flow-general・🌊

[clever axle]

it worked!

[bright crown]

aww yeah

also... are you cutting mid scene? VERY powerful way to save credits

[clever axle]

never did that, since i use the 8 secs.. how that works?

[bright crown]

like in this... its one prompt, two shots... cutting to the closeup mid scene, for 2x 4 second shots for 20 credits, rather than 2x8 second shots which you dont need all of

if you need the whole 8 sesconds, than this wont help you

but most of the time you dont need all 8 for a shot

[clever axle]

oh.. ive made that 🙂 very useful

[bright crown]

coo coo

ok, who wants a challenge? I got something I just cant get to render...

youd think it would be super easy! Its a Centaur... but heres the deal... you gotta put the human portion in clothing, like a baseball uniform, or mail man, anything, just not the brawny barechested mythical type

Ive never seen it done, ever

closest Ive gotten, to a kentucky derby winning centaur... and its all kinds of wonky and no way to get a full shot of him that I can find.

[spice phoenix]

So if you had the AI ultra plan and unsub'd you cant re sub and get the same 124.99 price?

[bright crown]

oh the fails are all spectacular! Ive wasted more credits on centaurs then everything else combined! thanks for giving it a go! lol

right? like Veo can create the craziest things, Alients, giant spiders, rocket ships, huge cities, but not a centaur

I acutally spent about an hour chattting with gemini about it, and why its so hard... my favorite line was when Gemini said... classic painters from antiquity even struggled with the centaurs form, its connecting two torsos... theres two rib cages if you think about it, so the anatomy is really difficult

ok, you gonna make me find my favorite fail, one sec

my two favorite fails

thats pretty great though

using veo 2 ingredients? or did you leave flow

ah, the aspect ratio gives it away lol

slow clap

nicely done FM!

thats great, the movements realistic

still impressive

thanks for playing along, I appreciate it!

I love your third fail attempt up above, where he has his own reigns

look up above at my fails

the yellow one is pretty realistic, I just cant get the body to show

centaurs have become my measuring stick, ill keep going back to them as the models progress in future, and when it can easily do them Ill know they've really upgraded

thank you!

Im really impressed at how once you assembled the cenaur, it does to the motion well, like the arms swing properly and stuff

yeah, it pops a horse head on a LOT of the time

its the clothing that throws it, you can do barechested ones, from antiquity style

even the ones in harry potter movie didnt have clothing

uh, are you in the right discord server?

[crisp basalt]

https://tenor.com/view/blonde-woman-saying-oh-tiktoker-saying-oh-sabrina-brier-confusion-astonished-gif-4516544766285635914

[dry frost]

It's how they carry more mail.

[half dew]

hi

[gentle sparrow]

The more I use Flow - the more I wish there was a way to leave comments on clips or flag them (good, bad, funny). That woulk make it easier later to go through them

[twin raft]

anyone know a way around that made a frame (of myself) using AI and Flow is insistent that this is a celebrity so wont allow me to upload it to create my short story which is the entire reason Im paying $200 a month for this service that I cant use. I emailed and havent heard back its been days. Is there a way to get around or to have an image clearly not celebrity approved?

[soft leaf]

Does anyone know if you automatically get refunded when "No audio generated"?

[wooden monolith]

No refund for No audio generated

I wish Flow tells us which part of the prompt is violating the policy. I already revised the script several times, and still can't pass the violation warning.

[soft leaf]

What the heck?

[soft leaf]

Does "No audio generated" means something is violating the policy? I just figured it was a bug.

Or do you just mean independently, which sure, sounds like a good idea.

[wooden monolith]

Sometimes it's a bug on their part, in my case, sometimes it's because of a child character speaking.

[soft leaf]

I feel like this would be easy to do

[wooden monolith]

They are so random. The first time my child character was able to speak. Then the next they say child are not allowed to do some speech.

[fallow sedge]

Wesley Snipes 😄

[sterile junco]

this is redicuslous.

Okay not only are videos failing to generate... now im even getting this message that i have to wait cause im generating too many images.. like wth this is my first generation lol... ive waited about 15 minutes to try again.. same thing.

am getting tired of this.

Staff doesnt give a f...

[halcyon spear]

does anyone knows a way to upload a JSON with embeded reference images into the flow UI?

[fallow sedge]

That's not possible, as far as I can tell. Your best bet is probably the “ingredients” feature, but it depends on what exactly you're trying to achieve.

[halcyon spear]

im trying to create a short film with consistent characters but i will try the ingridients ,method. thanks 🙂

[fallow sedge]

Note that “ingredients” currently supports Veo 2 only, so you won't get audio.

[halcyon spear]

thats a bummer as i do require audio as well

[wary raptor]

Guys I need help I need to make a video prompt of a red cardinal bird on the ground then flying off into the blue sky in a cinematic way but the Ai comes off weird or strange looking & slow motion please help!

[fallow sedge]

I guess that Veo's training data contains many documentaries containing birds taking off in slow motion. That's probably causing the problem here. Veo thinks that birds always take off in slow motion.

I tried to render a chameleon that catches a fly, but Veo also prefers to show this in slow motion, even if you want it to be rendered in normal speed. Same cause, I guess.

[wary raptor]

That’s the nicest one I seen thanks!

[fallow sedge]

But it's also slow motion.

[teal ether]

does anyone have the prompt for those asmr fruit cutting videos?

[wooden monolith]

Hi, anybody else havent got their bump on ai credits?

Im confused. Is it only for those who have made "payment" after the announcement?

[peak hatch]

I think its next month but not 100% sure

So right now u cant generate realistic characters?

[wooden monolith]

@main saddle said it will be credited within 48hrs

[wooden monolith]

Not sure what this has to do with mg Q 😅

[peak hatch]

This was my question to the community 😄

[fair gull]

Does anyone have any tips on how to get a decent English speaking voice with a spanish accent? I'm using a generated image of an animal and the client needs it to speak with a spanish accent for the scene. Every generation comes up regular Americanized accent.

[wooden monolith]

I can generate realistic characters no problem

[rancid ridge]

Create an Panipuri video

[peak hatch]

@rancid ridge Its not free and u have to go google flow to generate
https://labs.google/fx/tools/flow

[soft leaf]

I see how I can upload assets to Flow, but can I also download assets? Or browse assets?

[soft leaf]

(I was thinking of downloading assets to correct saturation drift, and then re-uploading them.)

[umbral thunder]

Hi @twin mortar!

It depends on your subscription renewal date. Current AI Ultra subscribers will see the increase upon their first plan renewal.

If your subscription was renewed before August 18, 2025, your credits will likely be doubled the following month.

More details on this article: https://blog.google/technology/google-labs/flow-whisk-ai-credits-update/

[soft leaf]

Hi Lígia, I subscribed to Ultra for the first time in response to seeing the announcement about doubled credits, but I still haven't received doubled credits for this first month. Can you help me?

[twin mortar]

Ok thanks. My subscription renews to the full £250 payment on 9th September - so is that when they’ll be added?

[soft leaf]

Tech Support keeps telling me every day to "wait 24 hours, and then they will definitely be there", but it hasn't been true yet. 😞

[wooden monolith]

When was your last payment?

The support i talked to said only those who paid after the announcement will get the 2x credits

[digital escarp]

I thought it got bumped to 2x credits because the 3 month intro price of 125$ is phasing out for all the early subscribers. So they bumped it up to more credits to entice folks into staying despite having to pay 250$ a month now.

[tropic jewel]

bro

whenever i tell veo3 to say some speech

it makes the peoples role switch

idk how to explain it

but the people's speech should be the other way round

[covert spruce]

Visuals:

A close-up of an old handwritten letter placed on a wooden table.

A half-filled cup of tea beside it, steam rising.

Rain pouring outside the window, faint thunder.

Camera slowly zooms in on the ink smudged by water with Soft piano with rain sound effects

[digital escarp]

Best bet is to assign names and positions in your prompt structure. And be super rigid. That’s what I’ve done and it’s been working

[tropic jewel]

i did bro

i did and it still aint working

[covert spruce]

(0:00 – 0:15) Close-up of an old handwritten Hindi letter on wooden table, half cup of tea with steam, rain outside window, thunder, camera slowly zooms into smudged ink.

[soft leaf]

My first and only payment was on Aug 18, I signed up in response to the announcement about doubled quotas, since that finally brought it into the realm of cost-effectiveness.

[tropic jewel]

wdym

wait dm me

[coarse wasp]

hey do the monthly credits stack?

[soft leaf]

I paid after the announcement, but I'm worried that I paid too soon after the announcement, and that I'll never get someone to get me squared away, heh.

[wooden monolith]

You will definitely get it then

[wooden monolith]

No, it resets

[radiant mural]

really? when you buy credits it says they are valid for 12 month.. @fierce valve can you confirm?

[thick smelt]

Hi guys, are you also having problems generating images on Flow? For me, it says that I'm generating too many too quickly and to try again.

[bright crown]

what are your settings? 4 x veo quality?

[peak hatch]

does veo3 quality have better prompt adherence then veo3 fast... Not looking to burn credits experimenting 😅

[bright crown]

Thread

[bright crown]

well, my gut says its constantly changing, but I would definitely use veo3 fast for any unpaid, learning type work... also, you can cut to a scene mid prompt and get two 4 second spots in one 20 credit shot... Ill tag you in an example from yesterday further up this channel...

[thick smelt]

veo 3

[peak hatch]

i have same problem but u can just generate in whisk or imagefx or googleai studio

[thick smelt]

I tried, but it doesn’t keep the same style across different images.

[crisp basalt]

oh?

[worldly yacht]

i just saw this and have no idea what it is , i might get credits to see what it does

[crisp basalt]

it doesnt use up your ai credit btw so dont worry,

[worldly yacht]

so i can test it with 0 credits ??

[crisp basalt]

yeah

its basically some examples on how to use ingredients to video feature, camera control etc...

[worldly yacht]

im trying to figure it now , doesnt let me add lets say the hulk or deadpool , says something about not allowing real people

[umbral thunder]

Hi @twin mortar!

I found more information on this matter, and actually the increase in AI credits should happen the first time the plan is renewed and also after August 18.

I am checking if there is a deadline to roll out the credits to all Ultra subscribers.

[worldly yacht]

hello 😄 dont us pro users get some love too ?:D

[umbral thunder]

Hi @worldly yacht!

At least for now, the extra credits are only available for Ultra subscribers.

[worldly yacht]

i hope so soon 😄

Batman taking a swim with the batmobile submerged in the water

[full flicker]

As stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now.”

[twin mortar]

Thank you 👍

[dusky basalt]

I'm very sorry to bother you all. I would like to ask whether the creators of the videos generated in Flow have the right to use them for personal commercial purposes, such as as promotional materials for novel homepages.🥹

[twin mortar]

Can’t see why not. Plenty of us using Veo for commercial purposes.

[dusky basalt]

Thank you for your reply. I just didn't find clear terms from Google's official channels, so I have some concerns.

Based on the information I have found so far, the content generated through the Gemini or Drive portal is not a big problem, but the content generated through the Flow portal in Labs is still unclear.

[dusky basalt]

So can it be understood that: it is feasible to gain profits through the traffic brought by generated content, but there are risks in selling the generated content independently.

Thank you for your answer, it is very helpful to solve my current confusion👍

[frank fox]

[rough halo]

so do pro and ultra now have the same access to everything in flow? since ingredients has been added to pro, what sets them apart?

[raven kettle]

Hello I'm definitely newer than new and I have no idea what this is for, what I can do or why I'm here to be honestI got here via Google Flow because my text to video prompts kept getting error message saying this may violate our policies blah blah blah.

[dusky basalt]

Pro：1000 point/month
Ultra：12500 point/month

[dusky basalt]

Maybe you mentioned some keywords involving children/violence/real people; or maybe you didn’t mention them, but the AI ​​recognized that the scene you described implied the above?

[sterile junco]

Guys i have a question, if someone could answer with a tip id be very glad.
Im trying to generate a video where i need the character in this case a "racoon" to whisper... however it doesnt matter if i use keywords like "quietly" "softly" "whispers" and so on the character just doesnt want to whisper. even tried "asmr"... but it only generated 1/10 videos where he actually sort of whispers.

[raven kettle]

Cuss words maybe?

[rough halo]

That I know. I mean with regards to access to models. Is there still anything that’s exclusive to ultra? Like how ingredients was before it was brought into pro. And ultra instead to 25k btw

[lavish field]

anyone loosing sound in their vids now?

all the videos i make now have no sound , first time I've seen this, maybe a bug but if so its a costly one.

