offline and tweaking the prompt actually fixed the problem
yeah just a small stuff fixed the problem
How to fix it? I've been experiencing the same thing this past week (regardless of prompts). Before I could queue 12 videos at once with no problems, but nowadays I'm lucky if I get 4 videos in 10 minutes.
In this case it wasn't a throttling issue, it was a safety filtering issue.
hi guys
if you are an Ultra user using Low Priority mode, you should be able to get a few generations per minute, assuming you aren't running into other safety blocking issues
Hey everyone
if you see a message like "You're requesting generations too quickly..." it means that you are exceeding the rate limits. In the standard mode (Veo 3.1 - Fast), there are much higher rate limits, although credits must be used.
hi
Hey again , is anyone able to tell me if this i2v prompt too much for veo ? I cant make it work . I want my pov cam to walk to a door on the left of the image but it never does.
First Person View (POV) shot on Panavision Millennium DXL2 with Cooke S7/i lenses, moving with a very slow, heavy, and charismatic walking rhythm, representing a powerful unseen protagonist directing their path strictly towards the dark arched doorway on the left side. The camera advances deliberately towards the two women standing at that left entrance, passing closely between both, before crossing the threshold. The walk takes place on a damp cobblestone street in a gothic district shrouded in mist. The scene is lit by deep crimson red and amethyst violet lights with volumetric fog, creating a moody cinematic atmosphere with high contrast and shallow depth of field.
Thanks for some info, but are these limits going to return to previous figures? Just a week or two ago, Veo 3 on Flow was the best video-generating tool, but nowadays it's not really viable for anything more than a video or two, here and there. My Ultra subscription is ending in a week and I doubt I'm going to renew if the functionality isn't going to return to previous levels.
We are working on restoring, but it's unclear where the limits will ultimately be set. We want our users to be able to easily generate enough for their professional workloads. On the same hand, if our growing set of users are generating an incredible amount with free models, that puts a large burden on our compute (amidst a growing set of compute needs at Google from Gemini 3 and Nano Banana Pro).
Thanks for some clarity on the issue 👍
@inland pier can you guys please fix the system i get flagged for like everything like i am paying for a product here like everyone here is. trying to create cinematic videos is kinda getting harder and harder with the system flagging every little word like sad.. mad.. happy and so on
I hope restoration work resumes soon 🙏 as the progress of my work for my clients involves a very long wait.
This prompt issue is that even its too much info is not structured properly. You give ai too much credability and chance to hallucinate. My suggestion would be to try with json. If you are unable to create a proper one urself, go to gpt, tell gpt the pourpose of the prompt and ask for a super detailed prompt for google flow image or text to video. This way, you will get what you want with quite a low amount of generation. Either gpt or gemini. Both works
Also provide the image for i2v
ty will try it rn. im having difficulty to find the proper structure indeed !
Lmk about the result
I asked gemini to make the json but also failed . Maybe you have a specific json structure ? I asked gemini to collect info in our chat and give it my image so it generate the json but .. still the same issue.
Dm
i did dmed you
first time trying nano banana pro. i used a picture (a still frame from a video) to make a thumbnail out of it. when i upload it, it only shows a black picture. when i just proceed and cut out the whole (black screen) and tell it what to do with it, it seems to have recognized whats in the picture since the new generated picture seems to fit. but i actually wanted it to use the exact picture i uploaded. how do i do that? and why does it show a black picture (uploaded)? bug?
😄 and when i want to download the generated picture im also only getting a black picture 😄
Or am i maybe wrong and nano banana pro has a different purpose? (related to a usecase within veo 3.1 / flow)
Anybody else being hit with the "too many generations at one time" message? I have no other generations pending
same.. since yesteray
Using VEO and its misidentifying adult faces as minors. Where can i get help on this?
its on normal, or on "lower priority"
?
it's on normal, it's not happening every time so I'm thinking it's some sort of temporary bug... though, if we have it on normal, are we still able to do more than 4 at a time?
Can you share a screenshot of your upload where you see the black image?
If you are on normal (not low priority), you should be able to get more than 4 generations at a time (but still rate limited at a higher level). If that's not what you are seeing, feel free to DM me
Cheers bro, any suggestions on the best place to learn how to master the prompting. I think my prompts are pretty good but i typically generate multiple times 4-5 times for the sound quality. It seems to me no matter how good your prompt is, the sound quality is just random.
Thats correct, but i dropped from 40-50 gens to 4-5 per scene
@inland pier when is fix for Swedish language and Alphabet planned?
Language support or better swedish accents (but still spoken in English)?
