flow-general・🌊

[main saddle]

please add this to #1319424187730755636 if you haven't already - definitely want this seen by the team!

[fierce jackal]

I have the free pro trial atm

[idle crest]

same, thanks to tempumail and revolut

[fierce jackal]

Got mine from being in early access

[prisma sphinx]

hedra is only image to video, not video lipsync. kling has true video lipsync. but yeah sync labs is great too

[fierce jackal]

And 1000 is how many avg generations?

[idle crest]

Nah I was on early access too

[warped arrow]

I created a scene (8 secs) with an audio, Extend won't work, since it will only use Veo2, so audio will be gone. "Jump to" will allow Veo3, but it is not producing the audio at all. Any advice on how to accomplish that?

[frozen kraken]

Hello

I'm new here

[warped arrow]

It's the other way around, 10 is for Fast and 100 for Quality and Veo3

[amber granite]

I really need help with this audio thing.
I have tried severally to refine my prompt I was thinking it might be the reason why my videos comes out without an audio I have wasted over 6,000 credits just trying to get it right but nothing seem to work each and every video I create I select the highest quality and text to video to generate a scene.

Below are some prompt I’ve used generated by chatgpt.

Prompt: Miracle picks up a bright red crayon and turns to the camera. “I could color all day if you let me,” she says with a grin. Behind her, a shelf full of books and markers adds color to the cozy classroom. Light filters through the window as soft background chatter plays.

Prompt: Miracle stands in front of a shelf stacked with books and crayons. She holds up a red crayon and says, “My name’s Miracle . I could color all day if you let me.” She smiles at the camera as soft classroom sounds play in the background.

Prompt: Miracle walks to a small shelf filled with books and crayons. She picks up a book, holds it to her chest, and says, “My name’s Miracle. I love to learn!”

Prompt: Inside a bright preschool classroom, a cheerful girl named Zaza stands beside a chalkboard that reads “Welcome to Miracle Learns.” She waves with a big smile and says, “Hi friends! Welcome to Miracle Learns!” Her voice is joyful as light filters through the windows and the room glows with warmth.

[prisma sphinx]

So it was human error on my part.

Veo 3 with frames works in Canada, if you paid for the Ultra account.
Maybe if you're on the free tier it defaults to Veo 2

[warped arrow]

oh that is true, but probably better to say how much credentials needed per Quality, since you can generate 9 Veo3 , and after that, you don't really have 100 Fast generations anymore.

[fierce jackal]

What does refreshing gens mean

Oh ok cool

[lusty rampart]

Yo guys

[warped arrow]

I created a scene (8 secs) with an audio that I want to extend the scene, Extend won't work, since it will only use Veo2, so audio will be gone. "Jump to" will allow Veo3, but it is not producing the audio at all. Any advice on how to accomplish that?

[unique abyss]

Putting aside the question about what is desirable, I don't find it perplexing at all.

It's consistent with most everything Google has ever made--they have always seemed far more interested in that changing the product to be more intuitive, they hardly ever create a bunch of documentation. I certainly wouldn't expect them to do a ton more here, even if that would be great, given that the nature of the model, and the very nature of the Flow could both easily change in 3 months or six months. (Counterpoint, YouTube creators get a fair amount of support, but YouTube is far more stable a product, and far, far more financially significant at the moment)

This gap might be an opportunity for an author or Youtuber to solve--provide a clear one stop shop for answers.

But if third parties don't, it nothing suggests Google will do a ton here, other than refining and improving FAQs a bit. That Google has a Discord with nice people who respond to questions is already batting above historical average. More would be great, but it's a hard thing to actually expect.

[lunar harness]

anyone have any advice, workflow or stack that would help me generate videos with VO in another language other than english. i was thinking generating the video on veo 3 and then using something else to generate the vo ? thoughts

[light robin]

Anyone happen to know how I work Flow for Veo 3?

When I press create with flow, it takes me to something else.

[fierce jackal]

Yeah video prompting I need to learn a lot for, I’ve mostly been a MJ user so I’ve had it easy lol

[crimson hound]

Anyone trying negative prompting in flow ..outside of ai studio??

I may burn a few credits and see how it goes

[fierce jackal]

Does it still use your credits if Gemini can’t produce a video cause too many request?

[harsh shell]

anyone know what to do if I was deducted points for generating a video in flow, but no results are generated?

[fierce jackal]

Ty

[latent elbow]

Can people fix the OCR?
Original + Prompt:
A low trust society is one where people have little confidence in others, institutions, or the government. Characteristics include: High levels of corruption Weak rule of law Reliance on personal networks over institutions Widespread suspicion and fear of being cheated Inefficient public and private transactions It often leads to slower economic growth, poor governance, and social fragmentation. A high trust society is one where people generally trust each other, institutions, and the government. Characteristics include: Low corruption Strong rule of law Efficient institutions High civic cooperation Smooth economic and social interactions It supports economic growth, stable governance, and social cohesion. Make an infographic

Veo3:

If the 'OCR' can be fixed, it would be good. Please AI ML it more.

This one seems alright too, would be good for presentations, although, some sort of 'interactivity' should be included:

[swift herald]

is flow down?

[hollow frigate]

How do you extend a scene so its longer than 8 seconds but keep the audio? it keeps switching me back to fast mode

[lunar harness]

i will look into it. i just tried prompting veo 3 to create the scene without audio and vo and it generated random vo -_- is that how it is right now hit or miss?

[analog sedge]

If the output doesn’t have audio in it the credits should be refunded

[crisp drum]

How can i ask it to make a video with voice ? In what channel and with what prompts ?

[warped arrow]

probably wishful thinking, but would be nice to have a "recycle" option, like for a grace period of 2-5 minutes, you can recycle(delete) a test scene/prompt, then get a fixed credit of 5 or 10 (even if you used 100 to make that video)... just getting a portion of the credit back would be nice, lol

[craggy egret]

It would be cool if there was a public roadmap for what's being worked on...

[lunar harness]

seems like my best bet is going to be generating video with veo 3 and dubbing using sync.so elevenlabs for the vo

[quiet sphinx]

every other generation or upscale seems to be failing. are we all crashing the servers lol

[warped arrow]

sometimes, refreshing your browser, or logging out and closing the browser will work.... I know it is equivalent of "turn off you computer" fix lol... but it worked for me for a couple of times, even for having failing generation issues.

[quiet sphinx]

thanks. This seems to fix it temporarily, especially for upscaling

[lucid dawn]

ok. since yesterday something has been happening that I totally don't understand. I'm writing a prompt (which by the way worked normally before), it's about dialogues or character monologues. I'm putting them in quotation marks. since yesterday each character, instead of speaking normally like a human being, sings, and there's generated music playing in the background. what's going on?

[halcyon vector]

Day 2 of praying for ingredients in veo 3

[fresh steeple]

Is there any info on the Veo3 quota on Gemini with a PRO plan?

https://tenor.com/view/puss-in-boots-puss-in-boots-eyes-puss-in-boots-the-last-wish-gif-27592566

<@&1173646484311384114>

[rugged walrus]

Hi! After subscribing to Flow, when I try to create videos, this appears and won't let me do anything. Does anyone know the solution?

[fresh steeple]

Yeah those are Flow credits. I’m asking about the Gemini website which spends no credits.

[fresh steeple]

This seem to be time based rates

[placid sparrow]

flow isnt working from this page...

[fresh steeple]

When did it break for you?

[placid sparrow]

wdym?

[olive vault]

Anyone knows if Flow don't work at Brave navigator?

[tame patio]

hello everyone

im getting this "You must be a Google AI subscriber to generate. Feel free to look around in the meantime." but im pro user

[fresh steeple]

it should work since is Chormium based

[fresh steeple]

Are you in one of the supported countries? Some have Pro avalable but no Flow access

[last crane]

Try going to https://labs.google/fx and click "LAUNCH FLOW"

[stray nacelle]

Vector illustration of 3 Gen Z teens hanging out on their cell phones, one looks upset at what they see on their phone

[last crane]

This is Discord, not Veo

[stray nacelle]

sorry

Can you creat in Veo in Discord?

[last crane]

Nope, just via Gemini and Flow

[tame patio]

US using VPN

[warped arrow]

Is upscaling download only available per scene? and not in scenebuilder?

[idle crest]

I took advantage of the student email thing and cant wait till june 1rst. I got it today

Since credits fill up monthly it has to gotta be

[rugged walrus]

no puedo ingrsar a flow por esta cosa que sale de la nada tot

[last crane]

Vaya a https://labs.google/fx y luego haga clic en "Launch Flow"

[crisp grove]

Sybau

[last crane]

?

[crisp grove]

Stands for shut yo bih ahh up

[last crane]

Yeah I know what it stands for, why did you say it?

[crisp grove]

I was excersising my free speech

[last crane]

This is Discord, not the United States of America

[crisp grove]

Does it matter

[last crane]

Yes, you don't have freedom of speech on Discord

[crisp grove]

No way, is my speech restricted

[last crane]

Yes

[crisp grove]

So then what happens now

[last crane]

I mean you haven't said anything that violates Discord's policies or this server's policies

[rugged walrus]

no me funcionó 😦

[last crane]

¿Aún te aparece la misma ventana emergente?

[rugged walrus]

sip

[last crane]

Intente hacer clic fuera de la ventana emergente o desplazarse hasta la parte inferior de la ventana emergente para encontrar un botón para cerrarla.

[rugged walrus]

lo he intentado mil veces y nada 😦 a usted le salió lo mismo?

[last crane]

No, a mí no me pasó lo mismo.

[rugged walrus]

bueno, muchas gracias por intentar ayudarme 🙂

[last crane]

Yeah, I'm pretty familiar with attention-seeking individuals like themself, thanks though :)

[last crane]

Echa un vistazo a este canal, puede que tenga alguna información que te sea útil: #1375910551397597316

Algunas personas dicen que alejar la vista puede ayudar a hacer clic en el botón "comenzar".

[rugged walrus]

siii me funcionó!!! muuchisimas graciass uwu

[last crane]

:D

[main saddle]

hi everyone- I'd like to remind everyone to adhere to our #rules-and-guidelines・📋 as well as our content policies. Our goal is to cultivate an encouraging environment where people can examine emerging tech with a creative and responsible spirit. Those who do not comply to our server rules, guidelines, and content policies will receive a warning and can be removed from the server if the behavior continues.

please re-review the #rules-and-guidelines・📋 and keep conversations healthy and productive. thanks!