[proud cipher]

This time, I would like to share the video I made with veo3.

https://youtu.be/arsmcS2gLD4?si=OPk88fcGE-RRlIdD

[inland nymph]

Hy everyone

[wild jetty]

Hi do you make scenes in flow and edit it somewhere else? Have a problem with transitions between each 8 seconds of video

I was wondering what software you guys use to edit, make transitions etc…

[lavish field]

yeah most of us do edit in other programs , the one i use is capcut , due to it being very easy to use

[lavish field]

https://www.capcut.com/

[wild jetty]

Thank you very much ❤️

[lavish field]

more than welcome

[undone briar]

The blocking of uploading "photorealistic" people is ruining this service. This was literally generated with AI.
Will this be fixed ? Otherwise I'm cancelling my subscription

[oblique gust]

Ultra now get 25k credits a month

[rose latch]

so hard to control

I tell it "he goes to the driver seat and drives off" it makes the guy teleport from the passenger seat

[wild jetty]

Appreciate it 🙏

[clever axle]

I wish to share my generations, but i only do degeneracy

[lilac narwhal]

If you want a joke, they blocked the "photorealistic" to an extent that even this is blocked, like.. Yeah even I am not sure if it's a real person or not 🤡

[boreal cargo]

a racoon whispers to microphone, ASMR style

[dusky basalt]

Thank you for keeping my knowledge base updated😁

[dusky basalt]

When is it worth upgrading? (GPT's answer)

1. You produce a large number of videos monthly (Flow/Whisk) and prefer the quality and stability of Veo 3 Official. Ultra's 25,000 credits significantly increase available capacity.
2. You truly need Deep Think for long-chain inference, script splitting, and shot planning.
3. You can use 30TB of storage plus YouTube Premium (large library, multi-person collaboration, and cross-device backup).
   Hope this helps.😁

[sullen lance]

Hello, I need some advice on how to best write a prompt so that the person I want to speak to speaks. 😄

I always write exactly what I want them to say, but they do the exact opposite 😄

[rough halo]

Thank you. It does. So as far as access to models, they’re both identical @dusky basalt

[dusky basalt]

I don't know the standard for swear words. After all, the prompt words I reviewed with Gemini still triggered several violation warnings.🥹

[dusky basalt]

You can describe your needs in the form of a story in Gemini or other AI tools and let it help you convert them into prompt words.

[rough halo]

Yeah I found good results specifying the subjects as person A and person B, and keeping those descriptors consistent in my prompt.

I’ve also had good results saying stuff like “blue eyed man” and “dark eyed man”. Maybe I got lucky with those but it seems like giving your subjects a descriptor tag, and referring to them consistently with those tags in your prompt helps.

[worldly yacht]

thats amazing

[jagged jackal]

That's what I do with decent to good results. Not 100%, but still gets pretty close.

[broken prism]

"A vast circular holy city surrounded by tall green mountains, a wide ring road encircling the entire city, inside the city futuristic urban planning with modern train tracks and flyovers, scattered natural blue water lakes shining under sunlight, at the very center a massive Biblical Solomon’s Temple made of golden stone and white marble, glowing divine light from the temple, the whole city designed in Unreal Engine with hyper realistic details, soft god rays from the sky, cinematic atmosphere, Jesus Second Coming holy city view, epic wide-angle shot, 8K ultra realistic render."

[bright blaze]

What's the best tool to create website nowadays?

[burnt oriole]

I wonder if we pro users will ever have an increase in the flow credits

also would love it if we could have a small quota for free users

[slate pivot]

what prompt did you write, pls insights 🙏

[clear cipher]

study and learn, you are capable

[slate pivot]

how?!

[sonic zealot]

https://www.instagram.com/reel/DNoMY5stIw_/?igsh=bHZhenltcXZqNzVs

buddah4k

Moments Made 4 U ❤️ 8/25/25

Video by @graystandards

-I wanna give a special thanks to @filyotf4k for mixing, mastering, and helping putting this project together❤️ my brother
-Also wanna give a special thanks to @armanimar4k for sacrificing his time to help put this together and also the Features I chose for this EP @prodbyspook who n...

Instagram

[clear cipher]

result

[wary raptor]

Help I need this video in 9:16 format and at night time

[quaint rain]

hey @clear cipher are u facing this error on flow is it just me?

was working fine a while ago but suddenly cant create video on veo 3

[clear cipher]

It is not yet available to expand the video with audio

I was a week behind this too

[quaint rain]

i am not expanding the video just generating new one with veo 3. sometimes it works sometimes it doesnt. wondering if other poeple facing that too?

i just created the video an hour ago with audio and veo3

but now not working

[clear cipher]

it must be an instability

[quaint rain]

try it some other time?

[clear cipher]

yes yes, sometimes it must be a full cache

[quaint rain]

cool thanks mate

[clear cipher]

I am not official technical support, I am just advising you as a friend

[quaint rain]

yeah i understand mate

[twin raptor]

Which text-to-image model is used in flow's video creation page? is it imagen 4?

It feels a little different

[clear cipher]

I really didn't understand your question

😦

[radiant mural]

when i try to create "from elements" it doesn't allow photorealistic portrait in my region. someone knows which countries are allowed to upload photorealistic pictures of people?

[clear cipher]

but this doesn't stay in the prompt?

mine has

[forest raptor]

the flight feels wrong, but it looks adorable on the ground

[blazing torrent]

Hi guys

How do i earn credits for veo 3?

[clear cipher]

Here I have almost 2 years of studying prompt engineering

[clear cipher]

It's in Portuguese but also has American content.

[peak hatch]

Does generating images in ingredients to video on flow work for anyone?

Any idea when it will be fixed because i saw multiple people complaining

[peak hatch]

You earn money, and buy credits 🥹

[quaint rain]

not true, was able to use it a while ago

[peak hatch]

Go back to your generation where you think you used it and double check if its veo 2 or veo 3

for me it gave a warning that it will automatically swap to veo 2 if it does not support veo 3

[quaint rain]

yeah i have couple of generations from veo 3

[peak hatch]

I mean ive had experiences in the past with different generative ai where if ur stubborn u eventually get it ur way, so i guess u can spam requests ?

[quaint rain]

i have spaming a bit but doesnt work, probably will try it other time. wants to upgrad to ultra but wanna make sure its useful for me

if its keeps having this issue not worth it for me in Ultra sub

[peak hatch]

yeah i bought ultra like an ape im kinda disappointed now tbh, veo still burns a ton of credits on mistakes like other video ai's and its more expensive per generation and does not have unlimited generations

And has a lot of restrictions, but its quality

[quaint rain]

oh thank u so much mate for the feedback. yeah it doesnt make sense to upgrade at this point tbh

[peak hatch]

@quaint rain I wouldnt say all that bro, its probably still the best one out

[quaint rain]

Quality wise its the best but so far not a smooth experience

its the best model i know

[clear cipher]

input prompt:
Now she stands against a neon-lit rainy city backdrop, droplets catching ambient light around her silhouette. Her expression carries a soft melancholy, with lowered eyelids and a subtle downturn in her lips, reflecting the urban solitude around her.

[quaint rain]

hey buddy which platform u using to create prompts?

[clear cipher]

Opal

[peak hatch]

When a generation is stuck for a long time on 99% what does it mean?, it could be a sensitive topic im trying to generate a gun?

[clear cipher]

I exchange for pro

[peak hatch]

Dont understand what u said

[worldly yacht]

The adventures of super and batman 😄

[jade relic]

Hey all - I love this community and the tool, really having tons of great time working with this. I have a problem - I suddenly can't access my FLOW app, despite having an ULTRA sub and credits left.

What's the best way to resolve this?

[worldly yacht]

can someone make 1 video for me please ?:D i ran out of credits

[jovial wave]

i just wasted like 2000 credits trying to get rid of the black bars in a widescreen it keeps making, help me please

[sonic zealot]

https://www.instagram.com/reel/DNoMY5stIw_/?igsh=bHZhenltcXZqNzVs

buddah4k

Moments Made 4 U ❤️ 8/25/25

Video by @graystandards

-I wanna give a special thanks to @filyotf4k for mixing, mastering, and helping putting this project together❤️ my brother
-Also wanna give a special thanks to @armanimar4k for sacrificing his time to help put this together and also the Features I chose for this EP @prodbyspook who n...

Instagram

[kind bison]

share your prompt

[sonic zealot]

If everyone could let me know their thoughts on this video. Everything is Ai

[wary raptor]

BAHHHUMMVISIS

[sage rock]

Hi everyone! 👋
I have a quick question:

I’m working on an animation project in Flow-Veo, going from a single image to a video.
My doubt is:

👉 When generating video samples, is it better to keep them (even if they don’t turn out the way I want), or should I delete them from the workspace as I go?

I’m asking because, based on how other AIs usually work, they “learn” from the interaction with the user. So I wonder if in this case it makes sense to leave all the test videos in the workspace, or if Flow-Veo doesn’t work that way.

Thanks a lot for your time and any insights! 🙌

[edgy thorn]

Maybe its a good idea to download them, if you edit on premiere, you can use those files as a new shot, maybe an expression or somethin in the shot can be helpful in your final edit.

[sage rock]

Yes, exactly! That’s what I’m doing — I download the clips and keep them, so later I can pick the best takes and edit/compose them into something better.

My question is more about whether it makes sense to leave those projects in the Flow-Veo workspace. I’m wondering if, similar to other AIs, Flow-Veo would adapt or improve as more clips/videos are generated, or if it doesn’t really work that way.

[edgy thorn]

Hi! I have a question... I am working on a project, and I have to create a shot, but this shot is a for a commercial I can't find much information on the google terms if I have an image created in Elements envato with AI, and then use it in Google Flow, can I use it in a commercial? If not, it will be necessary to buy a membership on Veo3 to get a license for using the final shot in a commercial way?

I hope you can help me understanding this part of the terms, for the moment I pay for the Google Pro account, not the plus, but if its necessary to pay for plus to get a commercial license I can afford it for my project.

[edgy thorn]

If you create a lot of videos... its better to clean your space, I started with Image and now I have hundreds of images and then I can't find old stuff I need now.

[sage rock]

Hahaha, it happens all the time.
I’d recommend two things:

If you still remember the prompts (at least roughly), you can use Ctrl + F to search by the words you used.

As you download files, start organizing them into folders with titles/names that make sense to you — even adding dates.

As designers/editors, we all know the struggle when we have too many projects and end up naming files like:
final… final_final01… final_final_edit_05 😂

Organizing by name/title + date in folders makes things way easier in the long run.

[tepid quest]

/prompt

[amber ether]

Guys I keep getting the same error "you are doing too many things" when I try to prompt. I don't do a thing though. Just open flow and type text and when generating I get this message.

Also how do I change the language to eng? 😅

[placid hatch]

Hi everyone! I have a question about using ingredients: why is it that when I add a character, the face looks quite accurate up close, but when it’s shown from a distance or smaller, it looks much less similar? Is there any tip or workaround for this?

[radiant mural]

no I mean when I change the creation mode in the top right corner of the prompt window from "text to video" to "elements to video" (there is also the third option "frames to video) - what I want to do is to give a upload a picture of a character and then have him in the video. but when I upload a realistic picture of a person I get an error that states it is not possible to upload realistic person images from my region (which is germany). so i was wondering if that would be possible from other regions and if someone knows which countries that would be.

[fallow sedge]

Veo does not “learn” from your previous generations (thankfully, I might add). So, whether you keep or delete old generations is just a matter of organizing them for yourself.

[frail tangle]

Are there any users who paid before August 18 and are currently using Ultra who received double credits?

[orchid geyser]

It’s time to make veo 3 unillimited... !!

[burnt oriole]

I'm getting this error while generating images using the text to image feature from ingredients mode

(pro user)

[burnt oriole]

omg is this actually happening on flow?

[orchid geyser]

No but, we wish and i think it’s time from google to do it

[burnt oriole]

how do you expect it's time? 💀

veo 3 is still very costly even though they plant a bunch of more TPUs

[orchid geyser]

When you see the competition as sora of gpt chat they have a service that offers generations of unlimited videos included in your subscription...

[turbid lintel]

maybe veo2 unlimited is good

[solid cargo]

I'm getting a message saying that this feature is not available in VO3, does anyone know how to fix this?

[peak hatch]

same, cant generate images there anymore

[peak hatch]

Veo 3 isnt compatible with ingredients yet. but veo 2 is so if u dont need sound its not that big of a deal..\

[solid cargo]

noooo, and the frame to video tool is compatible?

[signal jetty]

i want to generate 4 images before creating a video, but it doesnt let me

oh same like above

[quaint rain]

same here, can anyone from Ultra Sub can confirm if they are facing this problem?

[past tundra]

hello, i need Support, because it won't let me do things supposedly due to my location. I want to convert an image to text.

hello, i need Support, because it won't let me do things supposedly due to my location. I want to convert an image to text

[quaint rain]

Hi @main saddle alot of poeple having problem here generating videos on veo 3, wondering if we upgrade to ultra would it fix it?

Attached SS, sometimes i am able to generate videos tho.

[signal jetty]

i have ultra, doesnt work

[quaint rain]

thanks mate for confirmation

[flint summit]

My Ultra plan expired this morning. And with it about 6000 credits evaporated. Oh well, I expected that. I just didn't have enough good ideas to use them all up in time. That's why "doubling" Ultra credits from 12,500 to 25,000 doesn't really do me much good. I just can't use that many without a rollover of unused credits. So now I'm back on Pro with 1000 credits and that's fine for now. Maybe if they perfect Veo3 with the features we really need I'll go back to Ultra someday.

[flint summit]

Its notoriously bad at car interiors. It will put taxi passengers behind the wheel, or put the cab driver in the back seat or on the wrong side.

[sage rock]

Hello! Great, thank you very much for your answer. Greetings

[fallow sedge]

The “Pro” subscription is probably more appropriate than “Ultra” in that case. And if you've used up your 1000 credits, it is possible to buy more credits separately (I think 10 $US for another 1000 credits, or something like that).

[fallow sedge]

The “ingredients to video” feature does not support Veo 3. You can only use Veo 2 with that. If you need Veo 3 (because of audio support), you must use either “frames to video” or “text to video”.

That is independent of your subscription, i.e. it doesn't matter whether you're on “Pro” or “Ultra”.