Text and Swedish language nanobana pro has big problems with it stil
So both
oh Nano Banana? are you trying to prompt in Swedish, or you want rendered Swedish text?
Both
Have tried everything
It so crazy that it is even a thing
But I guess Sweden is only 9 million people
yeah at the moment there is a translation layer (to english), but I can discuss with the team about whether it should be removed
Cool did not know that. Think Elias has flagged it, because I wrote an email Logan and he forwarded it to Elias. But tried the other day rendering a thing but both promt was useless renders. I had cancel my subscription but I bought it again but still really bad render and exclusion flow and nano bana pro
So I just want it fixed but is probably higher up who has look at it
I'll discuss with Elias in the morning
Excision* the video had bad rendering with video and text and language and quality was not great but choice highest
Much appreciated, I don't know who contact for nano bana pro but tried render som stuff in Swedish language when they have a sentence and å ä ö in diffent grammatiks. You will see it renders are very very very bad.
@inland pier can you try it out see if get same outcome
Here is sentence
Can you please take look at it
Try this promt : välkommen till oss på kihlströms transport och lastbilscenter anläggning för att provköra.
Hemsida: kihlstroms.se/v3
Provkör gärna idag 🙂 och upptäck möjligheterna. Du kan också uppleva och testa användnings områdena men bilarna i terrängen eller vinter snön
yeah i see it translate. i tried the prompt "En skidåkare på en stuga håller en skylt där det står 'Välkommen till berget'" (english translation: a skier on a lodge holding a sign that says "welcome to the mountain"). It showed a man wearing ski clothes holding a sign that says "Welcome to the Mountain"
Cool cool 😅 I like thinking 👌🏻🥳🫣 😅
I have idea
Its all Swedish companies names
Let it render it then look at words or outcome maybe it good test a/b test
Like translation and then for rendering. But think probably you have go for grouping and costumers. Maybe look at old and new Swedish customer data. So it up to date
This there is alot of customers missing in Sweden.
Did you try this line? Can you please take look at it
Try this promt : välkommen till oss på kihlströms transport och lastbilscenter anläggning för att provköra.
Hemsida: www.kihlstroms.se/v3
Provkör gärna idag 🙂 och upptäck möjligheterna. Du kan också uppleva och testa användnings områdena med bilarna i terrängen eller vinter snön. Bästa kombon där du bestämmer själv. @inland pier
Can you upload in any channel?
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!
@hidden latch
Yes
Realistic smooth walk slow motion
For multiple times, audio genration has failed in the video, anyone facing this issue right now?
We are looking into this!
I used to face this issue before (with Veo 3) but for me I think it's resolved with the 3.1 drop
which is the best way to upscale an image?
is this related to the flowstateguide.com or something different by google?
what is diffrence between flow and this video generation option in gemini itself
woah you got ultra plan
Gemini has a direct Veo 3.1 generation while flow gives you more options to make it like a film. You can add ingredients, frames, generate images within the same tool and then combine them all together into a film with the scenebuilder! The scenebuilder and ingredients are my favorite features!
I didnt knew that flow can do more than generate videos
Yes you can now generate images with Nano Banana pro
btw is this direct veo 3.1 only for pro+ or free
I have pro but I have only 3 videos per day limit which means it must be very expensive I wonder is ultra plan worth buying if you are rich
On gemini, it's for pro and ultra. Pro gets 3 veo 3.1 fast per day, and ultra gets 5 veo 3.1 quality per day If I remember correctly.
Flow gives 180 credits per month for free users as well
Ultra gives you unlimited Veo 3.1 fast generations on Flow
makes sense this thing is do damn expensive
I thought damn is censored too lol
Yepp lol. Also I sometimes use Google Vids (which gives you 10 videos per day for pro users)
you have pro or ultra or free?
I have pro
Yepp. Flow gives 1000 credits per month for pro users. So I sometimes use Vids 10/day so if you utilize all the videos, it would be 300 videos/month
it saved me in college a lot
so for videos if you are pro (say from student plan or the like) you get that many generations?
Hahaa I also use Google Labs tools and gemini for studying as well!
3
1000 credits on flow which is equivalent to 50 veo 3.1 fast videos/month or 10 veo 3.1 quality videos
esspecially if you are lazy and not good at math but ambitious
Hahha true!
you can relate?
That may seem like a lot but due to the gambling nature of 'creative' generators it is not too many. (for one acceptable video you may need like 10-20 gens to land on a decently acceptable result)
btw what is nano banana
Yepp
like Ive seen people glazing it
It's the state of the art image generation/editing model from Google
yep Im studying CS so alot math is required
Very true! That's why I use Vids sometimes too
I made something like this
now objectively these are most of the time lifeless or uncanny as veo and even nanobananas doesnt really know how to do expressions, even so it seems not very artful, but it can be a fun toy to mess around with.
obviously seeing where we were and where we are, it is a huge step, especially for open source as people can study and improve them, this tech can be used in many other fields that need it more.