[analog sedge]

These kids using another language and that kid using the tiktok language should both be banned

[jolly cypress]

Hi, has anyone experienced issues creating vidoes with Flow and Veo2/3 ? I have a plan but still impossible to create any text to video. Whatever the prompt, even vey short, I have the message ''Failed Generation''!!

[last crane]

Why should the individuals speaking Spanish be banned?

[last crane]

I don't think others are experiencing this issue- are you logged in to an account with Google AI Pro?

[analog sedge]

This is english discord. Ask the mods to create a new group if you want to speak your language

[broken cairn]

Wait what? ¿Porque no podemos hablar español acá? Its simply another language, which Google has supported in almost every page, and a lot of the United States has people who speak Spanish .

[analog sedge]

Nobody else understands your language and it feels spammy

[jolly cypress]

Yes, I have Google Ai pro account but nothing works! Whatever the kind of video I try to create or the different options I always got the same message ''Failed Generation''

[main saddle]

hey all - as per #start-here・📍, "Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now."

we won't ban folks for speaking other languages, but we will ask to keep convos organized in threads for non-Eng discussion until we can offer support in other languages.

as a reminder, this channel is for discussing Flow. thank you.

[jolly cypress]

I talked with Google chat but it seems that nobody knows !!

[main saddle]

hey Dave, welcome to the server! this could be for a variety of reasons. please add this to #1319424187730755636 if your specific blocker hasn't been added there yet, so the team has visibility.

[real field]

Seems they've tweaked the voice model in the last 24 hours. I would hold off on generations till they get it sorted because it's awful.

[fierce jackal]

Another question, is the veo 2 and 3 credits/gens shared?

[analog sedge]

The credits run out so fast. Wish the team would add an unlimited plan even for like $500 a month.

[real field]

Each model utilizes a different amount of credits.

veo2-fast: 10 credits per 8 second video (no audio).
veo2-quality: 100 credits per 8 second video (no audio)
veo3: 100 credits per 8 second video (with audio if you're lucky).

[echo briar]

Its google, they own the most visited site so they could afford to make it free

[minor rune]

Most people cannot afford a 250$ sub plan so...

[minor rune]

It's just weird that veo 2 quality and veo 3 are the same price

[echo briar]

i wish they could just make it free, they dont need to follow in openais footsteps and make it overpriced

[fierce jackal]

So like my question is if I make 10 videos with veo 3 does that mean I can no longer make more videos with veo 2 quality

[minor rune]

You are in USA? Or in the 71 other countries currently supported

[real field]

The GPU time required is substantial. There will never be a $500/mo plan for unlimited generations unless they significantly throttled the amount of generations you could make in a day (then it really wouldn't be unlimited).

[minor rune]

I think it's two different things

So you should still be able to make veo 2 videos

[brisk swan]

hey guys the 1000 credits reset right?

[real field]

Originally veo3 was 150 credits per video but I think the product team discounted it further to 100 credits per video.

Keep in mind that everything is discounted right now. In August prices will double.

[last crane]

Thanks for clarification Georgia :)

[brisk swan]

Ultra AI is 50% off right now, should i get it.

[minor rune]

I think they will veo 2 quality and introduce a veo 3 fast to keep at 100

[real field]

You can make as many veo3 and veo2-fast/veo2-quality videos as you want so long as you have credits.

[last crane]

How much do you use these AIs, and what for?

[fierce jackal]

Ok thank you

[minor rune]

And the veo 3 quality go back to 150 maybe

[brisk swan]

For my youtube channel

I think its worth it honestly to get ULTRA AI, otherwise their is no point in using high qaulity.

because of the deal

[last crane]

If your channel is monetized and making more than 250 USD per month then I would definitley go for it

[brisk swan]

Yea pretty much just asking how important is high qaulity versus fast qaulity

[minor rune]

I would say maybe try with one month in pro

[analog sedge]

True, hopefully we have some breakthrough in the next year or two

[last crane]

Depends on what your channel's audience wants

[minor rune]

And if the result is good for you, you can upgrade your plan to the ultra the month later

[last crane]

Quantum computing 🔥

I saw Google was making some advances there

[brisk swan]

lol

[analog sedge]

If you are looking to make money forget it. For now it’s just for fun

[last crane]

Or possibly buying credits at a set rate (like 1 dollar = 1 credit type of thing)

Obviously that ratio is unrealistic

[brisk swan]

What do you mean zac, you can use the stuff you make with flow in your youtube videos, its worth it.

ultra gives you like 13000 thats enough to make all high qaulity generations for like a month and i would immagine then it would reset

[analog sedge]

I bought the ultra yesterday

[brisk swan]

you because its 50% off

[analog sedge]

Almost finished all of it

in 1 day

[last crane]

Are you just using it for fun?

[brisk swan]

well thats your fault

[valid sentinel]

Hello Guys , Just got .
Im Trying to acsess flow , but for some reason the welcome message is stuck , even after refresh even in icognito I just cant get to scroll down and press get started .

is that happen often ? Is there a turn Arround I would greatly appreciate your assistance .

@main saddle

[last crane]

Several users are having similar problems: #1375910551397597316
I believe if you zoom out a bit you can get access to the 'Get Started' button

[analog sedge]

Alot of failed generations

[valid sentinel]

@last crane , so simple yet didnt think of it it worked . THANKSSS

[analog sedge]

Have to try many times to get the desired outcome

[brisk swan]

Zac

are u using text to video

The best way to use it is to find a cool image, buy the rights for it and then animate it with frame feature

like i made a cool robot and it turned out amazing without even high qaulity

You are just not using the flow correctly

Just dont try making it dance

[analog sedge]

im just using it for dopamine hit

[wary dove]

hi folks! how do you generate sound across multiple scenes? I can create a clip with audio via VEO3. But when I add to scene and "jump to," the audio is removed from both the original clip and the newly added one. 😭

[analog sedge]

That way you might be able to save more then

[real field]

I like many others blew through my ultra credits in a single day. I have since then purchased about 10,000 additional credits but can blow through 5,000 in 4 hours now.

[brisk swan]

Yea normally artists will give you rights for as little as like 5 bux and u can then use that with flow

[lusty relic]

Not sure if anyone has gotten this but I paid for Google ai pro and when I go to flow website it just says that and when I click subscribe it takes me to the 2nd image

[last crane]

No problem :)

[brisk swan]

Zac try the frame feature it works so well its by far the best even with fast qaulity

[real field]

I think they definitely should do this or allow you to attempt to restripe the video with audio (I don't know if this is even possible or at what point audio is striped to the video).

[lusty relic]

if anyone has a solution for this could you plz help me out?

[analog sedge]

It’s ok im just brainstorming new ideas for now

[brisk swan]

want me to give you my secret webiste

daviant art!

🙂

[analog sedge]

I already use midjourney so no short of images

[brisk swan]

they got the best pics

never heard of that

can u message artists directly on their too?

[analog sedge]

You haven’t heard of midjourney? Lol

[brisk swan]

will check it out!

[analog sedge]

Text to image model

[brisk swan]

oh

yea thats cool too

[analog sedge]

Or why dont you just use imagen 4 if you are gonna subscribe to goole anyway

[brisk swan]

Yea i do use it too

[fierce jackal]

Midjourney is another ai art generator

Oof

[brisk swan]

I KNOW

[fierce jackal]

I was way late to this convo

[brisk swan]

obviously i know this

[craggy silo]

Hi

[fierce jackal]

That’s my bad, discord hadn’t auto scrolled down lol

[sharp dome]

really glad that the this discord server is being filled with colorful people

[craggy silo]

How can i do a video with AI

[sharp dome]

it used to be a bunch of internet weebs before. these few last days I see so many colorful wonderful people pouring in

anyways

how are you guys creating content Non-Stop?
are you guys that rich?
or are you just burning off the last credits

I only have like 2,000 credits left...

I use it like it's my last air

[fierce jackal]

I wish I had that much disposable income tbh

[sharp dome]

creating something only if I feel like necessary

[sharp dome]

me too

[brisk swan]

Catgot u have ultra ai?

[sharp dome]

yes

[brisk swan]

do the credits reset every month right

[sharp dome]

yes because you pay every month

it's really not a reset.... you're literally paying for it

[brisk swan]

oh i guess your right

[sharp dome]

more like its an auto-buy

you get 125 gens each month

[brisk swan]

on veo 3 high qaulity

[sharp dome]

that's about worth 125 dollars worth because payung 20 bucks gets you 20 gens

veo3 is really expensive. the monthly subscription is your accessibility (costs) plus some usages

you have to pay even more if you actually want to use it rather freely

[brisk swan]

Yea but the stuff it was making for me looked the same as in the cineplex.

[sharp dome]

cineplex? what is that

[brisk swan]

like u know the movies

[sharp dome]

yeah it's like movies

by the way if you are salty I mean I hope you're salty because I feel that way

[brisk swan]

LOL what do u mean

[sharp dome]

I put all my credits and creativity into this short film I made

[brisk swan]

LOL HAHA

[sharp dome]

https://youtu.be/sopOUiphIY0?si=ebABb4zKS_4s2UiB

VEO THREE?!!! MORE LIKE VEO THREE HUNDRED DOLLARS!!!

[brisk swan]

thats hilarious

[sharp dome]

I think I could be really good at prompting

I wish I had more opportunities to explore what I can create with the veo

if I had an access to unlimited access to veo, I will not stop create content for on behalf of googles team

ARE YOU HEARING ME GOOGLE? I SAID I WILL CREATE CONTENT FOR YOU NON-STOP!!!!

the dev team occasionally reads through the chat

I am hoping maybe they will read my chat and secretly grant me more access

[last crane]

I don't think they can do that even if they wanted to unfortunately

[sharp dome]

YOU CAN GIVE ME UNLIMITED ACCESS TO VEO! I WILL NEVER TELL TO ANYONE!!!!

[sharp dome]

you never know

well it cost nothing to try...

[last crane]

They're social media managers / community managers I believe

[sharp dome]

there are actually developers

there are community managers but also developers

occasionally they come here to answer stuff

[last crane]

ah cool

[sharp dome]

now u know this cool fact u r welcome to join me to shouting for more free access

[last crane]

didn't know that

[sharp dome]

PLEASE GIVE ME MORE ACCESS I WILL CREATE MORE CONTENT!

I wanted to make continuation of this plot

but alas I only have ... 2k credits

[brisk swan]

Why did you use all your credits on 1 video

[sharp dome]

there was a learning phase ... I spent a few thousands

then I made this video using most of my credit

[brisk swan]

Honestly its an amazing video!

[sharp dome]

then I'm now fiddling with whats left

[brisk swan]

but now you cant upload for a month lol

[sharp dome]

that really makes me sad

you know how art students are always poor?