[quaint rain]

em trying frames as well same problem

could create videos sometimes and then same error

i just created a video this morning then started to have the error again

[fallow sedge]

Please show screenshot.
I'm using “frames to video” right now with Veo 3.

[quaint rain]

was able to use it this morning

i thought if i upgrade it would fix it but thats not the case, poeple with ultra having this problem as well

[peak hatch]

I can actually use veo 3 with frames to video too

But not ingredients to video

[fallow sedge]

Strange … it's working fine for me, using it right now (and I'm only on Pro subscription, not Ultra). I guess that the Google Flow engineers need to fix a few bugs …

[quaint rain]

been stuck since last night

once u get the taste of veo 3, u can use any other platform 😛

[main saddle]

fwd from @sage rock 🙂

[silver bolt]

not long ago someone posted a very detailed and good looking prompt guide, but its gone. anybody got idea where i can learn mastering the prompting ?

[main saddle]

@crisp basalt posted one previously, was it this? #flow-general・🌊 message

[sage rock]

“Thanks for forwarding this, Georgia! 🙌 I actually had a similar post before and someone from the staff replied, so just pointing that out.”

[crisp basalt]

There’s also another in-depth guide (a bit more technical)

Google Veo: A Technical Deep Dive for the AI Director by Pier Paolo Ippolito (he's a GenAI Solutions Architect at Google)

https://medium.com/google-cloud/google-veo-a-technical-deep-dive-for-the-ai-director-a6fdf6438a3f

[crisp basalt]

Super handy and packed with good info

Skimmed the article earlier, and its actually really solid. They even included a “Challenge & Solution” section that walks through common issues and workaround etc..

[worldly yacht]

anyone got spare 20 credits to make 1 video for me please ?

[jade crest]

Will Flow have Banana?😂

[peak hatch]

I found your issue, i think you cant add camera movement on veo 3 yet if you add a frame

[peak hatch]

Good reads i think i vaguely remember your name from hailuo servers like 9 months ago

[sonic zealot]

https://www.tiktok.com/t/ZT64tqGCv/

Miniature construction workers sculpting an electrical marble of the logo G.S

[jagged jackal]

Nano Banana or Potassium?

[junior horizon]

you can only use first frame, are you also adding last one? if that is the case, there you have the reason

[quaint rain]

yeah buddy think found the problem can only use single frame atm for veo 3

[quaint rain]

was using 2 frames which is not available yet on veo 3.

[jade crest]

nano banana

[wicked ravine]

anyone else having problems generating images? keeps telling me i’m making too fast requests

this has been going on for a few days

[waxen bone]

Hi there! Why do I have a VEO watermark on a PRO account? How can I get rid of it?

[placid path]

Hello, need some help on creating video prompt about isekai anime. I'm new on veo

[orchid geyser]

You set a very good example! You defend the project as it should be! We try to bring up the improvements and you insult

You find it normal to pay almost $300 to have only 10 min max of video generation per month counting all the bugs and generation errors

the report?

But you're really a fool, this server serves to improve the service, so obviously we'll say what's wrong

Good for you

[frozen quail]

image to video doesnt work in gemini for the free 3 videos.

[river idol]

can somebody help me with my veo3 3 video please dm

[elfin dust]

Hello friends, I would like to know if it is possible to do this type of work with a photograph, keeping the elements as they are and just a type of timelapse.

A photo I took that I would like to transform.

[long prism]

Hello, I have made a commercial using VEO 3 / Flow. Where can I post it to get work / noticed ?

[tired oyster]

https://www.tiktok.com/t/ZP8BU3wkB/

Hello, who can help me get this prompt script please?

[mossy dew]

I would like to ask a question... I have some AI images generated using Flux from Black Forest Labs of a female singer. I were hoping to test out some images and run it though VEO 3 weekend trail but I am getting this message, below.

"We do not allow uploads of photorealistic people in your country or territory at this time. Please try a different image or send feedback."

I have attached one of the images to show what it looks like. The idea is to use VEO 3 to make a music video of a female singer in Jazz format

I know that I am able to generate the videos via other online portal like Repilcate.com but each clip costs 6 quid and I have about 20 to 30 clips to render out. This is to help with the editing stage of the project. I were hoping to work directly with VEO 3 main website like Flow to render out the clips but I am being blocked.

Could someone help to solve this issue?

Thank you

[kind bison]

They need to hurry up and release Nano Banana -I mean GemPix (I think that's supposed to be the official name - or something similar). Tired of being teased like this! Using it on LMArena is such a pain.

[idle depot]

https://vm.tiktok.com/ZNdXxNkjh/

[zinc ridge]

how to prompt better in flow? animation

[crisp basalt]

can u clarify what you mean by “prompt better”

Are u aiming to generate the video in a specific style, like 3D or 2D animation, or something else

[zinc ridge]

yes for 2d animation

i need to use the frame to video right

it's better

veo3

[crisp basalt]

uh that’s image to video yes if you want to make sure it actually generate video on that specific animation style using frame to video

[zinc ridge]

then i just need specific prompt

[crisp basalt]

Specific as if?

I don’t think you need to think much when using frame to video, you just write down the prompt as if like you’re telling / directing Veo 3 on how you want that image to be animated

[zinc ridge]

look at this

is this good

[crisp basalt]

Yeah it looks fine

Tho idk what’s your initial idea that you had in your mind for this so yeah but overall it looks fine to me

[rocky escarp]

this is the message i get when generating an image on flow, video generation is fine, i would appreciate any help

[crisp basalt]

I think it’s an ongoing issue right now? I can’t generate image either in Flow

[edgy thorn]

I usually organize per project, I make at least 5 to 8 projects a month, and since the prompt are made per project, the final prompts are not that important to keep, since I work every project with different perspectives, but Its a good idea to tag or even using Premiere as a manager, maybe bridge, or any other app to find with personal tags.

[mossy dew]

I read that changing your VPN wouldn’t change how you upload the images onto veo3.

I may have to use other portals to generate the clips. which will make the project more expensive. Maybe I’ll reach out to Google and ask if there’s a work around or something we can solve together but if this may be fruitless.

Thank you for your reply.

[mossy dew]

Try this?

https://www.reddit.com/r/PromptEngineering/comments/1mxb375/how_i_cut_my_ai_video_costs_by_80_and_actualy_got/?share_id=WXUgOoIph75RndacQp_6T&utm_content=2&utm_medium=ios_app&utm_name=ioscss&utm_source=share&utm_term=1

[limpid ember]

Hey everyone! I need to create some UGC type of ads for my business that's why I bought the Ultra plan in the first place but I can't upload images nor generate them here. The tool essentially is useless for me because of these two errors. Any help? Any other tools? Can I get a refund? I know Veo 3 is great but it's so confusing to use but I have a short deadline and can't do nothing with it

I'm based in Latvia by the way, which is in the supported country list and I also can't figure out how to generate videos longer than 8 seconds and in 9:16 format. I understand I need a reference image to be 9:16 but that crosses out as I can't upload any images.

[near zenith]

who can make a video for me via Frame to video AI, i will send the frame-image

[limpid ember]

Where can I find the instructions?

[elder knot]

Yo I had a question

I want to make an instagram post commercial for my story

but the size of the video or whatever you call it doesnt match and its like in the middle on the screen and the rest is black or like blurry anyone knows how to make it so thats the perfect size ?

[worldly yacht]

Ask the gemini to add to your prompt if its for instagram story , tiktok , i think its 16:9

[elder knot]

when I generate do I just add this like make sure the frame is 9:16 or whatever ?

[worldly yacht]

Copy your prompt you have written in Gemini , and add at the end - I want it to match Instagram or tiktoks scale

[worldly yacht]

For Reels, Stories, and Live videos: 1080 x 1920 pixels with a 9:16 aspect ratio. This is the full-screen vertical format.

For standard in-feed posts:

Square: 1080 x 1080 pixels with a 1:1 aspect ratio.

Portrait (vertical): 1080 x 1350 pixels with a 4:5 aspect ratio.

Landscape (horizontal): 1080 x 566 pixels with a 16:9 aspect ratio.

[worldly yacht]

Video Prompt: Dog in the Park (Instagram Reel)
Aspect Ratio: 9:16 (1080 x 1920 pixels)

Visuals:
The camera starts on a hyper-realistic, slow-motion shot of a golden retriever shaking its head, sending a spray of tiny water droplets flying in the golden hour light. The 9:16 frame is filled vertically with this detailed action. The camera then smoothly pulls back, tracking the dog as it dashes across a wide, sun-drenched field. The focus is razor-sharp on the dog as the background blurs, creating a dynamic, cinematic feel. The shot ends with the dog leaping into the air, its floppy ears and tail a blur of happy motion.

Sound:
A slow, dramatic synth beat begins as the dog shakes its head. The audio is layered with hyper-realistic sound effects: the sharp whoosh of the head shake, the individual plink-plink of water droplets hitting the ground, and the soft thud-thud-thud of the dog's paws hitting the grass as it runs. The music swells into a triumphant orchestral score, peaking as the dog leaps into the air. - It should end up being like this - sorry for the spam , tryna help 😄

[elder knot]

Thank you a lot man

[worldly yacht]

hope it worked for your vids man !

[peak hatch]

Any idea when ingredients to video releases for veo 3?

[elder knot]

I tried but it still doesnt make them in the frame

[kind bison]

What are you talking about, LOL? Google folks were teasing with banana emojis all over social media. They are building up the hype.

[worldly yacht]

ill give it another check later and see how to do it exactly and tag you

[forest pebble]

How is 1 video enough? I'm sad! 1 is WAY TOO SMALL!

Maybe I should type prompts in the flow prompts channel so someone can make videos for me if they have the money!

[forest pebble]

Wait, the channel's gone?

Is there a way to share prompts please????

[clear cipher]

which?

[forest pebble]

Easy ways to share prompts

[clear cipher]

what help do you need?

[forest pebble]

Why did the flow prompts channel change to tips n tricks? Now it makes it harder to share

[clear cipher]

Well, I don't know either

I thought it was help with your Opal or prompts

[forest pebble]

it's flow/veo

[clear cipher]

if you need help with this I am willing

[mossy dew]

There are other vendors like replicate.com where you can bypass Google directly, but it would cost much more compared direct to Google like flow.

[placid vector]

How do I get longer than 8 seconds like in these tik tok videos?

[soft leaf]

If I get a veo3 video with no audio, but it isn't tagged with "No Audio Generated" or whatever the usual phrasing is, should I report that as "Low Quality", or "Inaccurate"?

[vast crow]

What is the best way to pay for unlimited video generations? I’m using the Gemini Ultra plan rn and I’m still running out of credits

[fast forge]

this was probably asked a million times already but does anyone know if Google's current promo of 3 free VEO 3 videos this weekend also apply to accounts with existing subscriptions? TIA

[clear cipher]

Let me introduce you to my newest and most improved invention. I call it The Digital Drug.

[empty sail]

1. En la calma de una tarde, el profeta Jonás, hijo de Amitai, escucha una voz inconfundible: la voz de Jehová.

[icy dew]

Has anyone tried opal to create video? It's all over soc meds. I wanted to know how was the quality? I'm guessing it's not veo 3.

[worldly yacht]

Whats the prompt you used for the Hydrolic press and the crystal ball ? that was NICE

[clear cipher]

I did it with the prompt "sistemas".

[clear cipher]

I don't recommend it, focus on refining your prompt to get an amazing result in any AI.

[icy dew]

Thanks. I agree. With that i noticed json prompts gives better result.

[clear cipher]

Exactly, focus on refining and polishing your prompt.

I am willing to teach anyone for free

[worldly yacht]

Thank you i will check it out !

[clear cipher]

Of course, if needed, I’m at your disposal.

Sorry for the spam, my apologies for sending the same videos across multiple channels.

[worldly yacht]

i tried the Json method the natural language one , and they are always hit or miss

we cant use that in our country yet :/

ill have to check about that , not sure the reason whyh

Yeap true , hopefully we it too at one point , so im not waisting all my credits when a image to video can do more for the details and everything i need

ehh not really sure about that 😄 ill have to check it out tho

[clear cipher]

What matters is the process of refining the prompt.

[worldly yacht]

not too sure 😛

[clear cipher]

focus on refining your prompt

[worldly yacht]

Not sure if this is allowed , its a regeneration of Deadpool after getting his head smashed by the hulk

The joker DNA card

[minor rune]

https://x.com/joshwoodward/status/1959411289622770125?s=19

Josh Woodward (@joshwoodward)

@GeminiApp Free Weekend expands!

A lot of you asked us to double the Veo limits for the next ~24 hours, let's do it

Free tier: 3-> 6 pack
Pro: 3-> 6 per day until 8/24 at 10pm PT
Ultra: 5-> 10 per day until 8/24 at 10pm PT

X · Aug 24, 2025, 12:23 AM

[fast forge]

got my question answered https://x.com/GeminiApp/status/1959408375869190466?t=BiYsZ3INM1I_IWpJndm5Pw&s=19

Google Gemini App (@GeminiApp)

Good idea. Let’s double it for everyone.

This weekend only, ending tomorrow (8/24) at 10pm PT:

👉 Gemini free users, you now have up to 6 free video generations for the weekend.
👉 Pro users, your pack has doubled to 6 per day for the weekend. 
👉 Ultra users, your pack has

X · Aug 24, 2025, 12:12 AM

[clear cipher]

Yes, Google's current promotion offering 3 free Veo 3 video generations this weekend applies to all Gemini users, including those with existing subscriptions. According to 9to5Google, free users can generate three Veo 3 videos by Sunday at 10 p.m. PT, using the Veo 3 Fast model, which is faster and maintains the same 720p resolution.