I think you should post it in #flow-videos・📹 otherwise it will get lost!
Getting expressions is really one of the toughest lol! You should try gemini for promoting. It will improve the video quality like a lot!
tho some videos look terrible and not what I wanted despite wasting 100 credits for quality idk what is diffrence between two tho like this one
I wanted them to play on machines accurate like for hammer this one where you hit hammer to ger score or cut things
this is quality video despite that we can see a lot of generation issues
can someone explain what is diffrence between quality for 100 credits and normal for 20
haha definately quality video
this is what the planet's energy grid is used for I see.
sometimes I get better results for 20 credits than for 100
so whats diffrence
first cat is supposed to be only one second blocks should be in the backs and third the hell are those items they use also what happens to this cat neck
I hope like gemini 3 came out veo 4 will also come one day and will not have these ridiculous generation mistakes
Is it me or I don't see much difference between fast and quality 😭
I just see some bright colors!
Also I love 3.1 fast
same like I wasted credits
try select use pc version
these three dots on right up corner try
maybe try to use diffrent browser
ohh yesss I heard many complain about the trial thing to get the perfect shot!
I think thats what made them provide unlimited fast generations for ultra users
Also I heard flow is not perfectly optimized for phones. So maybe it's best to use it on a desktop
love flow
so cool!
Yeah it is too heavy and they (GOOGLE) have to optimize it if they want AI to be consistent content on social media
yess maybe we will get a flow android app in the future?
I saw some references to a mobile app in the google forms sent in #flow-announcements・📣 and in chats
Can I say something? I pay €250 a month for Flow, and there's no 4K on Nano Banana pro, but there is on Freepik and others. What's the point?
This doesn’t look too smooth and clean
may be weird to ask but how to remove the watermark ? i.e. from gemini images without cropping.
use chrome for iPhone
I dont know, but i heard that Veo 3.1 fast lower priority with zero credits usage, has worse quality, and also a daily usage limit. I can't tell if it's true or lie
Ok what’s the max resolution or whatever Google flow can make on images?
@inland pier just wanted remind on the Swedish text and alphabet
How did you do it and what was the prompt I never get this good quality output :/
Hello from the Flow team! Veo 3.1 fast with lower priority does not have lower quality - it's the identical model as the 10 credit one.
I used a reference image as first frame, and edited the video after in editing software
same here but just keep trying
also there is a new problem this morning, none a single viedo a generate has audio
Same here veo isnt working only free generations 1x4 every other generation fails
this is starting to be unusuable for a professional level
Kling bringing some killer updates today...better get this shhh together
how much kling cost
lets just say you can get a whole truck of generations worth for the price of flow
i love flow and all, but its turning to have issues on 80% of days
i pay 250$ for flow every month..
yeah specialy in the last 2 weeks
its was good at some point
why would generations ever fail due to high demand if I use the 10 credit version of 3.1
like whats the point then bro
all we can do now is just keep trying and hopefully they fix this soon
Even Quality generations are failing. It's really a pity.
Who has Ai ultra subscription ?
They say that veo-3 is goodbye ....)))) Well, we'll see, we'll see.....
I’m using Veo 3.1 (10 credits) and half of my image-to-video jobs keep getting stuck at 99% and fail. I tried 10 videos and about 5 of them failed. Seems like the image-to-video pipeline is bugged right now
everything is crashing. Constant errors/too fast lately. People are using it on professional level, paying 250$ for ultra and getting this.
It’s impossible to create anything, even though I’m on the Ultra plan. What’s the point of having lower priority and regular Veo Fast if neither of them works? Platforms like Syntax.ai are draining all of Google’s resources, and you guys are doing nothing about it — why?
I've been trying to generate one video for 35 minutes. Forty attempts, four of which produced crap results, and 36 crashes after reaching 99%...
Ok is it image 1, image 2 and image 3… or is image 3 and image 2 and image 1…. So would the face be image 1 or image 3? So I can reference it in a prompt
hey there, do you have any nano banana prompting guide you recommand ? ( I mean I have google but is there any specific one that give you better result ? )
why this is still failing like generation fail and error like you making too quick even im not making anything in the process is it still not fixed prperly
see this is reall bad really frustrated now
same thing happening to me for the last few days
they say they will fix this by november 28th but still keep hapening to me this crap is not wrokign fine real waste of money
yeah it's pretty frustrating especially when you're paying 250 a month for it lol
And when it does get fixed theyll provide credits that expire if you dont use them
Creating videos in Flow is getting worse every day...