I think I kinda get it now

creating an art costs a lot

[brisk swan]

2k is still enough for fast qaulity

im on pro and only have 710 right now

but i might upgrade because its on discount where I'm at.

[sharp dome]

fast q is good at animating meme

not as glorious as v3

ve3s like a heroin

whenever I get failed gens... this is created usijg v3, I get really salty

I wanted to create young girl patting the calf and saying

oh calf! u will one day be my delicious burger!

snd the mom says no to her

but instead I got this

[twin fox]

I generated two Veo 3 videos on the Gemini app and reached the "daily" limit until the 29th. Is this normal?

[twin fox]

No sound?

[brisk swan]

Yea flow is a different program then ai studio

or gemini

You gotta have google pro to use flow i think

[sharp dome]

its a bug 😦

[minor rune]

I have the same

[sharp dome]

lol

[brisk swan]

if u didn't tell me thats ai i would think its real

[sharp dome]

yeah its actually a short movie quality

[brisk swan]

wanna see my robot

[sharp dome]

do you want this dog to live?

[brisk swan]

what kind of sick content is that

[sharp dome]

Rick and Morty like kinds of humour

lol

[brisk swan]

lol

[sharp dome]

GIVE ME MORE AI CREDITS OR I AM GOING TO KILL THIS FAKE AI DOG!

[twin fox]

You need AI credits to kill him though 🫣

[brisk swan]

PLZ

DO NOT GIVE CATDOG MORE AI CREDITS 🙂

[digital yoke]

Can some tag me when ultra i available in sweden

[brisk swan]

Catdog watch and learn

watch and learn thats how u do it baby

thats with fast qaulity too

pretty awesome hey

[sharp dome]

the irony lol

[brisk swan]

Catdog i must go defend super earth in hell divers 2

hopefully i will see you around! good luck

[fickle shard]

Nooo😂

[acoustic swan]

Hi everyone! Is there any way to evaluate the prompt before generating the video to avoid spending credits unnecessarily? I’m trying to create a video, but every time it generates the result, there’s always something I don’t like — and while I refine the prompt, I just keep spending more and more credits…

[sharp dome]

no

[sharp dome]

you will not get the perfect gen unless u spend every credit you have

You will need to use frame to video if u want "the perfect shot"

get ur image genned first using whisk or Midjourney

then gen the video based on that frame

[jolly cypress]

seems to be not so easy to create a 3mn video...

[muted cape]

you also can't set a fixed seed. so you're going to be playing the seed game quite a bit. there won't be much point in fine tuning your prompt at this rate.

i wish there was a way to evaluate before generating though. that would be nice. just isn't really how this all works even if you were doing this locally. though if you could do it locally, obviously cost wouldn't be as much of a concern

the only way you'll end up with what you're after is by generating the same thing dozens or even hundreds of times. this is just how it works. again regardless of whether or not its veo3, veo any version any other text to video, image to video, or even text or image to image models. you keep on generating

you can't even adjust the denoise value between img2img or img2video so you can't control just how much it looks like the original. you're stuck with whatever you get. so play the seed game, generate again

[flint summit]

How do you delete a scene in scenebuilder? And how do you get it to follow the artstyle of the previous frame?

[atomic pewter]

Anyone having the issue of not being able to click get started? Just bought membership

[crisp basalt]

I'd use Extend

[dusk kayak]

https://youtu.be/5YeHFcIRNz4

[twin fox]

That's actually a very clever idea.

[hard cipher]

anyone know how to get consistent character and scene in text to video using the audio ?

[fading dirge]

does the 5 daily videos generated on gemini cost credits?

[uncut schooner]

wow, there are so many bugs with veo3

why cant veo3 be used with extension??

[muted cape]

Does anyone know if there's a content filter or something? No, I'm not generating pron lol. It's just a night club scene with a woman in lace being seductive no foul language, etc. It's for a longer video that's horror/suspense. I saw a puppy and a knife with a threat to kill it in another video someone shared so I assume that isn't the case...but 3 failures in a row now. Maybe bad luck

oh darn it does say in the message about policies. but it's fleeting. i didn't see it

[proud mesa]

I have a scene and i keep asking it to Jump to a new angle of them but it only changes the character action. Does anyone have any tips for the prompt i need to write so i can get an over the shoulder, or a birds eye view of the same scene?

[muted cape]

lets see if I can tone that down a bit. it doesn't have to be racy or anything. pretty messed up you can threaten a dog with a knife but night club seductive woman in lace is a trigger

ah i did use the word "sexy" i guess maybe "seductive" would be ok

puppy killers gotta get their policies in order there 😂 (yes i'm teasing)

[warped arrow]

you can use "jump to" with Veo3, but it seems that the audio won't work doing that, so don't waste your credit doing that... i think that feature is still experimental so still a lot of limitations.... my theory, the tool can't remember/access yet the audio from the previous scene, so audio can not continue by doing extension or jump to

[uncut schooner]

https:// x.com/0xAskar/status/1926490989415326116

yeah exactly, idk why the audio cant work either with jump to lol
i made a whole thread about this in case you have some other answers

[muted cape]

i'd really just make the clips and stitch things together in video editing software on your desktop. you'll have far more control

[ruby star]

Can we add voice over for a different language than english in prompt?

[muted cape]

good question, all i know is I got some other language somehow before. though it may not have been a real language as I can only speak one. but I have heard other languages

[ruby star]

I saw other content creator do that, they speak on their own languange. but when I tried it, I got a warning message "Only english" or something like that

[warped arrow]

but what if you generated one guy talking with a particular voice, if you create another clip to continue the 1st clip, same guy with same voice, how do you know you will be able to replicate the same voice?

[muted cape]

they should keep their messages around. i walk away or change tabs and come back to failed generations and i have no clue why

[muted cape]

well that's easy for now at least. veo3 can't do that

[warped arrow]

yeah hopefully they would be able to fix or implement extend or jumpt to using Veo 3

[uncut schooner]

the problem is that these videos are by definition not consistent frames or individiuals

[warped arrow]

most of the samples I am seeing right now are multiple disconnected scenes, edited together

with people talking

even the scenebuilder won't work with audio, if you try to connect 2 Veo3 clips... but anyone can correct me if I am wrong, i think I tried it earlier today

[muted cape]

correct. that's where you'll have the most success for now. with scenes that don't need to be the same

[uncut schooner]

yes exactly, thats why these "pharma" like commericials work well, because each scene is disjointed

and you can overlay music and images in imovie or video editors

[muted cape]

the other thing i'm trying to do right now (if the thing doesn't keep nagging about policy, which i'm totally not trying to violate here) is take a video and bring it into a different tool that does support video2video. and use scenes from there that don't require dialogue

[uncut schooner]

im just pissed i paid $125 for this when i cant even use 1) ingredients, 2) extension, nor 3) jump to!!!

[muted cape]

yea... i'm a bit conflicted. while i'm extremely excited about veo3... i ran through my credits in less than 48 hours and i bought more. yet I feel a bit burned because the features aren't there yet

there's not enough professional controls for a professional price tag. to be blunt and simple there.

[uncut schooner]

yeah, i dont mind spending the $125 but im def not re-purchasing credits until they fix this

[uncut schooner]

precisely, well said

<@&1173646484311384114> would love to have yall read this feedback from above and read my X thread about it too that i posted above

[spring wave]

https:// x.com/0xAskar/status/

[uncut schooner]

i assume yall are aware of these limitations but if not, it is very much needed to be addressed in order for this to be genuinely the model for video generations

[timber copper]

Hello

How can we pay to use flow?

[uncut schooner]

send some ethereum to askar.eth

[timber copper]

Ok

How much

[uncut schooner]

just 0.25 eth

but in seriousness, go to flow and upgrade account to google ultra

@muted cape what have you been building such that youre burning through your credidts so quickly if you dont mind me asking

[muted cape]

was testing the tool a bit. trying to create a commercial for the company I work for. which would require the continuity thing so it was tough. lots of generations to try and work out subtitles (didn't really find anything that worked). then I created a few videos for myself for fun with some more credits I purchased. i'm now creating a series of "AI nightmare fuel" or "AI takeover" kinda theme there https://www.youtube.com/@HackAfterDark

i'm not a content creator per se, but I figured I'd post some of my content to share because it's a bit of a hobby. i've been generating AI images (and some video) for 3+ years now. locally, online, lots of different models. i've trained LoRAs to share with others, etc.

[crisp basalt]

maybe post it in proper channel #1319424187730755636

[uncut schooner]

this is funny

my also real first usecase was making a commercial for my company too lmao

[muted cape]

my boss was looking for something that was fairly unrealistic. but i gave it a try

the technology isn't quite there yet but it's getting close and better

[crisp basalt]

Just a quick PSA: if your Flow output’s busted, like low quality, wrong subject, or just plain cursed. don’t forget you can flag it. It’s not just for show, it actually helps the team track and (hopefully) improve stuff 🙂

Click the three dots > Flag output > pick what’s wrong (inaccurate, low quality, offensive)

[muted cape]

I always hope they credit my back when I flag the content 😂

[crisp basalt]

And always check the faq page https://labs.google/fx/tools/flow/faq .

[mental mulch]

I hope we get credits back if we cancel in case we accidently generate

[muted cape]

yea, create a project you can't title. hope you remember things by date.

[crisp basalt]

you can lol

did you not explore the tools

you can literally do it from main page

You cant rename ur projects from this view, obviously. Switch to grid view if u wanna rename all the ones youve made.

[edgy pasture]

Does anyone have tips on how to make my mini film better? I wanted to add scenes in between but I don’t know how to make the characters and video more consistent

[muted cape]

thanks, you can't do it from within the project or when you go to create it. super slueth skills there, right on!

[muted cape]

brutal when the video comes out pretty darn good but there's no audio 😦

[real field]

Hopefully there will be a "regenerate audio" button for 10 credits.

[muted cape]

that would be pretty nice

[winged karma]

hello guys

Quick favor to ask! 🤝

I've been trying to generate a video with this prompt about Chopin and winter scenes, but no luck so far. Getting consistent failures and wondering if others have the same issue.

Anyone free to test this out? Would love to know if it's just me or if the prompt needs tweaking:

In the winter of 19th century Paris, the renowned composer Chopin sits at his piano, creating music. Snow falls gently outside his window as he recalls the winter landscapes of his homeland Poland: across vast snow-covered fields, horse-drawn carriages carry travelers through the countryside. The silver bells on the carriages create melodious sounds, reminiscent of the musical notes in Chopin's études.