For those with existing Google AI Pro or Ultra subscriptions, you can continue to use your daily allotment of three Veo 3 Fast generations per day. The free weekend promotion does not affect your regular usage. However, if you're a free user, this is a limited-time opportunity to try out Veo 3 without any cost.

To access the free generations, open the Gemini app and look for the “Video: Generate with Veo” option in the prompt bar. If it's not immediately visible, tap the three-dot menu to find it.

This promotion is set to end on Sunday at 10 p.m. PT (1 a.m. ET / 6 a.m. BST / 3 p.m. AEST on Monday).

[lavish field]

lots of issues with no Audio now , same here , which kind of makes VEO loose its edge on the other AI Genereators out there

[clear cipher]

Yeah, I was having that problem too, the audio wasn’t the same quality as the video, I went crazy trying to fix it.

[lavish field]

im trying to work around it now but all i see are wasted credits ..

[boreal willow]

How do u manage oversaturation when creating from photogram to video??? It started with this saturation, shadows ... and ends up with this. Even when I explicitly say in the prompt to change anything in his aspect it does .... I'm just putting some dollars in the trash

[clear cipher]

Because you need to focus on refining your prompt to get a good, controlled final result and stop wasting credits.

[lavish field]

lol yeah im not new to this , i have even given a testing prompt for example Ultra-realistic close-up of Mick the scruffy grey koala holding a stubby of beer.
He leans forward and mutters in a slow, gravelly Aussie drawl: “Got a beer, mate?”
This stripped-down version is great to test if VEO is even in “speech mode.” If you get sound here, you know the longer burrow description was choking it out. maybe ..I also tried using 'Audio -Mick etc...but sometimes Veo it often ignores it because “Audio -” isn’t a format VEO recognises. It treats that part like extra description rather than dialogue. sometimes though it works ...very incosistant of late

[clear cipher]

The audio is generated within the first 2 to 4 seconds of the video, so if you don’t have one defined, it really lowers the quality.

[lavish field]

yeah but i think it just glitched out , wasn't even saving scenes, restarted relogged and scnenes were back ..see if the sound kicked back in

yeah just a glitch but expensive one lol wasted test prompts ..just had to log out

[worldly yacht]

Because the sound was messed up , i add some from clipchamp

[verbal roost]

Try this prompt mate

Medium close-up of scruffy grey Koala, wearing a mechanic shirt, namtag reading "Mick" holding beer bottle, looks at camera and says "Got a beer, mate?" in slow gravelly Australian accent

[lavish field]

thanks champ, I got it going again had to log out and back seems Flow gets “stuck” in visual-only mode and won’t trigger speech until you restart fresh.

[boreal willow]

You're right. Is google thinking to lower a bit the prices any time soon? Do we have any news a of any coming release live Veo 4 or something similar??

[clear cipher]

It’s always good to clear caches.

[lavish field]

yup

[safe mango]

How to buy credit for flow ? I spent on my all credit. My plan is Ultra

I get error of 500. a message is "An error has occurred.
This page cannot be loaded."

[kind bison]

You really don't know what you're talking about. So, you're getting a block from me.

[autumn sorrel]

Has anyone managed to create videos in -90 camera degrees using ingredients to video, if so please share the prompt I want to make it real and through editing rotate the video and go from 16:9 to 9:16

[mossy dew]

Can’t you just rotate it in post production by rotating 90 degrees?

[subtle hearth]

Hey I can make a better ai video that any of you guys

Yes sir

[autumn sorrel]

Has anyone managed to create videos in -90 camera degrees using ingredients to video, if so please share the prompt I want to make it real and through editing rotate the video and go from 16:9 to 9:16

[jovial tendon]

the flow filters are way too strict

[idle depot]

https://vm.tiktok.com/ZNdXsb1wY/

[topaz trellis]

SCENE INSTRUCTIONS:

• Type: Worker perspective
• Framing: Horizontal 16:9
• Camera: Static
• Lighting: Soft neutral
• Background music: Subtle tension
• Emotional tone: Serious and empathetic

SPOKEN TEXT (Portuguese - BR):

"O trabalhador sente o peso de salários baixos e da falta de valorização. Mas isso não significa que o empresário seja o inimigo."

AUDIO SETTINGS:

• Language: Portuguese (Brazil)
• Speed: 1.3x
• Tone: Serious and empathetic
• Background music: Subtle tension
• Singing/Rap: No

FINAL CHARACTER LOCK REMINDER:

Do not include any on-screen text, captions, camera UI, or overlays.

[faint adder]

Hello can you help guys?

I bought pro of google flow and i am using it in my iphone but when i try to download the video made by flow on google website
It doesn’t download

[south magnet]

I think it's related to how iphones download files. I don't think this is best place to ask. Try to find solutions on the internet

[ripe umbra]

PHP: Version 8.3
PHP Extensions: gd, dom, xml, curl, and others.
Composer: The standard PHP dependency manager.
3. Troubleshooting Steps Taken:

We have spent a significant amount of time trying to set up the environment. Here is a log of our attempts and the results:

Initial State: The environment started without PHP or Composer installed.
PHP Installation: We successfully installed php8.3-cli after the AI diagnosed it was missing. php --version confirmed PHP 8.3.6 was present.
Composer Installation: The AI successfully installed Composer locally as composer.phar.
Dependency Installation Failure: The AI ran php composer.phar install. This command failed, reporting that several required PHP extensions (ext-gd, ext-dom, ext-xml, ext-curl) were missing.
Extension Installation Attempt: We provided the correct command to install these missing extensions: sudo apt-get install -y php8.3-gd php8.3-dom php8.3-xml php8.3-curl. The command appeared to run successfully in the logs.
Verification Failure: Despite the successful-looking installation, php composer.phar install failed again with the exact same "missing extension" errors.
Root Cause Discovered: The AI investigated the PHP configuration directory (/etc/php/8.3/cli/conf.d/) and confirmed that the .ini files for the extensions we tried to install were never created. This proves that the apt-get command, while appearing to run, is not actually configuring the environment correctly.
Final Block: The AI is unable to manually create the configuration files because it does not have permission to write to the /etc/ directory.

Thank you for your time and assistance.

[formal pier]

why can't i upload an AI generated photo of human to flow?

it says can't upload real people in your region

[spice phoenix]

https://youtu.be/f9TK_WkK2MM - Ep 6 Ihope these are getting better

[tough dome]

Same thing for me, Google was supposed to do something about it, but still nothing.

[late rampart]

Hello community, does anyone know how to contact Flow support? I’d like to find a way with them to remove the restriction on realistic characters. When outside of America, this restriction heavily blocks creation. So maybe with an agreement that clears them of any responsibility, they could unlock it for my Flow account or set it as if I were based in the U.S.

This restriction is greatly hindering my projects and I am losing a lot of time on my work, even though I have great projects to develop.

[cobalt junco]

image generation in flow is broken

[late rampart]

It's annoying, especially since it refuses me a model that has nothing human about it.

is not a humain hahaha

Yes is true, but it's also not normal that this model is refused, the security is probably too strong.

You're right, it's an exceptional tool, I can't wait to see veo4 which will undoubtedly have context memory, which will allow you to avoid changes as soon as the face or an object goes out of frame.

I'll figure it out and try pointy ears etc. I put it on blue skin but it's ugly thx for your help

[late rampart]

there is no sound, it's less good

[tawny ice]

Anyone know how to prompt for an extremely wide angled aerial shot I want to camera to look as if it was shot from a helicopter or drone following a character across a wide open space, so far the results haven't anywhere near what I want as far as camera angle

[late rampart]

but on veo I also start from an image, then I describe the scene and the elements, the only advantage is that veo3 generates the sound and the voice and behind I do dubbing with lip sync

[clear cipher]

ULTRA-CLOSE MACRO. A hyper-detailed, translucent emerald green glass kiwi, subtly etched with tiny obsidian seeds, rests perfectly centered on a polished light maple cutting board. The sleek stainless steel knife, gleaming, hovers infinitesimally above it, generating a barely audible friction. With precision, a soft, crisp metallic tap echoes, followed by a brief, ethereal resonance, preparing for three deliberate slices.

Tension builds as the blade slowly, deliberately presses into the glass. A delicate, almost imperceptible internal crackle begins, hinting at the gradual fracturing. The blade glides with continuous, gentle friction through the yielding glass. A satisfying snap is heard as each segment finally separates, followed by a delicate cascade of tinkle sounds from shimmering crystalline fragments scattering, catching the soft, diffused light. Ethereal 'glass smoke' drifts from the freshly cut lines. The three perfectly formed segments settle softly on the board with a gentle thud, each piece emanating an internal, ethereal resonance. This immersive ASMR experience, captured in ultra-realistic 8K, focuses solely on these pure ASMR sounds – crisp metallic tap, continuous friction, clear resonance, satisfying snap, gentle crackle, delicate tinkle, and soft thud – with absolutely NO human speech, voiceover, or spoken words at any point. No background music or jarring elements, ensuring a consistently smooth and reassuring auditory and visual experience for deep relaxation.

[clear cipher]

This here is his final prompt. So, you see it’s not about the length of the prompt, but rather the refinement of the information, because if you look closely at the prompt, you’ll see that the information is refined, not just thrown in any random way.

[clear cipher]

I had this result, and this whole prompt was generated with just “glass kiwi”. And that was it it produced all this. So the refinement process has to be the main focus. The video generation itself should be the last step. Because, in this case, you need to refine your prompt so that the quality is good in any AI.

Me, personally, my goal is not being able to reproduce the same quality in different AIs. Like, not focusing on just one AI. For me, the result is only good when I can reproduce that same quality in another video-generating AI.

For example, this same flow I showed that generated this video I realized I could make one more upgrade to it, which is what? Adding a secretary. So, in addition to having agents that refine the prompt, it will have secretaries to help them refine it. That way, I’ll get an even more refined result and be able to reproduce the quality in any AI.

[jovial vessel]

what happened why too strict now?

"The scene starts with her moving closer to the camera to examine the white stuff on her cheek. She says, “They’re so much emptier—do you see that? Do you see how good they look?” She stretches her cheek to let the camera get a closer look at her pores."

what's so wrong about this simple prompt?

[fallow sedge]

“white stuff on her cheek” is probably too naughty for a U.S. video AI. 😄

[jovial vessel]

I removed it yesterday, I even shorten the prompt to "she moves closer to the camera", that very simple prompt and I still get the error? what the heck? the image is just a face of a person. this is so frustrating, our team spend thousand of dollars on this tool and giving bad results now

[clear cipher]

It’s not refined.

[late rampart]

Am I the only one who doesn’t understand left and right in VEO3 prompts? When I tell my character to grab something with the right hand, it grabs it with the left hand. But when I give a very simple action, like “Raise your right arm,” it does raise the correct arm. I don’t understand why VEO behaves like this.

How can I make it understand what’s left and right?

[clear cipher]

Include this in the refinement process

can I see the prompt?

[late rampart]

2000+ caractere

[clear cipher]

Even so, it doesn't have control over a simple camera, I'll review it

[late rampart]

It is possible to control the camera after you have to know which way it turns, spectator or character

for me is spectator

[clear cipher]

Enter your detailed fantasy scene description. Let me test it.

[late rampart]

image to video

[clear cipher]

What did you think of the result?

[clear cipher]

@late rampart

[clear cipher]

I'm not sending the link here because I don't know if the person approves or not.

I always send no DM

[clear cipher]

Explaining a bit better, now it takes any description you create not just about that character and generates a video based on the prompt you sent.

Now with a good description, you can get a refined prompt for anything.

[clear cipher]

another example

[late rampart]

not bad i love this gun hahaha

[late rampart]

its VEO3 but whit Opal API ?

[sleek prawn]

I've been teaching myself Veo3 by creating a stylized graphic novel movie trailer about a story I created about my wife as a mild-mannered bartender named Camille who steps into a thrift store, finds a mysterious cowboy hat, tries it on, and gets transported into an alt-universe cowboy superhero world where she emerges as "Cami the Kid," a hero known to the townsfolk, though she has no idea what's going on. As she's taking in her confusing welcome, a shadowy villain welcomes himself and confronts Cami, suggesting that she lost her memory. He blows a dark blowing bubble and when it pops, a cascade of memories from various adventures as "Cami the Kid" rush back. Among the final memories in the montage is a dark version of her who seems to be commanding a legion of shadowy humanoid cowboy creatures, a town on fire with the shadowy man holding what appears to be a person, and a shocked Cami, watching as the town burns. When the scene returns to the present, the Dark Cami reveals herself in the shadowy manly figure that confronted Cami the Kid. She doesn't know it, but the line delivered by Dark Cami suggests that she may be Cami's "other half" that summoned her to this world to finish business. I haven't finished it yet, but we'll see a hip-shot of our hero hovering her hand near the revolver in her gun belt. A pause. Then a decisive grab. Black screen. BANG. "The Adventures of Cami the Kid" logo appears.

Here are some stills from it. I'm incredibly excited to see how it's going. Of course the character consistency is an issue, but I'm taking it as a limitation I'll lean into. I will clean up some shots, and add AI voices from Eleven Labs or something better than what Veo3 can do. I'd like to throw it all into an editing software.