Today flow is worse
Many people will have already stopped paying for the month, it's a waste of money 😫
General question: is it worth using the 'Flag output' option? Failure to follow the prompt is the one thing I would flag most of the time, but if flagged output doesn't actually feed back into the model and improvements to prompt adherence is actually completely outside of our control as users, I'd rather not bother.
we need to do something about this they not fixing this \
they say 28 they will fix this but they give us some credits and gave a new low effort thing nothing fixed
in november first weeks we make videos by putting 5 in the qeue now we cant even make one propely
flag is like "the model allowed an output that clearly violates the content policy and/or the law, google needs to fix this immediately"
at least that's how I see it
i think they just don't care abot us at this point how such a big company like this doing something for their users also paid ones specially with ultra accounts wth man
didnt post for 2 weeks bcs of this
from last week is always error because of ai movie challenge from google. and one of the requirement is must 70% of video must use gemini veo
same man
Hello if you're encountering errors and /or bugs or would like to leave a general feedback about Flow please don't hesitate in utilising the #1319424187730755636 channel with as much context as possible. It helps to keep the team organised and it’s also where all the technical issues are logged.
The Flow team is working hard to address all the issues posted in the bug and feedback channel.They also read and appreciate all the feedback they receive from all their users regarding the products, so we kindly ask for your patience as they work through the requests
Thank you !!
Hello @granite willow 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!
Anyone know?
That one person reacting to this message be like
what happend , error for 2 hours ?
Hi everyone! Can you guys explain to me if there's a advantage in using white background pictures to compose flow videos?
🥲 
We'll see, I always get really nice results during Gen 48, then the day after its crap again. I can't help but down and lol, look at how the skew the data visualization
normally i would do like 'image [1] is doing the action of the image[2] etc etc'
I always refer back to the Vertex guide for the prompt template stuff
Need better support for the Swedish language.
why i can't generate ?
day by day not getting better but getting worse 💀
mmm i have a big issue as last week, failed video generation and credits gone
i lost 380 credits last week, thought problem was solved, but now more 140 credits lost. sent email for assitance few days ago, no solved yet.
They keep asking me for screenshots, I even sent videos,
It’s a problem that comes and goes ...is anyone else here in the same situation?
hey, trying to contact support for antigravity. who's the best person to reach out to?
There's a channel called vibecoders in the Google Developer Community! You might be able to get some support in there by other fellow vibecoders that use antigravity 😄
vibecoders💀
is that a dc?
thanks 🙏
Can someone help me with prompting this video?
appreciate any help prompting this video above
But the gun and human hands should be excluded and not shown in the video
Doesn't matter how much I try, it still shows up in the video
🙏 🙏 🙏
Hey is there someone from support I can directly reach out to via email? I'm stuck in a never ending conversation with an AI bot and I really need some help on my billing. I'm currently on the ultra plan.
another day another....
yeah same, me too
Hey! I reported on the 27th that the website has consistently failed to generate videos with audio in portuguese language from portugal, something that it has been generating fine up until then. It has been almost a week, my credits haven't returned and it still isn't generating properly. I have to try everyday and waste credits because no one provides feedback, even after I talked directly to a Google One agent. What's going on?? Can someone please help?
Yeah I think they are just backed up, I’m getting the same error. This is what happens when the demand is more than the supply things start slowing down. Annoying yes but it’s nothing earth shattering… unless you are down to the wire and can’t spend those credits or have client work to finish oh damn that right there is where it hurts.
super happy with this 3d miniature render. In this post I explain how I made this video (via Flow ofcourse!) https://www.instagram.com/p/DRxC0NejM90/
cj_gta_sa.gif
oh s... here we go again
Same here
Other 140 credits gone ,last week 380 and despite reports and proofs i m waiting for response
damn...more errors today eh
HI, a quick question, does prompting in JSON give more accuracy?
Kling 01 looking better each day, though not sure how they are with content moderation and false positives
kind of, you can do stuff where you basically prompt what happens at which second, though of course its still AI and might ignore you
this chat gpt thingy gives you a formatted prompt based on what you describe: https://chatgpt.com/g/g-68a6dba6e48c8191a6ae71fed1b15952-veo-brainstorm-crew
please open a thread in #1319424187730755636
fixed or not?
no, not ifxed, tested right now -.-.