The travelers wear expressions of nostalgia, their eyes reflecting deep longing for home. In the distance, the beautiful silhouettes of Polish villages can be seen. The carriages gradually disappear into the horizon of the snowy plains, while the bell sounds linger in the air, inspiring Chopin's musical composition and transforming into beautiful melodies.

Cinematic lighting, period costume details, close-up shots of piano keys, sweeping landscape views, warm indoor lighting contrasted with cold outdoor scenes.

Much appreciated! 💙

[frigid pier]

bro wthhhh
the first shot is Blockbuster movie quality 💯

[wise cypress]

rethink how you prompt "creating music" is a weird thing to tell an ai video prompt

[crisp basalt]

i hit limit on veo 3 right now , but ill try on veo 2 with refined prompt:

Close-up on the aged, polished ivory keys of a grand piano as Frederic Chopin's elegant fingers, adorned with period rings, press down, creating a sequence of poignant musical notes; the warm light from a nearby oil lamp casts long, soft shadows across the piano and Chopin's focused face, his brow slightly furrowed in concentration; the camera slowly pans up to reveal Chopin in his 19th-century Parisian study, dressed in a velvet jacket, his gaze drifting towards a frost-covered window; through the window, a slightly blurred view of softly falling snow against the muted lights of Paris; the sound of gentle piano music underscores the visual; the scene dissolves into a wide, cinematic shot of a sweeping, snow-laden Polish landscape under a pale winter sun; horse-drawn carriages, their runners hissing softly over the fresh snow, move across the vast, white expanse, the delicate tinkling of silver bells attached to the horses' harnesses carried on the crisp winter air; a medium shot shows travelers inside the carriages, their faces etched with a nostalgic longing as they look out at the passing scenery; the distant, picturesque silhouettes of traditional Polish farmhouses and church steeples are visible against the horizon; the sound of the carriage bells fades slightly as the camera focuses on the expansive, silent beauty of the snowy plains, the scene then transitioning back to Chopin's study, the melancholic piano melody continuing, seemingly inspired by the recalled winter journey and the lingering echo of the carriage bells.

[wise cypress]

btw you have tooo much stufff for one prompt need to cut and stitch @crisp basalt

[crisp basalt]

cool, thanks for the feedback

[crisp basalt]

For just an 8s vid, that prompt kinda tryna do too much , too many scenes, transitions, and deep feels to really land in that short span tbh. All great ideas, but cramming that much emotion and scene change into 8s might feel rushed. The pacing’s important, and some parts might end up skipped or come off disjointed

im gonna try refine it again

refine it to focus more on chopin inspiration i think

[minor rune]

Add this to the suggestion box

[lone obsidian]

#flow-tips-n-tricks・✏️ would anyone try to test the prompt that i made?

[minor rune]

Why don't you test by yourself?

[lone obsidian]

Google wouldn’t give me my ai credits bcs its not supported in my country

[minor rune]

For that check your DM

[empty thistle]

Google announced they released veo 3 in Australia, but I still can't access it. Does anyone have any advice for accessing veo 3 in Australia?

[crisp basalt]

veo 3 or flow?

veo 3 and flow aren’t the same thing

if you want to use the veo 3 model, you can just use the Gemini app.

but you can also access the veo 3 model through the Flow website.
basically, Flow is just another way to run the same model, but with its own workspace and tools.

[gilded shadow]

At first when I saw Veo3 videos and created a few, I thought it was already ready primed for social media creation, maybe even commercials. Veo 3 was obviously a huge leap forward. After using it enough, it's still not there in terms of adherence to prompts. It'll always improve and never quite "be there" I realize. Every once in a while, it will match the prompt exactly or through randomness, create something brilliant. It does seem like a lot of repeated tries to get it just right. Anyone else feel that way?

(It's still amazing, revolutionary, and I love using it)

[lone obsidian]

How do people make long versions?

[split breach]

just got pro, but I can't use Flow; I'm just stuck at this screen

[empty thistle]

My sound is not working for veo 3, any suggestions?

[sly mason]

Trying zooming out and see if u see the button to click "Get Started"

[gilded shadow]

What mode are you using? Did you set the Quality to Highest Quality (Veo 3)?

[empty thistle]

Yeah just fixed it, didn't know where to locate veo 3, thanks👍

Is there a section in this server with flow prompt templates, or is it just recommended to have a look at prompts other people have made?

[crisp basalt]

deepmind website got many prompt examples

https://deepmind.google/models/veo/

Alternatively Flow TV is a great place to start

oh my god

[gilded shadow]

in addition to what @crisp basalt said, i find using llm ai good to enhance my prompts. it gives ideas and even full prompts. i usually edit them down becuase they're a bit too detailed and seems to confuse veo 3 at times. but it usually adds depth to my prompts. i use chatgpt but can use any.

[crisp basalt]

correct

im using both gemini and chatgpt to iterate my prompts

im working or should i say vibe coding this simple React app rn 👀

[mortal stone]

4 tries with no sound 400 credits gone 😠 :

[feral jetty]

how many credits will i get in flow if i buy ultra pro

[empty thistle]

12,500

[mossy venture]

i got google ai pro 1 but when i wanna generate something with flow it says i dont have enough credits

[tacit hull]

Where can I see how many credits I have? Or how to get them
Flow always says I have not enough credits to use but I have a Google ai Pro subscription since November 🤔

[mossy venture]

same problem i can only generate with gemini but its kinda weird

[tacit hull]

I was to slow 😄 just saw your text after sending mine

[crisp basalt]

Have you check your ai credit page?

[mossy venture]

if i click on upgrade it goes to google one ai page or something and says im already subscribed

[crisp basalt]

https://accounts.google.com/AccountChooser?continue=https%3A%2F%2Fone.google.com%2Fai%2Factivity%3Futm_source%3Dflow%26utm_medium%3Dweb%26utm_campaign%3Dflow_ai_credits_page

I’ve done doxed myself just now lmao

[mossy venture]

lol

why does my videos give no sound

[mortal stone]

we all have that i guess

[mossy venture]

only the first vid i generated gave me sound, the rest are just muted lol

[mortal stone]

would it be possible to add an option in "flag output" like "no sound" and have credits refunded once checked by Flow's team ?

[crisp basalt]

What was the prompt for the no sound video?

[mossy venture]

i did like a million different things, even asking to make sure there is sound etc

do you got a way to get sound in my video

[crisp basalt]

The thing is, Veo 3 will generate sound no matter how you prompt it for example, even if you don’t include any audio cues.

For example, I once generated a video without including any dialogue in the prompt, but Veo 3 added it anyway.

[mossy venture]

A Moroccan boy vlogging and saying, 'Welcome to marrakech everyone! Look at this beautiful village where my family lives

this was my simple prompt

just testing but it doesnt give me sound

even tho his mouth is talking

[crisp basalt]

The only reason it might not generate audio is if you forgot to select the Veo 3 model before hitting submit. Sometimes after generating a video, the model selection reverts back to Veo 2. I think they’re already working on fixing that

[craggy remnant]

hello team!! I m currently in france and would love to test flow! but it says not available in my country yet, any idea how to access it?

[crisp basalt]

Another possibility is the safety filter, it seems they have an internal system that blocks audio generation if it detects anything that violates their guidelines, though the video will still be generated (which I really find this to be abit annoying )

[mossy venture]

im not generating this on flow btw im using veo 3 on gemini cuz everytime i try to generate sum on flow it says upgrade to get credits but when i click on it, it says already subscribed

[crisp basalt]

Hmmm

[crisp basalt]

Did you use Gemini web or app on mobile?

[mossy venture]

web

im on pc

[crisp basalt]

VPN and US Google Pay profile

[mossy venture]

but the first vid gave me sound

[crisp basalt]

Was the first vid include prompt that involves Moroccan boy

[mossy venture]

no

just some random gorilla talking but that one was horribly bad

he was munching on a banana while talking clearly

sounded more like a voice over

[crisp basalt]

try this

A documentary-style shot of a Moroccan boy vlogging, holding a phone and looking directly at the camera. He is standing in a beautiful, sun-drenched village near Marrakech, surrounded by traditional mud-brick buildings and narrow alleyways. The scene captures the authentic feel of rural Morocco. Basic audio includes the boy's voice speaking the line, 'Welcome to marrakech everyone! Look at this beautiful village where my family lives,' overlaid with the subtle ambient sounds of the village environment.

[mossy venture]

did you ask ai to write this prompt?

[crisp basalt]

of course

Keep running your prompt through AI until it actually sounds right and does what you want. Don’t settle for your first draft , iterate till its good to go 👌

[mossy venture]

do you got the pro plan or the cheapest

[fresh steeple]

Soooooo.... answering my own question here. We have 20 Veo3 generations per month with a limit of 10-max per day. I'm talking about the GEMINI website, not Flow. Interestingly, Gemini mobile App only offers Veo2.

[crisp basalt]

pro plan is the cheapest haha

[crisp basalt]

i made 7 vid Veo 3 on gemini, that was the max i could make apparently

[mossy venture]

why is there no plan that has no limit in generating\

[crisp basalt]

idk m maybe in the future?

[fresh steeple]

Hmmmm do you get the same kind of message as in my screenshot?

[crisp basalt]

yes

[mossy venture]

cuz in thought the pro plan was infinetly generating but its not xD

[crisp basalt]

i have to wait until june 6 Lmao

[fresh steeple]

weird they should be 10 generations

[crisp basalt]

its confusing tbh,

kinda weird, but every time I check, the wait time for the limit reset just keeps getting longer.

@fresh steeple you might wanna bring this up in official gemini server too, ill send u the invite link in dm

[fresh steeple]

Thanks!

[lusty relic]

is that an actual answer because i've spent the whole day looking for an answer to how much we can generate (in gemini website)

[lime crypt]

can i somehow access flow from europe?

really want to test out veo 3

[fresh steeple]

Yes, ran my own tests and another user here replicated. Also, there is no info anywhere hence our own figurint out

[lusty relic]

buy google ai pro subscription in gemini (i bought it with a french card) then get any vpn (i use surfshark) and put the vpn on to anywhere in america

[lusty relic]

i found this that says 10 then its gone

[crisp basalt]

u can also test it in gemini

[fresh steeple]

Nah is more than 10. They also get renewed

[crisp basalt]

is it really more than 10?

[lime crypt]

is there no free version or how much does the cheapest cost?

[lime crypt]

Really? How?