Pt 2 of 4

Pt 3 of 4 (the bubble pop, and the memory montage)

Pt 4 of 4 (the return, the transformation, the showdown). very rough right now. Mostly test generations. Have to clean it up to get consistency in wardrobe and generate a few options that I can use.

That was a lot, but I'm excited about it and would love to get any feedback or ideas for anyone who has edited these generations in a premiere or final cut (or had success with imovie).

Thanks!

[unreal viper]

REMINDER! #flow-videos・📹 for images and videos and #flow-general・🌊 for conversation about flow! thanks!!

[pure frost]

Did you build an automation to generate the same kind of videos as the glass fruit cut? which software do you use to create the automation?

[unreal viper]

#flow-tips-n-tricks・✏️

[sudden condor]

Sorry didnt see that thank you

[pulsar crescent]

Can anyone tell me why sometimes the sound doesn't play on my videos? Lips are moving but no sound. And why am I charged credits for these unusable videos???

[pure frost]

sometimes it happens to me as well

[cobalt forum]

Has anyone else noticed that they can't get the "Jump to..." feature to work? It used to work well, but today, it actually clones the first frame of the previous video rather than starting with the last frame. So it's completely broken.

[kind bison]

Yeah, I posted about that in the bugs-feedback channel the other day - it stopped working for me.

Also audio quality - especially with dialog - has gotten bad this week, with lots of noise and digital artifacts.

[tough otter]

Did the website kinda outdated now? Why the flow keep generating wrong phone looking (as example,google pixel 10 series or iPhone)

[olive obsidian]

Hi

[uncut marsh]

hi i can no longer generate a text to image anymore- it keeps coming up with you are generating prompts to quickly. which in this case im not. anyone else having similar issues?

can someone reply if there is a problem with text to image please?

inside flow platform

i asked in the wrong general chat it seems a few days ago

[uncut marsh]

I'm spending like 20,000 credits most days but this issue is starting to become serious as i now use another platform for text to image prompt but now it wont let me even upload most images as they say it violates their policies. it doesn't at all. 😢

[pine tendon]

Same problem.

[fringe ferry]

Hey if i want to make videos, but my country has made that i cant generate using realistic humans, any work arounds i can do?

[small venture]

does anyone have a good example of using veo3 for a software product?

marketing it

e.g. releasing a new feature or doing a product launch

[twin palm]

Don't Flow only gen video, do it?

[soft leaf]

Does anyone know if Flow accepts negative prompts? I've seen some examples with "Negative Prompt: <...>" at the end, and I thought that had worked for me, but lately it doesn't seem to? i.e. it's treated as a positive prompt. Not sure which version is real and which was my imagination.

[peak hatch]

there was that superbowl add but im sure people use it professionally but they are probably lowkey about it, general masses dont like AI

[clear cipher]

I created a system, my own custom flow, that does this for me it has agents that refine a prompt by adding psychological triggers.

[peak hatch]

@clear cipher theres a 150k$ agent competition for video generations going on, i dont know too many details but i can direct you to where it is

[clear cipher]

Where?

[peak hatch]

ill dm it its a competitors discord

[clear cipher]

I want to participate

[clear cipher]

yes

[pure frost]

I'm just getting into ai automation so I was wondering where these flows can be created, I saw something like n8n

[clear cipher]

Thank you so much for the heads-up I’ve already submitted my application and I’m already in the process of creating the agent.

[clear cipher]

I recommend you use Opal. Like you said, you’re just starting out, so I suggest using it because it’s much more flexible.

[pure frost]

Really appreciate the advice 🫰🏼

[clear cipher]

So, if I disappear or don’t reply to any of you, that’s the reason.

Haha

[clear cipher]

And if you need any help, support, or assistance, my DMs are open.

[pure frost]

Thank you so much!

[clear cipher]

So, if anyone needs help or anything like that and asks me here but I don’t reply, try sending me a DM.

[clever axle]

Flow is doing this with my character, it is so bad that i am not even mad

[wicked harbor]

Hello All, first day on google labs!

[scarlet tide]

welcome here! hope you'll have fun!

[bitter halo]

Seen a lot going round on TikTok of the fruit eating fruits generated on veo. Has anyone got any accurate prompts to generate similar things. Gpt and Gemini can’t seem to get the prompts correct

[blissful gulch]

hello also to you from problems with the text to image option?

[wind sparrow]

i have a question to all of you, flow text to image makes me a video of 7 seconds, i can have more seconds or that is the maximum?

[remote crypt]

I'm hoping Veo 3 will soon have the possibility of extending videos. Right now, it's only Veo 2 that can do this?

[hasty herald]

I'm looking for the funniest piece of flow content out there. Drop the links!

[edgy sun]

I noticed that too, it redirects me from Veo3 fast to Veo2 fast and causes the audio to be lost. Previously I could extend them, hope this is just temporally. And for the Jump To option redirects me to Veo 3 quality and that's too much.

[clear cipher]

bro, i finally managed to submit that project lmao
after a thousand bugs, video crashing, upload fails, file not supported, almost gave up but it’s done
it’s up, submitted, now we wait (unless the site breaks again lol)
if it all goes wrong just pretend you never met me 😜
thanks for the support as always fam

The project was, of course, a prompt refiner. I wanted to add more details to it, but I can’t. I don’t even know if my registration will go through. But it’s a prompt refiner, where it takes any type of prompt and polishes it to the point of giving real quality to that prompt.

This was the video I used as a quality demonstration.

[heady umbra]

how do we correctly report video that has no audio but should?

8 of the last 12 videos generated had no audio

seems higher than usual

[wicked jackal]

Hello. Can someone tell me approximately how many videos you can create with 1000 credits?

[heady umbra]

depends on the model chosen, there increments of 10. 20, and 100 so math it like that

[remote crypt]

Flow needs a full video editing suite, that includes a connection to 11Labs for audio.

[peak hatch]

Do you guys find it much better to use veo 3 over veo 2 if you dont care about audio?

It looks like it has better quality but i dont know if its placebo when i go on a arena site its 50/50% for me

[late rampart]

Has anyone found a way to bypass Flow’s security in Europe regarding “photorealistic characters”? Or otherwise, does anyone have Flow’s contact? I’d like to propose signing a waiver so that they could authorize me to create without this restriction.

I have a problem in France with this damn flow security so if someone has a secondary American Google account I am ready to pay by PayPal so that he buys some and I will use his account to generate my videos quietly. You would make my life. But is it allowed?

[remote crypt]

Flow told me that this was a photorealistic character, so I wasn't allowed to upload. 😛

not sure if I should laugh or whatever

but the photoreal filter is definitely wonky. 😉

[remote crypt]

I dunno, I find my Russian childrens tv from 1970 kinda funny. But I have an odd sense of humor.

you can find it in the gallery channel

[late rampart]

i teste and yeah the security target this image LOOOL

[verbal roost]

looks like my grandfather, seems real to me

Want to see what grandpa-pa looked like?

[late rampart]

is creepy for me :x

[kind bison]

I hope the "no prominent people" thing goes away soon. Today I was hired by a very well known person - not an A-list celebrity, but certainly famous - who wants me to create a series of videos with him as the main character. If images of him get flagged, it's going to be a major embarrassment for me (plus a loss of money). We need to have some way to prove that we have permission to use prominent people's likenesses.

[fluid sequoia]

@kind bison message me ill help you out 🙂

[regal pewter]

Higgsfield AI

[kind bison]

Yeah, I have gotten around the "no prominent people" thing on some 3rd party platforms - but recently I had one get blocked (not a prominent person - completely AI character), so it might have spread to them as well.

[remote vortex]

Same here in UK

[remote vortex]

That's actually mad, i can't even use images I own the rites of. That I took myself. Have even arranged usage rites with the models.

[vapid rover]

NARRATION IN PORTUGUESE
"Ji-hoon finally takes Soo-min to a quiet garden filled with cherry blossoms. Under the falling petals, he confesses: ‘There’s something I never told you. I’m the heir of a large company… but I don’t want to live that life.’ Soo-min’s eyes widen in shock, unsure if she can trust him. The scene closes with her torn between her heart and her doubts."
NARRATION IN PORTUGUESE

[vocal pasture]

hi do someone knows why i get this message ?

[clear cipher]

Yeah, the narrations end up sounding just like those TikTok ones.
Loll

[mystic forge]

Yea how can you stop that cuz it's annoying 😑

[clear cipher]

Alguém e brasileiro?

[glacial oxide]

@grave saffron Please dm me, i think your assistance would help me.

[remote vortex]

Hi i was politely asked to post here instead of 'general'

Have been working on flow for the last three days now with mixed results.

I'm getting a continuing message that states

"We do not allow uploads of photorealistitc people in your country or territory at this time. please try a different image or send feedback"
I'm in the UK

I get this when attempting to upload images I own and have taken, as well as images that have been generated with prompts made with gemini visual. Uploading these images to use as ingredients in scenes and visual prompts.

I've even mimiked the images used in the demos and it won't allow me to upload giving the same error message.

Surely this is a fault ?

Am I missing something ?

(Hovering over subscription cancellation)

[verbal void]

Really hope this gets fixed ASAP

I'm optimistic that gemini and Google can Unlock the possibilities shown in their own demos to people in countries like UK, if its allowed in USA then it should be also allowed in UK, hopefully soon

[clear cipher]

Well, honestly, I didn’t understand this new update. So here are two examples for you. I want you to guess which one I made with the new update and which one I made with ChatGPT.
image I used as base

Whoever gets it right, seriously, I’ll give you a demo account I have with a thousand credits to use on Flow.

Let the games begin.

[clear cipher]

somebody?

[peak hatch]

https://tenor.com/view/sad-gif-7523306793289960933

[clear cipher]

the choice is yours

[peak hatch]

Since the left one is bad the right one is chatgpt

[clear cipher]

mark your choice

[peak hatch]

Im too easy to distract 😄 i gotta focus bro

[remote vortex]

Thank you, realy appreciate the clarity. 🌟

[clear cipher]

delivered ok

[cobalt junco]

ive had that for a long time in imagefx

not that particular msg of course, but the try whisk button itself has been there a while

[analog sedge]

The right one is extremely low quality so that is gemini

[clear cipher]

it's over 👀

[peak hatch]

How do i access the new gemini image editor just announced ? through gemini chat?

[clear cipher]

even for me

[noble carbon]

Veo 3 - Fast uses 0 credits for Ultra subscribers

anyone see this for veo 3 fast with ultra plan or am i trippin?

wait it works im not paying any creds rn 😮

[hasty island]

I need some help with Veo3. I’m trying to make a video with a top-down shot of a glass with ice, something pretty classic. But whenever I write “overhead shot,” or Top-down it doesn’t understand and generates the shot from the front of the glass instead. Has anyone managed to do this already? Do you have the right camera prompt for it? Thanks!

[sudden condor]

Tome to flow

[barren terrace]

ভোরবেলা সূর্যের প্রথম আলো জানালার ফাঁক দিয়ে ঘরে ঢুকে পড়ল। মিষ্টি বাতাসে ফুলের গন্ধ, পাখিরা চিঁ-চিঁ করে গান গাইছে।
আজকের দিনটা আলাদা, কারণ সারা সপ্তাহের কাজের পর অবশেষে অবসর এসেছে।

রাফি বারান্দায় বসে চায়ের কাপ হাতে দেখছে চারপাশ। সামনের বাগানে লাল-হলুদ ফুলগুলো দুলছে বাতাসে। মনে হচ্ছে যেন প্রকৃতি হাসছে।

সে সিদ্ধান্ত নিল আজ দিনটা শুধু সুন্দরভাবে কাটাবে—
📖 একটা প্রিয় বই পড়বে,
🎶 একটু গান শুনবে,
🚶‍♂️ আর বিকেলে বন্ধুদের সঙ্গে পার্কে হাঁটতে যাবে।

দুপুরে মায়ের রান্না করা গরম ভাত আর ডাল-ভাজি খেয়ে মন ভরে গেল। তারপর একটু দুপুরের ঘুম।
সন্ধ্যায় আকাশ ভরে উঠল কমলা-লাল রঙে, আর সেই মুহূর্তটা যেন দিনের সবচেয়ে সুন্দর উপহার।

[muted matrix]

It's great to see more options with the Text-to-image generator with flow, it will be awesome if they're planning on adding Gemini 2.5 Flash Image as well.

[thorn wave]

I asked Gemini to create a json script as if the camera is looking down into a glass of ice and ran it through veo3

[worldly moth]

Just choose.

[hasty island]

wow can you show me the prompt of the first movie? please, thanks in advance

[turbid lintel]

now veo3 text to image generation have reference option

now veo3 is cooking guys

[hasty island]

i did it in gemini and it worked, do you think is better than chatgpt to work with prompt generation?

[soft leaf]

Is it just me or did the button for reloading a prompt from a previous render disappear?

I have something instead that looks like a button labeled [Extend], but it isn't a real button.

And then the only other choices are flag and delete?

[fluid sequoia]

is Veo 3 fast free for ultra users rn?

[tardy ridge]

is veo down?

[tardy ridge]

thats probably why its not working right now lol

[fair bone]

I am sooooo tired of this error

it is happeing with multiple animation types

Does anyone know who I can contact to get help? I am on ultra subscription and this is stopping my workflow.

[noble carbon]

is a 17 min export possible

been trying to download for over an hour

[wicked sapphire]

Hey- in flow, as early as yesterday, there was a button to reuse prompt. Now thats changed, and it doesnt look like i can do that anymore. Is this a glitch?