How is the Flow service working? Can it generate properly with Nano Banana Pro? Can Yo veo 3.1 Fast be used already?
hi all - here to troubleshoot (thanks for reporting in #1319424187730755636 too!) - quick question:
- what mode(s) are you experiencing "Failed Generations" and/or "Generating too quickly" errors on? (I2V, for example). feel free to reply in thread for organization's sake:
How to download the high quality of the generated output when using Nano Banana Pro on flow? My initial image was 1080x1920 amd the output from nano banana pro was 1365x768... If i use Nano Banana Pro in Gemini the generated output is 2752x1536
10 generations failed , i use 3.1 fast
this is my second time i have this issue (last week i lost 380 credits). sent email to support with proofs (screenshot and videosI ) i m waiting for their response yet...They reply very rarely and only after several days; they’re extremely slow to respond.
On top of that, the issue with failed video generations happened again today…
Impossible to use FLOW at moment
flow cant make this type of videos
just seing the remove option (outpainting) pretty cool!
Nice, hadn’t noticed that yet
From my personal experience, I haven’t gotten anything better out of JSON prompting. When I’ve converted some of my prompts to JSON, I feel like I’ve always had better luck with just natural language.
So it's a bit odd – unclear when I lose access to Nano Banana Pro (or when it's used)! Refreshing the page shows it was Imagen 4 (thought that seems unlikely in many of the cases)
Further, it's odd to buy more credits for video gen and not have extra access to Nano Banana Pro! I get why this is, but I bought credits to continue a project and then realized I couldn't – since 🍌Pro is part of my workflow
Hello! When you click the download icon on a video, you'll see the option to upscale to 1080p
I got a question? With my ultra subscription I can add more people of my family to my ultra plan
I can do the same with the pro plan?
i dont see yet but ill try again, thank you
yassssss thank you for finally adding object removal
They did lol
Hey I'm using flow in my ipad and I can't find the camera control option. Can someone help pls.
It seems like all the videos I generate from images are in slow motion. They always require a speed-up to 1.5\times or 2\times to be usable. Is this happening to anyone else?
Nano Banana pro image gens aren't working. Just says something went wrong, try again later. Happened after the update.
This been happening a lot
nice update , we have now more control . let build
bloody flow making me redo scenes that I dont need to ..I back out of add the scene and it looks like the video i just did is not there so I have to do it again only to find out it is there but just flow dont show it untill i refresh about 6 times ..wasted money
sorry not ranting but yeah i am just angry about time waisted ..
How do I use Video to Video Generation? Veo
omg this is so frustraating mannn
Nanobanana in flow is having problem for everyone ?
Yes
Use the latest tiktok video generated by flow.
A truly controllable template.
Used in tiktokshop.
anyone having issues with audio? My last 10 generations are coming out mute
common Flow...get you shhh together
every day new problems ..thank you flow team
Who created Consistent Characters and made the Video? Please send me a message. I want to see your profile and work
Is it possible to use a 6 image grid as an ingredient? Meaning, will it pickup on the info without roviding poor quality?
I haven't used veo 3 in about a week now due to noticeable drop in quality. Is anyone also experiencing this? Has the quality improved at all for you? I do not want to waste credits on clips if the quality is still noticably degraded.
So we get the object removal, but if I want to remove some Veo3 generated artifacts that happened in the middle of the video it's impossible? well done once again.
Way more crashes (maybe week ago it was already happening), but the quality sometimes is better, sometimes worse. While the quality in overall decreased, I am still able to make some progress. Sometimes just trying different time of day...
So still a noticeable overall quality decrease, that's a bummer. My style of video was already pushing veo 3 to its limits so with the quality downgrade its just unusable for me now 😢
I would love to test the remove feature, but me, an ultra user, has no credits to do so
Let us know what you think
Do you mean tiling the multiple images into a single ingredient image? that should work. you can use Nano Banana Pro to help with the tiling as well.
Yeah currently we support 1365x768, but higher resolution is coming soon to Flow
hey guys i have a question. when generating text to videos, can i upload some assets that flow can use in creating a video?
Unfortunately Veo does not support all languages. Previously (for quality reasons), we translated prompts to English, which means that user requests for spoken audio in other languages would come out in English. Recently, we removed this translation, so if you request that a character says something in French, Hindi, Spanish, etc., it should (most of the time) respect it. This does not mean all languages will have the same quality as English though. Please try the new multi-language Veo support and provide feedback.
Hi @distant steeple , yes you can either use an image as a starting frame, or (if you have an Ultra account) you can use an image as an "ingredient" to reference characters, objects, styles, backgrounds, etc
Waiting for 2K to give my feedback to the tester group.
Guys i cant see my old work from morning in flow library (veo 3 videos) its not loading? What should i do