[crisp basalt]

let me recount how many vid i generated in gemini

[crisp basalt]

If you havent claim the pro trial before, no'ws the time

[lusty relic]

not that i know of,the subscription costs 22 euros,theres also a free trial with the subscription for one month

[crisp basalt]

https://one.google.com/about/google-ai-plans/ check this if you are eligible for the free trial

[lime crypt]

Ok thanks will do

[fresh steeple]

Yeah, already generated more than 10

[crisp basalt]

ahh

yeah i think mine is more than 10 too

[lusty relic]

it says i can generate two more videos,is that cuz i generated too much already?

[fresh steeple]

I didn't see that prompt in my testing.

[crisp basalt]

ive seen it

and yeah you can only generate 2 more vid , then youre done

[fresh steeple]

This is how mine looks like today

[lusty relic]

done for the day right? or like forever

[fresh steeple]

How many Veo3 uses have you done so far?

[crisp basalt]

it will tell you when your limit will reset dont worry lol

[lusty relic]

pretty sure almost around 8 or 7

[lime crypt]

Do i need to make a new gmail or is it ok to use my gmail that i used so far (in europe)

[crisp basalt]

to claim the trial ? yeah

[lime crypt]

Yes, i can just use any gmail?

[crisp basalt]

But if you still wanna try Flow, just make a new google account , set your Google Pay Profile to the US country (make sure you’re still on a US VPN while you're doing all of this), and grab the AI Pro trial. You’ll get access to Flow after that

[lime crypt]

So I need a new google account

Ok

[crisp basalt]

for Flow yeah

[lime crypt]

Ok thx

[winged karma]

okay，thank you for your advice

[crisp basalt]

actually i forgot to get back to you on that

sorryuhhh

i was tryin to figure out how to split idea of yours

[winged karma]

I think the"Chopin" is a problem. Because when I remove it,the veo3 was work

[lime crypt]

I created now and im in flow, but it says i have no credits

[winged karma]

No problem. thank you so much

[lime crypt]

And it displays this message, but when i click on it it says I only subscribed

[crisp basalt]

did u set your google pay profile to US?

before you claim the trial

[lime crypt]

i have created a new gmail while being on the us vpn

Where can I see the google pay profile nation

[crisp basalt]

well , i think theres no solution then becuz u claimed the trial already-

https://payments.google.com/ this is what i was talking about

[lime crypt]

I thought hwen i create while being on vpn its ok

[crisp basalt]

nope. i did told u earlier ,

[lusty relic]

thats if u wanna make videos in flow,you can also do them in the gemini website with the us vpn and it allows you to make videos with veo 3

[crisp basalt]

yeah as for now, use gemini

[lime crypt]

hmm ok

And how do i do that, when i ask gemini it tells me i need to use it somewhere else?

[winged karma]

I love veo3

[lime crypt]

Found it

[shy depot]

When will Veo 3 roll out to the Netherlands?

[edgy pasture]

Thank you🔥

[frigid pier]

the body isn't morphing amazing

[amber haven]

WHERE is Character Controls (vid2vid) & Motion Master for Veo 2 ??????????

[crisp basalt]

in the land of hopes and dream

[amber haven]

They r listed on the website tho ?!

[crisp basalt]

yes

and at the same time , in the land of hopes and dream

[jade crest]

I’m currently a Pro user. If I run out of credits, can I top up like Ultra users do?

[idle compass]

no, I don't believe you can. Topping up is only for Ultra subscribers, I believe. Your access to Veo 3 via Pro is like a trial, with the hope you will upgrade to Ultra I guess. Plus if they did offer the ability to just buy more credits to pro users, then the ultra folks would be pissed i imagine.

[crisp basalt]

angry 💰 ultra folks

[jade crest]

There’s currently no third-party platform where you can try out Veo 3 either.
I’m a bit concerned that their review process might be too strict — even after purchasing, there could be many times when it’s not usable. 🤣

[crisp basalt]

its gonna be on fal.ai soon

[jade crest]

wow

[idle crest]

So May 19,2025’s BOOM in popularity is how Flow Became paid? Fricking 4chan locusts.

[idle compass]

I saw a third party you can access it... Let me see I I can recall it through my history

[idle crest]

@edgy holly regarding your chart, the boom in popularity on May 19th might have been caused by 4chan spreading the news of Veo3. Whenever that happens, a certain group of people within 4chan abuse the credits system, causing sites to lose money and rely on subscriptions. Not that it matters because thanks to YouTube there’s tutorials on how to get Veo 3 for free.

[crisp basalt]

Turns out, Veo 3 now uses prompt rewrite enhancement by default, even on the API. Found that out today

[idle crest]

This one right here

[crisp basalt]

so that explains the random dialogue in some of my video

[idle compass]

Here is a third party that has deployed it by the looks of it. https://veo3ai.ai/

[crisp basalt]

I do hope the Veo 3 in Flow doesnt have this on by default.

[idle compass]

My guess is that there is definitely some kind of prompt enhancement going on once you hit generate, via all platforms - incl via Flow or natively in Gemini app.

[raw swan]

where can i read this document?

[crisp basalt]

It’s a bit ironic for me , I used an LLM to improve my prompt, but I didnt expect Veo 3 to apply another layer of enhancement on top of my already-refined prompt

[muted cape]

the generations will always differ with the same prompt, there's a random seed

[crisp basalt]

oh its on vertex ai

[crisp basalt]

https://cloud.google.com/vertex-ai/generative-ai/docs/video/generate-videos

Also, if anyone was hoping for 9:16 ratio in Veo 3… yeah, you can probably forget about it lol

[muted cape]

sometimes llms can make prompts worse. sometimes better. you have to understand the prompting style a bit and it is model dependent. i train LoRAs often using llms to enhance or auto caption.

[brisk swan]

Hows this one in google cloud different then flow?

[crisp basalt]

more control to parameter

more technical

[brisk swan]

Nice is it the same qaulity?

[crisp basalt]

of course

[brisk swan]

Nice thanks

[knotty needle]

hi, flow is unavailable?

[brisk swan]

Flow looks better then the google cloud version.

[crisp basalt]

oh u tried it alerady?

[brisk swan]

No, im talking about the ui lol

[crisp basalt]

oh

of course flow is better haha

[brisk swan]

also good cloud doesn't have a veo 3

I only saw a option for veo 2

[crisp basalt]

its for selected users only

[dawn lake]

Does anybody know when it will be available in the Netherlands?

[shy depot]

Niemand heeft antwoord daarvoor haha

Maar ik heb wel een idee

[brisk swan]

Hey guys is it true google ads destroys your youtube channel

is it better to invest that money into AI Ultra with flow?

My organic AVD Is 50% 🙂

[mellow gulch]

Can anyone tell me why I'm having errors when trying to generate videos using VEO 3?

It´s my prompt:

A 3D wireframe model of a humanoid character rotating slowly on screen. The model is made of triangles and polygons, showing the mesh structure clearly. As it rotates, glowing vector arrows appear at key points (like joints or limbs), visualizing directions and forces. On-screen label appears: “Geometry & Vectors”. Background: a clean digital environment (dark with a subtle grid floor, like a game engine viewport). Background sound: subtle futuristic synth hum with soft tech clicks, creating an analytical and high-tech mood. Narration: “Geometry defines the shape. Vectors control movement and direction.”

[shy depot]

Go dms

[pearl dagger]

Maybe someone can help me:
I tried to use Google Flow today but it does not work. I always get "You need more AI credits to complete this request".
I have just subscribed to Google One AI Pro, so I have not used any AI credits yet. Any idea someone?

[last crane]

Try waiting a bit

And check your remaining credits

[dull mauve]

hey guys, whats the best work flow for wanting to create videos longer than 8 seconds with same characters. I work with crypto meme projects and I would love to experiment with Google Veo 3 and having a consistent character with the same characteristics in each video. How would I go about it?

Would something like this make sense?
Generate character in ingredient > create clips with Veo 3 > post production in external video software

Im a novice to everything video related so would love some guidance thanks!

[fierce jackal]

See yall next month, lol used all my creds on veo3 this month

[warped arrow]

i feel like the AI is playing dumb sometimes, so they can burn our credits quickly, lol... like I told my character to walk backward, away from the camera and it keeps walking towards me, lol

[fierce jackal]

lol I know the sound is experimental but I feel like it’s purposefully not adding sound to most of what I generated

[idle compass]

Extend doesn't really work yet with Veo3. You will lose the sound, and it will likely bump you down to Veo2 (the older model). Ingredients also does not work with veo3. Google need to be clearer on this. The only real way to extend your first 8 second clip, is to try the 'Jump To' feature in Flow or keep tweaking your prompt, and generate multiple 8 second clips that you can then stitch together in your own video editing app. But this is challenging, and you are likely to burn through your credits fast. This is why you will not find many extended videos being posted anywhere that feature the same consistent character/s, because currently it is extremely challenging and expensive to pull off.

[idle compass]

Try using Veo3 within the Gemini App instead. At the moment, so few features actually work in Flow with veo3 anyway, you will get a similar experience in terms of producing clips just using the gemini app.

[tepid geyser]

I have to disagree slightly — there is a noticeable difference between Veo3 in the Gemini app and in Flow. While the underlying model quality is likely the same (no turbo or downgraded mode in Gemini), the key issue is prompt handling. Gemini tends to automatically adjust or “enhance” your prompts, which means you often don’t get exactly what you asked for. In contrast, Flow gives you results that adhere more strictly to your original prompt. That level of control makes a big difference for me.

[dry kite]

Hi guys, just payed for google pro. but i still cant use veo 3, someone can help?

[reef wraith]

Need ultra

[fierce jackal]

You don’t need ultra, pro can do like 10 videos on veo 3 a month, just need ultra for more

[fierce jackal]

You are in flow right?

[dusky slate]

hi! are there any news of when START FRAME + END FRAME will be able to be used at the same time?

[reef wraith]

Veo3 is hard to work with right now. It's not quite there yet. But Im very excited about whats coming.

[echo sandal]

How do I access Veo 3? I only have access to Veo 2

[reef wraith]

Go over to Flow

And then make sure you have the right subcripotion level

[wind epoch]

anybody know how to get the veo 3 to generate the same people in a video after they generate them a first time?

[reef wraith]

Im struggling with that

[dusky slate]

maybe its a region lock? currently only working with start frame, not both.
if I choose start frame and end frame, the model switch to veo 2 automatically

[dry kite]

yes

[fierce jackal]

In the prompt box see if there’s an option for highest quality or what have you, will also have experimental audio

[dry kite]

its saying for subs only but i have 1 sub already

Google AI Pro fits for experimental audio?

Also, in quality its saying "Fast"

[tepid geyser]

huhh!