[wicked sapphire]

exact same issue

<@&1173646484311384114> is this being fixed?

it seems like new prompts can be reused, but yesterdays cant. You cant even select the whole text from those historic prompts to rebuild

the text field collapses when you try to select the text

Im in the middle of production. would be helpful to know when this will change

[soft leaf]

Same. I am able to select the text, but the linebreaks are removed, which makes them hard to work with.

[foggy rapids]

Guys i need help

[idle compass]

me too - soooo frustrating...

[twin mortar]

Everybody else using Ultra getting charged 0 credits for Fast video generations?

[fair bone]

it was working sometimes so i thought it may be an error but i got what the issue was

my account was made in EU , US accounts dont have the restriction

[fallow sedge]

The fact that the filter is currently overly strict is an error. This has been acknowledged by Google (see the bugreports channel), and they stated that the team is working on a fix.

[fair bone]

yeah, my videos were just anime characters not realistic

and that too in every different style i tried

[fallow sedge]

I just quoted the Google team. They acknowledged the behavior as an error. Don't get me wrong, but I think they know Veo better than you do.

[clear cipher]

Hey everyone, just a quick heads-up for those who usually reach out to me for help or support. I was banned from the Google Gemini server, so right now I'm only active here. If you see anyone claiming to be me over there it's not me. Just wanted to make sure no one gets confused or tricked. Thanks!

[clear cipher]

Just to be clear
I’m not part of any official Google team or support staff. Everything I share is based on my own knowledge and experience. I heard that my previous ban may have been related to this misunderstanding so I want to make sure there’s no confusion here. I’m not affiliated with any official support, just sharing what I’ve learned. Thanks for understanding!

[tired idol]

Great news, Ultra subscribers. no more credits are needed for veo 3 fast. make as much videos as you want + you get the 25000 credits

[terse lodge]

Why haven't I gotten my monthly credits replenished?

[thorn wave]

I just started using JSON scripting within Gemini. I have baselined it against chat. From a couple of videos I watched on it, they claim Gemini handles JSON better.

[zealous sand]

A cinematic logo motion for a luxury photography and videography studio, using the provided golden ribbon-style logo for SIAVASH Photo & Video Studio, involves starting with a pure black background. The process includes a glossy golden ribbon entering smoothly, twisting and wrapping in 3D space to form the stylized 'S' shape of the logo. Subtle light flares and reflections add a luxurious touch as the ribbon moves. Towards the end, the camera shutter symbol inside the 'S' shines with a soft golden glow. Finally, the text 'SIAVASH' fades in with golden futuristic typography 1 . This motion design aims to convey elegance and professionalism, enhancing the visual appeal of the studio's brand 1 .

[meager edge]

how to make character consistency without morphing the face

[hot silo]

Same!

[swift quartz]

Not a great perk if all my generations get the “No audio generated” error

[tired idol]

it happens but not often, hopefully they'll fix that

[deep vessel]

@signal crow Hey can anyone guide me how I can be ai researcher

I am about to start my degree

[warped crow]

I have lost so many credits I did the subscription when the video do generate I don’t give back credits I went through customer service and sent video after video after video and nothing they just keep asking for the same thing I sent 🤦🏾‍♀️

[thorn wave]

Yes!

[limpid ore]

Hello, I'm creating shorts for YouTube using Flow, but I don't know how to create videos in 9:16, can anyone guide me?!

[sonic cedar]

Keeps getting stuck at 99% when extending, anyone else?

[idle depot]

https://www.tiktok.com/@skan.ai/video/7543048814693797142

[silk nexus]

How do we get in touch with Google as I've just been using Flow and some videos create whereas others get to 99% and then just hang forever and I left them for a very long time (probably around 20 minutes) but then got fed up and refreshed the project page and they were no longer in there (the prompts or the progress) but I've still been charged for them!!
That's just not on. I need a refund but where do I go to do that? Thanks.

[elder elbow]

Go back and check your library. I was having the same problem but found them in my library page.

[silk nexus]

Nothing there unfortunately 🙁
Actually I've just now got a message saying that Flow is temporarily unavailable so I reckon there were problems with it definitely. I'll have to contact Google Support to see about getting some refund as I'm not having that.

[main saddle]

hey! thanks for flagging this, super helpful. I'll pass along to the team, rec also reporting in #1319424187730755636 if you haven't already

[main saddle]

we're investigating this!

[silk nexus]

Ah so this Discord is actually setup by Google? Excellent. I thought iwas maybe just someone had set it up for people to help each other out. Can't remember how I added it now so it probably did say but I’m getting old and forgeful! 😄
How will we go about getting a refund as usually when the videos fail the credits go back on again but these didn't so I don't know how you'll know how many credits to give back? Thanks.

[main saddle]

this is a realio-dealio Google server! if you see any more weirdness with Flow, please do report in #1319424187730755636 as we have a team who are fielding and reporting these to our eng teams.

for credits, I'd rec reaching out to Google One Support as we're unable to directly assist with accounts/credits thru Discord. keep me posted on what they say, pls!

say that 5x fast

genius.... almost too genius

everyone can take turns in voice lelll

you will go first! 🔥

[sacred dawn]

Is there any chance of a decent update on photos and realistic people in the uk and eu @main saddle

[silk nexus]

Okay thanks I'll do that and I'll get in touch with support about the credits. Thanks again.

[main saddle]

I am going to see what I can do 😂

[silk nexus]

And being able to use Frame To Video so we can upload our own images? We're perfectly legally allowed to upload images of people that we have permission to upload here in the UK and there are laws that if we use anyone's photo that we don't have permission to use or we use a famous person or do something bad then we can already be prosecuted for any of that but just blanket banning uploading images of people is really stopping creativity. And when we upload an image of a person that has been created by an AI image generator then that's an extra big kick in the face!

[sacred dawn]

@edgy holly OpenAI does it already with sora

Veo2 isn’t

[silk nexus]

Sorry you said "You are" but I'm not sure what you were referring to when you said that? Currently I'm not able to upload any images of people at all to Google Flow, even AI generated images of people?
As to our laws like I say, we do already have laws in place that say what we can and can't do and if I say wanted to upload an image of myself then I'm not allowed to but I'm perfectly legally allowed to if Google would allow it.
This just seems like a blanket ban just to maybe make things 'easier' for the admin side?

[sacred dawn]

100% Mark, let’s be blunt we pay the same

@edgy holly it’s a bit late here so debating will not be great. BUT it’s been relaxed in the API on vertex

There is no reason it shouldn’t be relaxed on flow and Gemini

Maybe, true, maybe not. But if a google workspace ultra user executes on flow direct all legal protections including an nda are already in place

[silk nexus]

It's late and I'm getting confused but you're saying those things are legally allowed so I'm wondering why in the UK we're being prevented from doing them then?

[sacred dawn]

That’s the million dollar question mark

[silk nexus]

Not really. There's nothing in our laws that says that I can't upload any images that I have the legal right to use. So images of myself, my family (with their permission) and AI generated images are all fine but we're prevented from doing so.
That's nothing to do with our politicians or laws as we don't have anything legally stopping us from doing whatever we like with images that we have the legal right to use.

[sacred dawn]

Then we should not be charged full price if the features are not the same

Your wrong, the laws are in place adherence to them well that’s something else

And don’t get me started on our politicians lol that is one area I’m not going into

Brats?

Bit low reall

Really

[silk nexus]

I agree with the features as DisruptLIVE mentioned above. When I purchased the subscription it didn't say anything about certain features not being available in the UK. I got in touch with Google support to see if I could get a partial refund but they said absolutely not.
The problem here is that in this country (and I suspect it's probably more or less the same in the US) is that if there is information which would be pertinent to a person making an informed decision regarding a purchase then legally it should be mentioned. It wasn't mentioned and over here in the UK that's definitely not legal to do.

[sacred dawn]

Mark don’t bite

[silk nexus]

I'll ask them again more forcefully this time then and see if they'll do anything.

[sacred dawn]

No call out of restrictive features no confirmation just warnings paying the same no desire refunds and slated in the server as entitled and a brat

The server of the provider

I think the term used was “entitled brats”

[silk nexus]

I'll admit I'm very confused then. I thought we were being prevented from uploading images of people here in the UK due to Google thinking we have some law that says we're not allowed to. If it's nothing like that then why has it been prevented then?

[sacred dawn]

That’s what I was asking mark

Someone to look at all the big reports and do a decent escalation so we get a answer beyond “we are working on it”

Bug ^

[silk nexus]

I'm getting old and definitely tired tonight so unfortunately still not following what you're saying at all. It's probably not you just me being tired though.
You're saying they're responsible for what THEY do? I don't quite get what you meant by that?
I know we have the new safety act that came into force over here but none of that (or any other laws) prevent anyone in the UK from using an image they have the rights to use and as that's the same in the US then I don't get what's different from a responsibility point of view?

[sacred dawn]

lol tired here I’ll continue this in the morning post coffee

[silk nexus]

I get that it's THEIR tool creating that something but none of what the US can create isn't legal here so I don't get it.

[sacred dawn]

Do you work for Google or are you affiliated with Google ?

Ah that’s better blocked him

Mark only by keeping asking and doing what we are doing will we get the elusive answer

[clear cipher]

I'll show you two prompts, which were refined on my opal, and were tested on stak

[silk nexus]

I'm definitely not following you at all now, sorry! Their model isn't able to tell the difference about what?
Let's say two countries (the US and another different country) have exactly the same laws (it must be based on laws somewhere along the line or everything would be available to everyone surely?) but the US gets all features but the other country doesn't then what's the difference?
I'm really not following at all.

[silk nexus]

I'm trying but I'm really not following what's being said at all. If two countries had the same laws but the US is allowed to do everything but the other country isn't then I'm struggling to see what the difference is that is preventing that other country?

[sacred dawn]

Problem over

Here you go mark

Here’s the link https://support.google.com/gemini/answer/14286560?hl=en&co=GENIE.Platform%3DiOS

So the question is does it work on flow now ?

It’s still blocked though lol

[silk nexus]

Thanks for the link but that's for Gemini. I'm talking about Flow where you can upload an image of a person or a scene with people in and then have Flow make a video from it. If I try that then it's a blanket "No you can't do that". I'm asking WHY are we not allowed to do that in the UK if our laws allow us to use images we have the rights to? Just the same as in the US they must have to have the rights to use an image and then they're allowed to do so. So the question still remains why one country is allowed to use that feature but not another!

I really don't know how much clearer I can put my question any more 🙁

[sacred dawn]

Yes so if it works in Gemini that means it will very soon work in flow

[silk nexus]

I hope that's true as I can't understand any other reason why it wouldn't. Thanks.

[idle compass]

I don't understand how you are allowed to upload images of real people into the new Gemini image editing app (Nano Banana) but not into flow. That makes no sense???

[past pagoda]

Can someone answer a simple question? does animation permits renew daily for free users?

[oblique gust]

But I did before and it worked

[hollow kayak]

/Only a man is singing, a woman holding the microphone and moving in rhythm with the music, on the street, people are listening to music, and a camera drone from above

[noble carbon]

is veo 3 stuck on 99% for u guys

[paper coral]

Nope it is not

[proud bluff]

Same here but u need to wait longer

[noble carbon]

ah it happens when the generations rn are against policies but if it isnt itll load

so if its stuck adjust ur scene

[proud bluff]

Flow is temporarily unavailable. Please try again later. That's what came up for me.

[woeful fossil]

<@&1366831844502274180>

[raw swan]

Hello everyone, is the Veo 3 Fast free credits for Ultra plan for a limited time?

[unreal viper]

thank youuu! we got it sorted fast thanks to this ping!

[past pagoda]

thanks

[lusty relic]

right when my ultra ends

[crisp basalt]

YOOO ultra user is getting good foods’n

[lusty relic]

they drop ts 🙏

[crisp basalt]

That’s amazing !!

[fiery radish]

Unlimited generation for ultra? Is it limited or permanent?

[tame ermine]

What users are saying is most image to video generations in flow in the uk for example are blocked as they look like real people it says which makes flow useless for anything beyond a simple clip, yet every other ai video platform works fine in the UK so it cannot be law, even orcs are sometimes banned as too realistic, I have travelled quite a bit but never encountered any orcs.

[spark lark]

Flow is amazing but after I purchased the Pro, the videos i create are only like 7 seconds

[stuck adder]

I wonder what the new updates are that’s coming

Maybe video to video to rival runway?

[turbid lintel]

probably veo 4

[stuck adder]

I doubt they have veo 4 already

[crisp basalt]

Too soon ?

[stuck adder]

Impossible

[turbid lintel]

no point giving veo3 for free if no veo 4

veo 4 will cost credit for sure

[crisp basalt]

Is it really free tho

[stuck adder]

Well that’s a good point, veo 3 fast is free

What else do you think it could be?

[turbid lintel]

veo3 free mean good thing is coming

i bet veo 4

[crisp basalt]

I thought it’s free for ultra users

[stuck adder]

I want 9:16 ratio, and video to video

[turbid lintel]

ultra always got the good thing first

the pay more

[crisp basalt]

Yall just be saying anything at this point 😭 it’s embarrassing

Anyway moving on!

[stuck adder]

Is that why they gave double the credits ? To bring in another model that’ll cost more to make up?