[reef wraith]

I get these errors a lot

[tepid geyser]

yeha

[lusty relic]

Flow still doesn’t work smhhh

[fiery radish]

I am waiting for the announcement where it says unlimited relaxed mode generation for Ultra subscription 😄

[minor rune]

where u from?

[lusty relic]

I’m from Morocco,subscribed with French card,using USA vpn (also Gemini works with veo 3)

[minor rune]

ok it's normal that it doesn"'t work

[lusty relic]

dang it

What should I do?subscribe with one of the countries listed?

[minor rune]

yes

[lusty relic]

Alr thx for the help

Wait does anyone have a link for the countries that have access?

[minor rune]

#flow-announcements・📣 message

[lusty relic]

Thx

[bright marten]

hmm... am I the only one getting a "Say Hello to Veo 3..." overlay/modal that I can't click out of? 😢

There's an unclickable "Get Started" button. Updated Chrome, tried inspect mode. No luck!

[tawny lark]

are u from ? i got the same am from Poland

[brisk swan]

When google AI Ultra is 50% off but still super expensive 🥲

[dusky slate]

maybe this one? #general・🧺 message

[shut harbor]

Same here

:/

[main saddle]

just updated my announcement with the link! #flow-announcements・📣 message

[proven crown]

any news on availability for the EU?

[latent briar]

is flow not working right now for y'all as well?

[lusty relic]

thanks for that

[proud rivet]

hi! anyone here with a Pro has ingredient working?

[vague solar]

yeah it's not working for me right now

[lean crown]

Hi everyone, i have a friend that want to generate a quick 30 seconds of a clip with Ai, i suggested Veo3 With Flow but i am wondering if i can integrate the text of the music so the lips of the fictional person in the video are in sync. Is it possible to do that? If its not clear just tell me haha

[latent briar]

mine started to work again

[lean crown]

& if i get the ultra version, am i still going to get the * unable to generate, try again later * ?

[vague solar]

alright, I'm gonna re-do

[latent briar]

you want to generate a clip of someone mouthing the lines.. then you want to dub it?

because that's the only feasible way that works I feel

[lean crown]

He did a music but he want to put it in the 30 seconds of the ai video so is there a way to telecharge the music so its the same

[brisk swan]

I started making trailers with flow and i wanted to share some amazing RESULTS with only using fast quality.

[lean crown]

What do you mean by dub ?

[latent briar]

here.. you basically want to generate a clip of someone singing along to the lyrics of a song

[lean crown]

Yes exact

[proud rivet]

i'd like to test ingredients before go to Ultra, anyone here already tested it? tried character consistency?

[real field]

You will need an additional tool to make the lipsync happen.

1. generate your videos in veo3 but have the main character not speak. veo3 is great if you want some background audio of what's happening in the scene.
2. Use Live Portrait in Comfy ui to make the veo3 videos lipsynced (there's quite a few steps for this process).
3. Master in Resolve/Premiere.

[lean crown]

Thanks man 🤝

[midnight kelp]

hey is veo 3 actually available for pro subscribers now? the website still says veo 2

[wraith basin]

when will it come to other countries?

[latent briar]

wait this so genius and will save me sooooo much time lmao what a shout

[warped moon]

What is the current limit of VO3 in Gemini for the Ultra plan, and how much is it to purchase extra credits for the Ultra plan?

[fickle chasm]

Had some fun with VEO 3 🍌 😅

https://www.youtube.com/watch?v=4h0b9AvUmzs

[latent briar]

Awesome. Just well done. That was Cinema

Please post that everywhere lol

[fickle chasm]

Thank you!

[flint veldt]

Running into that video generation limit bug in Gemini chat for Veo3, but it seems like Flow is not using Veo3 yet (no audio), can anybody else confirm that?

[latent briar]

Also you should really put a watermark on these

It’d SUCK to have people repost without getting credit

[rocky fossil]

Comfyui is going to have a lot of a learning curve, even just to get it installed. You are likely better off using a tool like hedra for lipsync. There is a hedra lip sync api on the tool Krea if interested

[flint veldt]

https://support.google.com/gemini/thread/345412627/you’ve-reached-your-video-generation-limit?hl=en

Not according to all the google community posts where Mods are confirming the "You’ve reached your video generation limit" is a bug, not a feature

Sure but most people including myself are hitting that limit at 4-6 video generations per day. I just found out from another user that actually I can use Veo3 in Flow, I just have to click the slider icon and I'm not encountering a limit issue anymore - so def a bug

well if anybody else needs help on the issue, this post seems to lay it all out:

That pesky “five videos per day” wall shows up only in Gemini / AI Studio. Switch to Flow (the dedicated filmmaking UI) and the rules change:

• Flow counts credits, not days.

• Ultra drops 12,500 credits into your account each billing cycle.

• A Highest-Quality Veo 3 render costs 150 credits, giving you roughly 83 top-tier clips per month—all usable in one marathon session if you like.

• As long as the credit counter in Flow stays above zero, you can keep queuing clips; the only brake is a brief concurrency throttle when the backend gets crowded.

Meanwhile, Gemini / AI Studio:

• Still burns credits and enforces a soft burst cap (most users hit a 3-to-5-clip ceiling every rolling 24 hours). That’s the banner that cries “out of ammo” even when you have credits left.

Bottom line: Heavy generation belongs in Flow—watch the credit meter, not the clock. And yes, despite the “US-only” fine print, Ultra works just fine here in Korea.

PS. In Flow, flip the dropdown to Highest Quality (Veo 3)—the UI still loads on the quicker Fast(or Quality, Veo 2) preset by default.

[mental mulch]

How do I buy more credits?

Do I have to be a ultra member to buy credits

[minor rune]

Yes

[hushed temple]

I cannot get sound in my clips to save my life...What I am doing wrong? Setting: Kitchen breakfast table
Timmy talks excitedly to his mom who's making lunch
Timmy says: "Mom! My teacher has fangs and white..."
While not looking at Timmy seemingly distracted the Mom says: "That's nice, honey."
Quick cut to Timmy's frustrated face.

[cinder geyser]

on gemini You've reached your video generation limit until (6th of june 11:37 am) not 24hours? is this a bug?

[long needle]

in VEO3 prompt (txt2video) is there a word limit? or optimal range of words that it performs best ... im assuming there is a prompt rewrite behinds the scenes.

[warm flint]

Per day? I hit mine last week and it says it's going to reset only next week

[kind palm]

Hey team! Does Veo 3 video generation come with audio? Can one also edit the in-video text? The clips I've generated so far don't have audio

[flint summit]

I'm getting more failed generations than usual. In fact I've only had 1 out of the last 5 work across all the models.

[cinder geyser]

some are correct audio some are fail or no audio

[craggy tiger]

hello guys i hope you are doing good, is there anyone here who can give me an advice on how to use flow in germany since its not available here? I know that there is a workaround but VPN wont work. Thanks in advance!

[somber ice]

When using vpn
Check through myip website if the vpn is actually locating you to America
I had the same problem using opera's vpn , when I switched to another vpn (it was urban vpn for me) and checked my ip from myip , it was now correctly locating me to America and then it worked for me

[tropic gale]

Trying to extend or jump to new videos usually just uses the last background and overlays something. Am I doing something wrong? As well as that these features are only working on veo2 and not veo3. Any advice would be appreciated.

[craggy tiger]

got it the only problem is i need to create an american google account but dont have an american number in order to verify 😦

[graceful nymph]

This is Veo3 right ?

[crisp basalt]

robby, there should be model indicator tag on the right

[graceful nymph]

I found issue

If you choose fast its veo2

[crisp basalt]

Quality and Fast is Veo 2

Highest Quality is Veo 3

[graceful nymph]

I wonder if thats why I was getting only 8 second videos

[crisp basalt]

The max length for a video in Google Veo is 8 seconds, though

If u see any1 posted Veo 3 videos that go past 8 seconds, just assume they were stitched or post-edited together after generation

[gray delta]

||A macro timelapse begins in a minimalist white studio, focused on a closed bud sculpted from slabs of Italian marble. Its outer layers are smooth and pale, veined with Carrara-grey striations, appearing like folded stone origami. The lighting is soft and angled, casting crisp shadows that reveal the weight and polish of each layer. As the bloom unfolds, the marble petals crack slightly apart, then tilt outward in smooth, tectonic motion. Each new layer reveals a different marble variety: soft pink of Breccia, dark veins of Nero Marquina, and luminous green Verde Alpi. The petals reflect light with a cool sheen, while matte veins give depth to each surface. The camera glides inward, revealing microfractures, edge chipping, and high polish. At full bloom, the flower holds wide and still—each mineral petal gleaming in harmony. The scene ends on a close-up of the center swirl, revealing fossil-like flecks embedded deep in the marble’s core.||

[visual dust]

para que sirve esto?

[lavish lava]

What do they mean by increasing capacity for ultra subscribers It still costs 100 credits per generation #flow-announcements・📣 message

Very vauge thing to say

[lavish lava]

It was already 100 credits per generation before this update I think it must be a delayed announcement on the server @edgy holly

[crisp basalt]

it was 150 tho

on day 1 release

[lavish lava]

For me it was always 100 credits

And I got aussie access a while ago too

[crisp basalt]

u were probably not there before the flow expand into more countries , it was 150 credit for veo 3

[lavish lava]

I did see a similar announcement on X which is why i think its a delayed announcement

[crisp basalt]

they lowered it down to 100

[frigid plover]

For those who are having trouble generating videos with audio...

Veo 3 does not support picture continuity with audio

[crisp basalt]

it does

[lavish lava]

Oh I see thanks. I signed up 5 days ago for ultra

[crisp basalt]

just not with speech

[frigid plover]

I found it in their FAQ. I just wasted 3k credits.

[lavish lava]

Ouch!

[flint raven]

hello

[crisp basalt]

its in the faq lol

[frigid plover]

Yes, the problem is that when using the tool to continue or generate videos from an image or something like that, VEO 3 does not work well. In order to generate with audio, it needs to be without any resources, just text for video.

Unfortunately I only discovered this late.

[crisp basalt]

it can do audio.. just not with speech/dialogue

for image to video

[frigid plover]

True, I noticed that.

[crisp basalt]

Out of curiosity, how many videos did it take before u realized the audio wasnt being generated?

[frigid plover]

I realized right away that the first time it failed, but I was persistent, I deleted it, reported it, and I kept doing that for a long time. I reviewed tutorials, researched... then I tried again, changed the prompt... until I found this information in the FAQ and the information matched my problem. I spent about 3k credits.

[crisp basalt]

oof,, well at least u flagged the failed output before u deleted it i guess

[frigid plover]

Yes, I did that for the most part.

[summer holly]

at 18:39 AI Search uploads a start frame to Veo 3 (generated video has sound). So is that new? Im sure Veo 3 didnt support that at launch. https://youtu.be/_7LxEhg63Zk

[crisp basalt]

api i assume

[muted cape]

Pretty sure you can't get Veo3 with audio using image to video. I'm also surprised the video had Will Smith in it there as others have had trouble with public figures and policy. All that said I have seen some bugs when I originally tried to use image to video, the extend feature, etc. So it's my assumption it isn't supported yet, but it really all could simply be a bug. Which would be pretty unfortunate given I went a completely different direction with what I generated because of the lack of functionality.

Not to mention the waste in credits trying.

[crisp basalt]

You can,

they stated in the faq

Environmental audio still generates for image to video veo 3 in Flow

[muted cape]

"Speech is only available in Text to Video mode"

[cinder geyser]

i tested on image to video has no audio...

[muted cape]

I'm still surprised the video in that YouTube video even had the characters attempt to speak. I didn't expect audio at all given their FAQ and everything I've seen.

[cinder geyser]

only enviromental sound no speech sound...mny credits are gone T_T

[muted cape]

I guess the attempt to speak is an environmental sound 😂

[grand zealot]

Are failed generation errors common? Apologies I'm new to using Flow. Trying to generate Frames to Video (from a Imagen start frame) in Veo3 - the Veo2 generation only generated one but still charged me for 2, and then the Veo3 has failed twice already.

[muted cape]

They refund for errors (unless there was a problem with that), if you click on your user menu button in the top right and then click on the "123 AI Credits" text it takes you to a transaction log. You should see usage and credits refunded there.

[grand zealot]

Got that part, was more questioning the failed generations.

[muted cape]

Did any other message pop up there? Sometimes they stack on top of each other but do animate.

[grand zealot]

Ah, I'll pay attention this time. Thanks.

[muted cape]

They might have some additional info. Sometimes things like potential issues with policy.

I missed a bunch of those at first because I switched tabs and looked at other things while waiting.

The messages are only fleeting with no other log of information.

Nonetheless, there are also bugs as well. You'll run into failed generations every now and then.

[grand zealot]

Okay, I think I figured it out 😄 Because I didn't specify what clothes it was wearing, and there is no clothing shown in the closeup image, it assumed I wanted nudity.

[muted cape]

ah, that could be it

[mental mulch]

if i generate veo 3 videos on gemini pro are they free

since i have pro

[muted cape]

not that I'm aware, I think everyone would be all over free veo3 videos 🙂

[mental mulch]

i mean i have google ai pro currently

[muted cape]

i had pro, upgraded to ultra, all veo generations seem to cost me money

[dusk juniper]

Hi! How do I check my credits? Where can I see my remaining credits and how much I've spent?

[crisp basalt]

wait oops

[crisp basalt]

https://one.google.com/u/1/ai/activity?utm_source=flow&utm_medium=web&utm_campaign=flow_ai_credits_page&pli=1

[dusk juniper]

Awesome! Thank you!

[gilded quartz]

Trying something

[pale sentinel]

Hey everyone! I have google one subscription but I can't use Flow and Veo3 for some reason. It says I need to have more points. But I can do it inside Gemini pro interface, but its limited with setting. Does anyone know why this happening?

[gilded quartz]

It's say you are out of credit

[gilded quartz]

I can't generate in Veo3
I am a subscriber
Only thing I can generate is veo 2 + without sound like this results

Anyone know why ?

[pale sentinel]

Yes, but when I click Upgrade, it saying I already signed in to google one and Google AI is included. I have this message from very first use. I mean I haven't generate even 1 video. But starng thing is that I can do it in Gemini pro

[forest raptor]

250 bucks is insane 😭

I wonder how much it'll cost when it'll be capable to make long videos. Maybe like an actual film.

Kinda makes the ai alternative pointless

[gilded quartz]

I think there is a way for free

[forest raptor]

is there??

[gilded quartz]

Yah

[forest raptor]

I'd have heard about it

[gilded quartz]

Student account
And the price will be 0

[forest raptor]

I am not a student unfortunately

[gilded quartz]

You can create one 😅

[forest raptor]

It can't be this easy

[gilded quartz]

It is

[gilded quartz]

@forest raptor
What you think about the video

[forest raptor]

Which

[gilded quartz]

This one

[forest raptor]

Lovely

[pallid jetty]

Does anyone know why this happens when I want to use flow?

[wicked galleon]

Where do I share my veo 3 creations ?

[lament ivy]

I am subscribed to AI Pro, but when I check my credits, it shows That My Monthly AI credits are 0. Why?

[stiff harness]

It says why in your screenshot.

[crisp basalt]

for real 😭

[hearty flax]

im getting the same problem

any help?

[dry meadow]

Text to video

[arctic surge]

動画を生成できるWhisk Animateと、Flow（Veo 2、Veo 3）のAIクレジットは共通なので、Whisk Animateで100本の動画を生成してしまったらFlowでの動画生成は出来なくなります。間違っていたらすみません。

Since the AI ​​credits for Whisk Animate, which can generate videos, and Flow (Veo 2, Veo 3) are the same, if you generate 100 videos with Whisk Animate, you will no longer be able to generate videos with Flow. Sorry if I'm wrong.

[tall kestrel]

veo3 can use image to video now？

[rancid sun]

I'm from EU and I managed to register to Flow and bought subscription (via VPN and incognito browser). It worked well for a couple of days until yesterday. I can't log in, google doesn't recognize the account.

Does anyone have any suggestions how to bypass that?

[forest raptor]

that's what you get for throwing 250 bucks like that 😒

[rancid sun]

thank you for your help. Very insightful.

[ripe aspen]

what i can do to work from Armenia? pls help 🙂

maybe can some one generate 1 video for me ? 🙂 pls

[urban estuary]

same, i really want to try it

[tall kestrel]

yes

it working

[proud radish]

when will Flow available in Thailand?

[shell mason]

is ingredients to video not working with veo 3? I can't get it to work, it always swaps back to Veo 2 when I try to run it.

[marsh fox]

any release date for EU?

[thorny hedge]

When will Flow available in Russia or Ukraine?

[frigid pier]

I thought Europe already has it?

[rose comet]

Is it true that only text-generated videos can have sound now? When veo3 was first released, I used pictures to generate videos and they also had sound.😩 😩

[midnight fox]

Is flow’s “extend” (veo3) available for Pro? Or Ultra only?

[placid sparrow]

It is currently available in around 70 countries I believe, it will be released for more countries soon

[placid sparrow]

Flow, our AI filmmaking tool, is now available in 71 more countries 🌍 through the Google AI Pro and Ultra subscriptions. for the full list of countries, see: https://labs.google/fx/tools/flow/faq,
Veo 3 is now available for Pro subscribers to try (create your videos with sound!), and we’ve increased capacity even more for Ultra subscribers. ✨,

[midnight fox]

Right, specifically asking about the extend features because I am having trouble blending two veo3 together and didn’t know if it was me or an ultra feature

[opaque rock]

Came here to ask that... I have no audio on Veo3 when using an uploaded image as the first frame

Kinda make it impossible now to create a normal video with the scenebuilder. if you want audio, each scene will generate a completely different video without the image reference.

[hard plinth]

same for me

[rose comet]

A few days ago, I used frames to generate a video with sound, but now it doesn't. Is this a bug? When will it be fixed?

[opaque rock]

honestly they should also refund the credits. I burned many credits trying to figure out what's happening.

[rose comet]

Yes, I experimented a lot with the sound problem, and found that the official statement said that only text can generate sound. In this case, it becomes extremely difficult to generate a consistent short film.

[shell mason]

I'm having trouble with dialogues to have the right person saying what they're supposed to say, often the roles are reversed, anyone knows how to fix this?

[rough ginkgo]

Welp exhausted all my credits on failed attempts to have audio in my frame to video in Veo3.

[swift thicket]

does flow allow to create in 9:16 or other aspect ratios? im prompting it to but its not creating in that format

[errant dirge]

only 9:16

[swift thicket]

its generating in 16:9 only now

[warm garden]

Can someone tell me why I had credits in my account yesterday, and today I log into Flow and it says I have nothing and that I even need to subscribe? I don't know if I'm the only one this has happened to.

[hardy wagon]

hi everyone, is flow/Veo 3 capable of doing style transfer? For example, if I upload a short gameplay video of GTA and ask it to recreate it in the style of Roblox? Composition should be the same, but the props/characters should have the Roblox style. It's not available in finland yet for some reason, as I would've tested it myself

[steady vine]

I'm still confused on how to acess flow itself

how do I access flow? I can only create videos through gemini

[rough ginkgo]

https://labs.google/fx/tools/flow

[wise cypress]

frames to video no audio 😦 😦 😦 on veo3

im just trynig to use one frame to start with

[brisk swan]

when is veo4 coming out

(That would be 1000 credits plz) lol

[wise cypress]

oh frames to video = no audio just saw it in the feedback

thats silly

[steady vine]

thank you, however it gets stuck on this page, how do I acess the tool ?

[vapid hull]

Upon video generation failure or flagged, there should be a way to rescue/recover the prompt. luckily I copy it to notepad on the side, but the the Flow itself should make it friendlier to not have to do this. Sometimes there are tweaks that are done in flow that I forget to copy over and then I have to retrace my steps and think which words I used for the prompt that is no longer available.

[humble coral]

hey guys honestly what can we do to use this Flow or veo3 AI just curious

i love memes and that was the only thing that im thinkin of

[wise cypress]

can i request my ultra account not be given early unstable version that waste all my credits its like a scam to take my credits for generations that are completely lacking audio

and it is account based if i use like an alat account on a pro membership it works just fine

[fierce nimbus]

Try to reduce the browser magnification. The button should be located below. I have tried to repro and am unable.

[fierce nimbus]

Frame To Video supports SFX but not speech!

[jade crest]

Why does it feel like Veo3 distorts cartoon characters and isn't even as good as Veo2? 😭

[steady vine]

thank youuuuu

[surreal dust]

Is flow with veo3 available to create videos in 9:16 format? 🤔

[broken pelican]

when trying to access veo3 it says i need ai credits while i am already subcribed, has anyone had this issue?

[silent carbon]

I have a question, i'm from the netherlands and i want to acces flow but i can't. it says "
Flow is not available in your country yet." does anybody know how to bypass it..??

Should i jst use VPN or something?

[steady vine]

do y'all know if with a ultra ai subscription I can do videos longer than 8 seconds ?

[remote grotto]

Is flow available in every country with the ultra sub ?