[turbid lintel]

yup

[stuck adder]

Surly it must be something big idk

[turbid lintel]

google now full swing on AI

they lead the race

[stuck adder]

They’re leasing right now for real

Leading *

If it’s even veo 4 I wonder how they could make veo 3 any better at this point lol

[gloomy helm]

Flow is nice but too limited for paid pro then to jump over 100usd for ultra is insane.

[stuck adder]

It can be used at production level once upscaled with topaz

[turbid lintel]

we need more than 8 second

[stuck adder]

What do you think this update could be?

[gloomy helm]

Idk

[turbid lintel]

btw veo3 already achieve character consistency

[worldly yacht]

Didnt us pro users get anything ?:D only the free and ultra users ?

[stuck adder]

Lolollll

I went on ultra with the discount

Well the thing is with the new image model I’m sure maybe video to video is coming cause that model they released is insane

[turbid lintel]

i can end my runway sub lol

save money

[worldly yacht]

we dont get love 😄

[stuck adder]

Yeah cause aleph is a hit and miss

And Luma already has video to video maybe Google wants to have that too

[spark lark]

i just burned through 1,000 creits in about 1-1/2 minutes lol Veo is a joke

or Flow or whatever it is

[turbid lintel]

sub ultra bro

[stuck adder]

How did you manage that? What mode were you using?

[worldly yacht]

we dont all have money for ultra

[turbid lintel]

but you will get free veo3

unlimited generation

[worldly yacht]

I know but some of us cant afford to do that atm

[turbid lintel]

understood

[worldly yacht]

If ultra got unlimited free , us pro users should at least get like 20 a day or something intead of 3

[turbid lintel]

they will give something for sure

[worldly yacht]

would be cool if they did

[turbid lintel]

veo3 for free mean something bigger is coming

cant wait lol

[worldly yacht]

yeap , keeping an eye out on the announcements

[idle depot]

https://vm.tiktok.com/ZNd4mBjas/

[gaunt mesa]

https://tenor.com/view/fry-shut-up-and-take-my-money-futurama-gif-11820664

Every day I think I’m going to spend a bunch on a different products “Ultra” subscription. Then Google comes by and says “unlimited Veo 3”

[placid path]

Just wanna share my first anime creation using veo

https://x.com/RonkeUpdates/status/1960099994108870950?t=9Haq-YxGPrKT8rVedAEwiQ&s=19

Ronke Meme Updates (@RonkeUpdates)

Ronke Rice Farmers: The Animation
Episode 1: The Kingdom of Ronin🔥

$RONKE $RICE SEASON!
@RonkeOnRon @RiceFarmerRonke @Ronin_Network

If you are a ronke rice farmer drop WEN RRF! WEN RRF!

X · Aug 25, 2025, 10:00 PM

[wicked crater]

I started using flow...

[night osprey]

HI

[past cargo]

Make a spider man animation video in which spiderman slip from a big building and then when he died rain starts with ruppes not a water

[lament flare]

Veo3 Gemini and Flow should be inspired by Grok to generate videos, with the possibility of generating from images in portrait format.

[silk nexus]

It was for me earlier on and I lost credits due to it never creating the videos but it still charged me. A few hours ago it also said Flow was temporarily unavailable so there was definitely a problem with the system somewhere.

[faint wyvern]

I generated an image with a nano banana, uploaded it to Flow > Frames to Video, and guess what? Apparently, it violates the guidelines because it’s too realistic. Seriously? You guys okay up there? How can it be “too realistic” if it was literally made with nano banana? And the best part—I tried earlier with Imagen 3 and got the exact same error. At this point, I might as well just draw it in Paint, maybe then it’ll pass your sacred rules.

[severe grove]

Well, this is a new venture for me because I have never done or used these sort of tools inside here in my business . Looking forward for someone to hold me by the hand and help me with a tour around these tools

[summer pumice]

flow is free now

[severe grove]

Thanks man🙏. I appreciate it

Okay. Well, I’m also a young tech entrepreneur and founder of TX. My software start up that I’m developing for crypto and forex traders . Do you perhaps have any sort of ways you could help us grow our user base?

Btw, I will be launching the beta version later this month as an MVP

[severe grove]

Okay ✅ no worries at all. But curious, when you say inbound and organic, does that mean you only work with physical businesses ?

[sacred crypt]

Hey guys, I'm using flow with the last frame reference of the image to create the next video, but it dulls the image quality. Any fix for it?

[twilit pebble]

hii! which prompt do you use it? If I can ask!

[clear cipher]

a test of 31, and a result of only 6 😅

[clear cipher]

first result, using ChatGPT benchmark

first result, using ChatGPT benchmark

[forest raptor]

The lack of originality is interesting. There was a time when EVERYONE was making "selfie videos of historical or mythological figures chatting", now EVERYONE is making "asmr videos of cutting glass objects"

[clear cipher]

using the same reference with different models

[clear cipher]

yes, because the image process does not prompt

I didn't understand

[swift anvil]

Hii

[stiff harness]

https://tenor.com/view/what-are-you-talking-about-huh-what-what-are-you-saying-stirring-gif-17820207

[oblique gust]

Hey, try two videos at a time. For example, use the first video last frame as the start of your second video. Then from your second video, start a whole new project or enhance the last frame of the second video, then you can start the third. I did it for my videos and I haven't seen any problem yet.

[rocky tangle]

Hey there,
Is Flow only available, if i get the Ultra?

[foggy linden]

Hi, I have a problem. I'm trying to generate my video but it says generation failed. Is this normal?

[sterile junco]

oh wow, google listened, now we can get VEO3-Fast for 0 credits with Ultra! this is HUGE!

[fair bone]

if flow work in EU

and also the watermark is annoying after having Ultra subs

[sterile junco]

wym? as in you switched from ultra to lower grade and now have watermarks? or?

[sterile junco]

does it not?

[fair bone]

You have watermarks no matter the sub

[sterile junco]

No? or is this something new?

[fair bone]

Its not new, but i think they should give us the option to not have the watermark

[sterile junco]

I dont have watermarks tho, so im not sure what you mean.

[fair bone]

VEO written on the right hand cornor at the bottom

[sterile junco]

I have an Ultra sub and i didnt get a single watermark on any video i generated so far and its been about 2 months now

[sterile junco]

Nope.

[fair bone]

ohh

you are right

I bought Ultra recently and i am in UK and its not letting me generate 99% of stuff because of laws in UK, so I switched to a pro account which was created in US and used that for POC, just checked on Ultra there are no watermarks, thanks

@sterile junco sent you a dm

[humble moat]

google veo 3 video oluştur turkish evde ciğer arnavut tarifi 9:16 gerçek hayat kaç dakika kaç scene saniye engilsh prompt

[rocky tangle]

Never!

[rocky tangle]

Nope. On Ultra you dont have any watermarks.
By Pro yes.

[sterile junco]

hm?

[rocky tangle]

Google let never the Veo3 FAST cost by 0 credits.

I get a list, you can watching at support by Google. Still need 50 credits.

[sterile junco]

um

you are wrong

thats all i have to say to that

[rocky tangle]

unlimited, no limits by usage.

Google AI Ultra - 25,000 monthly AI Credits
Product Model Number of videos offered AI Credits needed per generation
Whisk Veo 3 - Fast Up to 1250 20
Flow Veo 2 - Fast Up to 2500 10
Veo 2 - Quality Up to 250 100
Veo 3 - Fast Included Included
Veo 3 - Quality Up to 250 100
1080p Included Included

[sterile junco]

Ultra gets 25000 credits and unlimited VEO3 Fast costing 0 credits to generate 🙂

[rocky tangle]

Okay. Then it's new!
You are right. Because a few days ago it wasn't like that. Starting August 25, 2025!

[sterile junco]

Yep! its pretty damn cool!

now i can experiment more and not worry about wasting all the credits on failed generations and failed added sound 😄

last month i wasted about 10000 points becuase of these "errors"

[rocky tangle]

Honestly, I'm looking for other options too! You know. I don't think it's expensive, but Ultra is a lot of money.

[sterile junco]

for sure, if I wasnt using it for work and my boss didnt pay for it i would not spend 300$. a month for this 💀

[rocky tangle]

Yes. Its only for my own. Not work!
I use it for Social Media´s, or something create like funny stuff.

300$ is Grok 4

I like Veo3 very much.

[rocky tangle]

But apparently, Google's terms and conditions state that credits are only consumed when a video is generated. So, not per request, but per generation. So why did you lose so many credits!???

[deft helm]

Hey guys,
Seems like I suddenly get lots of mistakes using - veo3 fast mode.

Goes off-script.
Generating a vid without audio.
Adding a random animal sound mid scene.

Is there any known issue ongoing atm?
Wasn't having any issues the past 2-3 weeks, issues started yesterday.

[rocky tangle]

IDK

[sterile junco]

yea it gives an error after a request and then gives you an error "could not generate" and still takes your credits 😄

[sterile junco]

Maybe thats why its free now in ultra 😄

[foggy linden]

Hi, I have a problem. I'm trying to generate my video but it says generation failed. Is this normal?

[fair bone]

it hasn't happened with me, but try to write a more detailed prompt in sections for example
Action: The logo and title text remain static in their positions. Thematic VFX are introduced.

Camera: A very slow, subtle zoom in (push-in) toward the center of the title text.

VFX: Stylized pixelated footprints (neon blue, 8-bit style) rapidly walk across the screen from left to right, passing behind the text and disappearing off-frame. The background data streams continue to move subtly. The title text glows slightly.

AUDIO: The synth-pop soundtrack swells to full energy. A subtle digital stepping sound effect accompanying the footprints.

VOICEOVER: None.

NEGATIVE PROMPT: Do not move the logo or the title text. Do not obscure the title text. Do not make the background VFX too bright. Keep the camera movement very slow and smooth.

[deft helm]

Thanks,
I do make sure that the prompts are very clear, but looking at the #1319424187730755636 seems like other had similar issues atm.
Guess its an ongoing bug, hope Flow will make an official announcement soon to keep us on the loop

[warm solar]

Did VEO3-fast's become worse in quality? I'm on Pro, and was thinking of switching to Ultra. But after testing today, it seems the results are worse than before. Looks like the videos are coming from VEO 2.

[frigid forge]

I got a question anyone figured an easy way to quickly download all out flow video clips ?

[clear cipher]

There were 187 tests for these results, and basically everything disappeared

[hazy void]

"scene_identity": "shrestha interiors_8s_transformation_static",

"subject_and_action": "In a clean, white room viewed frontally, four VEO3 Designer paint cans roll in quickly from the right and stop at the center. One second later, the cans burst open with a synchronized, colorful explosion, precisely spraying paint onto the walls and ceiling. Immediately after the burst, the cans vanish in a smooth visual effect. The walls and ceiling are now perfectly colored. In the next moment, modern furniture slides in from the left, paintings and vibrant flowers appear with fast, elegant transitions. A round rug quickly rolls into place at the center of the wooden floor. The transformation completes in 8 seconds. The shrestha interiors logo fades in at the end.",
"environment": "Minimalistic, clean white room with a large window on the left side. Natural light enters softly. Floor is wooden and remains spotless throughout.",
"shot_cinematography": "Static, symmetrical wide shot from a fixed tripod. No camera motion. All action unfolds within a single frontal perspective.",
"audio_and_dialogue": "Clean, upbeat music that rises with the can burst. Subtle rolling sound as cans enter. A satisfying whoosh and splash with the paint explosion. No dialogue.",
"key_visual_elements": [
"Four paint cans rolling in from right",
"Colorful paint burst painting walls and ceiling",
"Paint cans vanish after burst",
"Walls and ceiling instantly transformed",
"Furniture slides in from left",
"Paintings and flowers appear quickly",
"Round rug rolls into center",
"Shrestha interiors logo fade-in"
],
"motion": "Fast-paced and precise within 8 seconds. Rolling cans, explosive paint burst, instant vanishing of cans, smooth decor entrance, rug rolls in, final logo fade-in.",
"ending": "Scene holds for a beat on the complete room as the shrestha interiors logo appears.",

"text": "Only the shrestha interiors logo at the end.",

"keywords": [

"Shrestha interiors"

"scene_identity": "shrestha interiors_8s_transformation_static",

"subject_and_action": "In a clean, white room viewed frontally, four VEO3 Designer paint cans roll in quickly from the right and stop at the center. One second later, the cans burst open with a synchronized, colorful explosion, precisely spraying paint onto the walls and ceiling. Immediately after the burst, the cans vanish in a smooth visual effect. The walls and ceiling are now perfectly colored. In the next moment, modern furniture slides in from the left, paintings and vibrant flowers appear with fast, elegant transitions. A round rug quickly rolls into place at the center of the wooden floor. The transformation completes in 8 seconds. The shrestha interiors logo fades in at the end.",
"environment": "Minimalistic, clean white room with a large window on the left side. Natural light enters softly. Floor is wooden and remains spotless throughout.",
"shot_cinematography": "Static, symmetrical wide shot from a fixed tripod. No camera motion. All action unfolds within a single frontal perspective.",
"audio_and_dialogue": "Clean, upbeat music that rises with the can burst. Subtle rolling sound as cans enter. A satisfying whoosh and splash with the paint explosion. No dialogue.",
"key_visual_elements": [
"Four paint cans rolling in from right",
"Colorful paint burst painting walls and ceiling",
"Paint cans vanish after burst",
"Walls and ceiling instantly transformed",
"Furniture slides in from left",
"Paintings and flowers appear quickly",
"Round rug rolls into center",
"Shrestha interiors logo fade-in"
],
"motion": "Fast-paced and precise within 8 seconds. Rolling cans, explosive paint burst, instant vanishing of cans, smooth decor entrance, rug rolls in, final logo fade-in.",
"ending": "Scene holds for a beat on the complete room as the shrestha interiors logo appears.",

"text": "Only the shrestha interiors logo at the end.",

"keywords": [

"Shrestha interiors"

[oblique gust]

i dont have any watermarks...

[signal jetty]

what do they mean with "final" ?

[clear cipher]

and final

[signal jetty]

When i download it?

[silk nexus]

I'm guessing they mean you can test your prompts out using Fast until you get one that works and then use that prompt with the full version to get your 'final' output. Only problem with that is that I don't believe you'll get the same output from the full version as you would with Fast even if you use the exact same prompt. I may be wrong there as have no way to test it but I've run the exact same prompt on Fast just to test and it's come out with a different output each time so wouldn't see it being much different on the full version?

[oblique gust]

not really. I am ultra user and i've been using VEO 3 fast. Been generating ton of videos since they rolled out that new update. Even when i saved the final video that i liked. It didn't charge me anything. Still 0 credits even when i downloaded the 720p or 1080p still 0 credits.

[hazy void]

"scene_identity": "shrestha interiors_8s_transformation_static",

"subject_and_action": "In a clean, white room viewed frontally, four VEO3 Designer paint cans roll in quickly from the right and stop at the center. One second later, the cans burst open with a synchronized, colorful explosion, precisely spraying paint onto the walls and ceiling. Immediately after the burst, the cans vanish in a smooth visual effect. The walls and ceiling are now perfectly colored. In the next moment, modern furniture slides in from the left, paintings and vibrant flowers appear with fast, elegant transitions. A round rug quickly rolls into place at the center of the wooden floor. The transformation completes in 8 seconds. The shrestha interiors logo fades in at the end.",
"environment": "Minimalistic, clean white room with a large window on the left side. Natural light enters softly. Floor is wooden and remains spotless throughout.",
"shot_cinematography": "Static, symmetrical wide shot from a fixed tripod. No camera motion. All action unfolds within a single frontal perspective.",
"audio_and_dialogue": "Clean, upbeat music that rises with the can burst. Subtle rolling sound as cans enter. A satisfying whoosh and splash with the paint explosion. No dialogue.",
"key_visual_elements": [
"Four paint cans rolling in from right",
"Colorful paint burst painting walls and ceiling",
"Paint cans vanish after burst",
"Walls and ceiling instantly transformed",
"Furniture slides in from left",
"Paintings and flowers appear quickly",
"Round rug rolls into center",
"Shrestha interiors logo fade-in"
],
"motion": "Fast-paced and precise within 8 seconds. Rolling cans, explosive paint burst, instant vanishing of cans, smooth decor entrance, rug rolls in, final logo fade-in.",
"ending": "Scene holds for a beat on the complete room as the shrestha interiors logo appears.",

"text": "Only the shrestha interiors logo at the end.",

"keywords": [

"Shrestha interiors"

[hazy void]

Bedroom studio Assembly
Prompt: description": "Cinematic shot of a sunlit Scandinavian bedroom. A sealed Brown box with the logo "shrestha interiors" trembles, opens, and flat pack furniture assembles rapidly into a serene, styled room highlighted by a yellow, RED throw on the bed. No text.",

"style": cinematic",

"camera": "fixed wide angle",

"lighting": "natural warm with cool accents with neon lights on the ceiling",

"room": "Scandinavian bedroom",

"elements": [

"Brown box (logo visible)",

"Bed with yellow throw",

"Bedside tables",

"wide screen monitor on the wall between the speakers",

"Computer with studio speakers on each side",

"Midi keyboard",

"Jordan poster",
”Prince album cover”
"art",

"rug",
”a piano chord chart hanging behind the monitor”
"vocal microphone",

"gaiming chair",

"over the ear headphones"

"acoustic guitar"
” “Shade” album cover
],

"motion": "box opens, furniture assembles precisely and rapidly",

"ending": "calm, modern space with yellow red accent",

"text": "none",

"keywords": [

"16:9",

"1804",

"Scandinavian",

"fast assembly",

"no text",

"warm & cool tones"

Bedroom studio Assembly
Prompt: description": "Cinematic shot of a sunlit Scandinavian bedroom. A sealed Brown box with the logo "shrestha interiors" trembles, opens, and flat pack furniture assembles rapidly into a serene, styled room highlighted by a yellow, RED throw on the bed. No text.",

"style": cinematic",

"camera": "fixed wide angle",

"lighting": "natural warm with cool accents with neon lights on the ceiling",

"room": "Scandinavian bedroom",

"elements": [

"Brown box (logo visible)",

"Bed with yellow throw",

"Bedside tables",

"wide screen monitor on the wall between the speakers",

"Computer with studio speakers on each side",

"Midi keyboard",

"Jordan poster",
”Prince album cover”
"art",

"rug",
”a piano chord chart hanging behind the monitor”
"vocal microphone",

"gaiming chair",

"over the ear headphones"

"acoustic guitar"
” “Shade” album cover
],

"motion": "box opens, furniture assembles precisely and rapidly",

"ending": "calm, modern space with yellow red accent",

"text": "none",

"keywords": [

"16:9",

"1804",

"Scandinavian",

"fast assembly",

"no text",

"warm & cool tones"

[warm solar]

Have you noticed any changes in quality for VEO3-fast?
I'm a pro user, and my usual prompts look like VEO2 quality when I prompt VEO3-fast, so I'm hesitating to switch to Ultra

[oblique gust]

still exact same for me. No changes. It is all up to how you do with your prompt.

[silk nexus]

No I was referring to what the person was asking where Google said you can save your money until the 'final'. I believe what Google means is you can test with Fast for free as much as you like and when you get a prompt that works well then you can use it within the full version which is when you would then obviously use credits up.

[oblique gust]

I don't think they charge you anything in final product either.

[silk nexus]

I don't think you're following what I'm saying. On Ultra, Fast is free for however many videos you want to do. We hopefully agree on that.
Now when you've got a prompt that works well with Fast you can then use that prompt with the high quality version where you WILL use up credits but hopefully get a higher quality version of that same prompt. Follow me?

[silk nexus]

Yep would be good to know exactly what they meant by it as I've tried quality and it's never been any better than fast quality.

I'm sticking with Fast but try out quality every few days to a week just to see if it's gotten any better. So far it hasn't so I'm sticking with Fast as I'm on the Pro subscription and buying extra credits when I need them.

[worldly oak]

Do any of you have automation from promt to generate video to download?

[silk nexus]

I don't think that can be done with Flow can it? Maybe with the API though?

[rocky tangle]

Like, 1.000 credits/ month!? I know. Yes. 25.000 by Ultra.
WOWW

Yes, i understand it. Someone here told me that, and i see, its for some days updated. Yes. Good news for me. But what is with watermarks in output videos!?

Thank you! 🙏

[timid orchid]

Hi im about to join ultra again, is this unlimited veo3 fast only for a month? or is this permanent? also do i lose credits once im happy with a veo3 fast for the final download? im confused, or does it mean i can spend credits for the higher quality veo once the veo3 fast looks good?

thanks has anyone confirmed its permanent or its a guess, i hope its not those 1 week thing where everything is unlimited than back to normal, even though the credits are doubled, for someone who works on 9 min videos credits run out fast.

[oblique gust]

I am ultra user. Trust me. They won't charge you anything. Even on the final product when you get high quality, it will still be 0 credit. I've been using it all day high quality videos and download in upscale all with 0 credits.

[silk nexus]

That doesn't make any sense. Google have said that Veo Fast is free not the quality setting.
Are you saying you're using the full quality setting (not Fast!) and not being charged? If so can you send a screenshot to show that as I'll be having words with Google if that's the case.

[oblique gust]

Oh you meant Veo 3 quality. I was talking all about VEo 3 fast with upscale download.

[worldly oak]

Not yet

[prime sparrow]

Just tried flow. Syncing issue is the there.

[silk nexus]

I did mention I was talking about quality a few times now 😉

[oblique gust]

in that case then yeah with VEO 3 quality they still charge you even in Ultra plan. But I don't use VEO 3 quality at all since I can get good quality with VEO 3 fast too. ><

[prime sparrow]

Anyone got the perfect sync?

With flow

[oblique gust]

lol sorry. I'm at work and been in and out reading without context haha

[worldly oak]

I have 20k tokens left, if someone has a good idea lets work together

[silk nexus]

No worries. I was wondering for a minute why Google are even bothering to charge for the Ultra plan if absolutely everything is free. 😂

[oblique gust]

i guess if that the case then they probably only charge for the services in order to use both for free lol. But honestly, even with VEO 3 fast at 0 credits. That is a huge win too. I've creating a lot of high end quality videos with VEO 3 fast. It's all up to how you prompt your JSON. Either way, i do hope more good things come in VEO 3. Longer than 8 seconds and more consistency tools.

[silk nexus]

I hope they give something to those of us on Pro.
I've been getting really good results too. See a lot of people saying that if they get 1 out of 10 good videos then that's about as good as they get. I'm getting at least 9 out of 10 that are good and the one that isn't I'm okay with anyway and can usually find a way to edit it so it's good anyway.
I've made my own Gem that I use to provide it with a simple scenario, characters and dialogue and it fills in everything else for me with random values such as age (unless I specify it), accent etc
Getting really nice and consistent results.

If you follow the whole of that conversation thread you'll see what I meant by 'free' 😉

Yep there's no way I was paying all that money to only get 1 video in 10 coming out okay! That's just for fools like you say.

[sacred crypt]

Any fixes for that?

[clear cipher]

Well, the only solution I see for this is to focus on text-to-video. Because when you use an image, it doesn’t start from the text it generates from the image. That means the quality depends entirely on the image itself. So you either have to rely on the image being very good, or you need to upload it in very high quality. That’s why I recommend focusing on text.

Ah, but for someone who’s just chatting, yeah, it might be better. I thought he actually wanted to learn how to generate video properly. Because if you’re generating video and relying on images, it means you don’t have enough talent or skill with text. So you just stick to using images. I didn’t realize he was a beginner. Damn!

[fluid sequoia]

google made veo 3 fast for free and doubled ultra credits

something big is definetly coming soon

veo 4?

[severe grove]

Ooh alright man. I appreciate the help

[fluid sequoia]

veo 2 --> veo 3 was 6 monthts

so we can assume the same timeframe for veo 4

[stuck adder]

I wonder what they possibly fix in veo 4 tbh lol

Veo 3 is great right now probably adding a native 1080p mode?

[lean narwhal]

One message removed from a suspended account.

[main saddle]

hope this helps! 🙂 #flow-announcements・📣 message cc @signal jetty

[maiden lily]

Thank you @main saddle

[jovial flare]

hey, I'm playing around with veo. I want to use frames and have the credits for it but can't use veo 3 with frames because feature isn't supported (?) and veo 2 isn't accessible to my plan. Is this right?

[halcyon vector]

you can use first frame with veo 3, not last frame

I'm pretty sure all the plans support using veo 2 unless you're on some trial basis

[jovial flare]

plans to add this I suppose?

thanks for the help

[halcyon vector]

yes according to one of their announcements they want to get veo 3 to use last frame and the ingredients feature eventually.

[hoary ibex]

I have a question, why do I have a problem?

[halcyon vector]

what's the problem it seems like you can use it. the fast costing 0 credits only applies to the top tier of subscription

[hoary ibex]

Ok, it means a subscription to a subscription higher than that

[halcyon vector]

"ultra" plan gets you free veo 3 (fast)

[hoary ibex]

okay , Thank you!

[clear cipher]

@edgy holly what's your opinion ?

[sudden condor]

Flow tornado thing coming soon

[clear cipher]

For someone who is just starting and has no skills yet, that’s how it is. You’re at the beginning, so saying that makes sense, and I understand.

In this case, I see the limitations in using images you’re stuck with them without any control or ability to make changes. Like I said, for someone who’s just starting out, that’s normal.

These are not superstitions I literally prove this all the time. I’m confirming that it’s real, not a superstition.

You can’t see the limitations because you don’t have that ability yet. I understand you because of that, so take care.

I understand you

If you literally can’t see the limitations, it’s already a bit late to learn. But if you’re interested, I can teach you.

neither do I

[clear cipher]

it makes sense

[main saddle]

fwding this question from @supple zodiac!

[mortal wing]

I’m trying to make these videos of ai baby getting interviewed. But for some reason the interviewer also answers the question, even tho I make it very very clear the baby should

[clear cipher]

When you have two characters in a prompt, you need to structure it with brackets. So you have to include the character’s name, followed by a colon, and then put the character’s dialogue inside brackets. For the other character, you do the same. This way the model understands that there are two different people speaking.

[silk hill]

hi in flow where can i find images i have generated with the new reference feature

i clicked out of it now i can't find them ahah

[clear cipher]

so haha

[silk hill]

yeah they seem to have evaporated

the model is really impressive btw

also when you save images out of it they don't have an extension so you have to rename them to .png manually (on win11 anyway)

[clear cipher]

answering your question and saving you time, you lost.

[worldly yacht]

Check your google drive and find the folder that says GOOGLE LABS AI something like that\

[silk hill]

theres a google AI studio folder but none of these images are there

[worldly yacht]

I will share screen and show you , Go into the labs lounge

[silk hill]

dont see anything like that in my google drive but i dont use google drive a lot

[worldly yacht]

you dont see anything like that ?