[viscid fjord]

no

[clear fox]

anyone else getting failed generation?

[remote grotto]

i wish i had XD

[clear fox]

same 8 seconds

[steady vine]

Oh okay, i've seen some people creating longer videos, but that must be a different tool then

The prompts I used for veo 3 in gemini apparently doesnt work with flow for some reason, an error pops up

I figuered it out, apparently the prompts needs to be simpler

[jade crest]

Will access to Veo2 be discontinued soon?

[warped arrow]

are you referring to just a single clip? beacuse you can extend or do a "jump to" option to a clip to make a scene longer, using the scenebuilder

[dusky hearth]

Hello all 🙌

[upbeat mauve]

I'm having the same issue, does someone knows how to pass through this screen? I'm from Brazil and I don't know if this is a bug or the tool isn't available yet

[steady vine]

diminui a tela, segura ctrl e roda a bolinha do mouse pra baixo q vc va iver o botao

[upbeat mauve]

top mano valeu hahahah

[steady vine]

oh really??? so you can do a continuation? I didnt know that

[broken cairn]

Not until Veo 3 supports all elements that Veo 2 does, like ingredients and Frames and extend, which I really hope are all added to the highest quality version of Veo 3, and not just fast.

Also, I cannot wait until we can get speech generation within frames videos

[idle crest]

same here 🙁

Im waiting for mine to reset

[flint summit]

Is it just me or are we all getting more "failed generations" than we used to? Is it just that they can't keep up with the increased demand as more regions get access?

[idle crest]

I shouldnt have to wait for a month for them to retop my credits, especially not when im on student plan

[warped arrow]

yup, watch their basic tutorial in youtube

https://www.youtube.com/watch?v=9nVEfjmDlVk

[idle crest]

ill keep using veo2 so HAHAHA

[feral skiff]

Hi, i just got a free trial for ai pro but cant access flow?

[idle crest]

I love veo 3 tho, and wish there was a middle ground, not everybody is richy rich like google is

I shouldnt have to wait a month for 10 generations,

about how much?

[surreal dust]

please help . Is flow with veo3 available to create videos in 9:16 format?

[feral skiff]

Does the free trial not have any generations?

[idle crest]

they doI got the 15 month free trial that only requires me to pay on the 2026, screw that im cancelling my subscription then

100$ is alot

[feral skiff]

I got a monthly free trial for google ai pro but i cant use flow?

[idle crest]

are you in us

It costs $200

[feral skiff]

yep

it says im not subscribed, but when i click subscribe, the other page says im already subscribed

[latent elbow]

<@&1173646484311384114> Can you put how many days are left until token reset?

[ripe mirage]

AI pro gives access to flow and even Veo3. It’s just a very limited number of uses.

[brisk swan]

Monthly AI credits
Next update: Jun 27
180 🥲

[latent elbow]

180?

[feral skiff]

I have 0 uses because I can't access it, because it says im not subscribed. But when I click subscribe again, it takes me to a google page that says I'm already subscribed.

[brisk swan]

Should i get AI Ultra or 15 pizzas

[ripe mirage]

Making that same calculation right now. Food or rendering.

My short is lacking several scenes…

[brisk swan]

Fast Qaulity Only

Reminds me of interstellar

[ripe mirage]

How Veo3 makes us feel…

[mighty kayak]

What was the prompt?

[ripe mirage]

In a 16th-century Bavarian town square, an enigmatic wizard captivates a crowd with a series of increasingly astonishing magical feats, culminating in a bewildering transformation that leaves the onlookers in stunned silence.
Setting: A bustling market square in a small, fictitious Bavarian town, nestled in a valley surrounded by rolling hills dotted with dense forests. The architecture is a mixture of Late Gothic and early Renaissance influences, with half-timbered houses, steep gables, and a stone church dominating the skyline. The time is late afternoon, and the square is bathed in the warm, golden light of the setting sun, casting long shadows across the ornate cobblestones.

[mighty kayak]

I'm facing problems with narration

It's not narrating at all

even when I'm specific

[ripe mirage]

That one was kind of an accident. I hit enter intending to paste the specific feat…which was going to be pulling a bunny from a bag of holding while he is actually searching for basically smoke bombs but I really liked it.

You really never know what you are going to get…

[tight oxide]

that is a really bright beak

[mighty kayak]

You have Pro or Ultra?

[ripe mirage]

Was supposed to be laser beams for a squirrel 🐿️ which you actually see at the bottom…learning prompting.

[tight oxide]

Ahhh thank you

[ripe mirage]

Pro, and try actually saying “A narrator says” followed by what you want. That has worked 100% of the time for me.

[mighty kayak]

Cool! Using flow? Because it can only generate 10 videos a month, right?

[ripe mirage]

Flow and Gemini

[brisk swan]

fast qaulity can generate 100

[mighty kayak]

So it's unlimited but gets slower?

[brisk swan]

My fast qaulity generations look the same as veo 3 high qaulity

i use the program to actually make cool* stuff

[mighty kayak]

There's no other option with Veo3

[ripe mirage]

My take on it is flow is for consistency of characters and Gemini can be for scenes that don’t need exactly that. Just my thinking.

[mighty kayak]

This was the best I got so far

[brisk swan]

you know what would make that video better

if he was wearing a cool astronaut suit like epic spaceman!

but thats just my opinion

[shell mason]

I'm having trouble with dialogue accuracy in Veo 3. Sometimes the wrong character ends up saying a line, or the roles get reversed. Has anyone found a reliable way to control who says what?

[mighty kayak]

Yes! Just figured that out!
If you want a long scene transitioning to scenes and stuff, you want to have a full video of your character talking, and in your video editor you just transition to the takes generated by Gemini

@shell mason if you need to create 2 different videos with the same character/voice, how would you do it?

[ripe mirage]

I suspect everyone is struggling with that. My tips fwiw (ie not much) be really specific in the prompting and try not to overload the dialog too much.

My best results have been from working with Gemini on the prompts. Collaborating with AI in advanced of the rendering of your vision can help you break out the different scenes too.

[shell mason]

I tried doing that with 'Jump to' in Veo 3, the result was great except there was no audio. This seems to be a common issue with the Scene builder, so I didn't try more than once

[placid sluice]

Is Veo down?

[mighty kayak]

That happened to me

[shell mason]

Thx for your feedback. I tried to improve my prompts with the collaboration of multiple AI, still having trouble, but I will keep trying 💰

[mighty kayak]

I think Veo3 still very early, they should be fixing these things this week

[serene cipher]

How can I get my characters to talk?

[shell mason]

this week? did they made an announcement?

[shell mason]

be sure to choose Veo 3, what kind of characters do you have?

[ripe mirage]

Veo 3

[mighty kayak]

They updated the whole pricing under a week, so I guess that was to get more users to test

[serene cipher]

Two human characters. I followed the FAQ Using Veo 3 Text to Video

[shell mason]

I've rarely had this issue with humans. Did you try more than once?

[serene cipher]

Yep

[placid sluice]

is anyone else having issues with Flow?

I keep getting failed generations

[mighty kayak]

Isn't that Veo2?

[shell mason]

if you're at ease sharing your prompt I could try to guess further, but make sure that it is really Veo 3

[serene cipher]

I assume it Veo 3

[mighty kayak]

Flow should have the possibility to create multiple branches for the same take, starting different paths

[shell mason]

it should look like this

[mighty kayak]

Do the 1000 AI tokens reset every month or day?

[serene cipher]

Interior of a 1970s appliance repair shop, warm amber lighting filters through dusty venetian blinds. A cluttered workbench filled with disassembled toasters and radios. Arnie, a grumpy elderly repairman with thick glasses and worn overalls, glares at a toaster that just burned a Pop-Tart. Dotty, an older woman in a rhinestone cat sweatshirt, enters holding the scorched Pop-Tart speaking to Arnie in a Southern drawl, says “This toaster’s conducting a séance!” Arnie turns to Dotty and speaks in a gruff voice and says “It’s not haunted—it’s dignified. Like me.”

[shell mason]

sry I chose fast on that image but of course you need to choose the third option

[placid sluice]

month

[shell mason]

From my experience it works best when you place the things to be said rather in the beginning and not towards the end

[serene cipher]

Oops! Thanks!

[shell mason]

also be mindful of those: "", they often cause weird subtitles. I put things to be said after a : without quotation marks

[placid sluice]

Am I the only one Flow isn't working for?

[stoic sky]

Why is it stuck here?

[placid sluice]

So I guess it's not just me

[timber mulch]

is flow comeing to germany ?

[stoic sky]

I think it's IP or location related issue.

[flint summit]

Maybe not down, but I'm getting way too many failed generations.

[fierce nimbus]

What do you mean by this?

[honest dagger]

yo how can i bypass the location issue for flow

i havea prompt

[storm valley]

anyone know how i can generate vidfeos where it remains true to the first video generated?

[marsh fox]

Nah Can't access it in Sweden unless with a VPN which causes trouble

[crimson ridge]

Hello guys i want to use flow for video and im from Lithuania what can i do to do videos?

[mighty kayak]

I’m using it in Europe, no troubles no problems

[crimson ridge]

pleaseeeeee help me :(((

[mighty kayak]

How’s that?

[crimson ridge]

what can i do?

i dont know what to do

[oblique summit]

Flow isn't working for me

Can't believe I paid $100+ sub just to try it out

[glad cedar]

It certainly is not functioning well enough to justify the price point, especially when it goes up to $250/moi

I am having success with generating speech for characters, but they always continue moving their mouths after their speech is over. Anyone know how to resolve?

Ya I figured... I have been seeing so many clean videos online, but I guess people are just getting more lucky when they generate. Or their prompts are a little more succinct

Thank you! I have not considered the third option

[crimson ridge]

When i try to join it doesn’t let me yes i use chatgpt 4.1 newest too

but i want to use flow for video generating

[arctic gulch]

Just me or is flow all jacked up today?

Every time “failed generation” or the video is jumpy, misspellings, etc

[grizzled prairie]

Won't generate anything at all for me 😦 So frustrating!!!

[glad cedar]

Same here

[knotty needle]

anyone have a problem for failed generation??

[glad cedar]

Failed generation after failed generation

[trail gulch]

nooooo not flow

[main saddle]

thanks all for your reporting - we're working on it!

[trail gulch]

it could’ve been anyone else why flowwwwwe 😭

[halcyon vector]

one thing to consider is that when the prompt is under 30 words, it gets sent to an LLM to "enhance it" somehow and format the prompt in a way that is ideal for veo, supposedly

if the prompt is over 30 words, it better be formatted "the right way" otherwise it will lead to a worse outcome