flow-general・🌊

[charred sail]

Believe me, when it's ready for serious filmmaking.. They'll have options for all of that stuff.

but you'll have to license those rights..

[red drift]

if I buy more credits so I still get my 1000 monthly credits on top of that?

[charred sail]

yes

they should add a plugin for gemini on discord so gemini could answer the basic questions 😛

[marble hamlet]

Why are you apologising for a service which limits the creativity of it's user base? Ask other users here who've had images or videos fail for no logical reason. And I'm talking about their policies, not problems with the model.
Yes, it's Google's service they are providing to us, and we don't have to accept it if we don't like it. But the user experience is important, especially with the cost in pricing they are asking for.
I love creating with AI, I've used many different services, but the constant issues is ruining the experience for me now. I work 7 days a week, I just don't have the time to deal with the crap that goes with it. All we want is to create.

[charred sail]

beause I'm a developer and I understand the process.

And I understand you're feelings too as a user

[marble hamlet]

That makes no difference to what I'm talking about.

[charred sail]

I feel them as well

You have an experimental technology. You can't take the brakes off. It needs the training wheels

you might have good intentions. But by allowing you to do it, they let all the people with ill intentions do it as well.

better safe than sorry

i mean do you know how much google has on the line with this?

one naked pic of trump on there and it's done

i mean i'd make a dancing singing trumpian downzy in a heartbeat if they let me 😛

btw i got a warning from tiktok for quoting a song that was playing on the video as it was singing it..

that seems way more egregious

[marble hamlet]

The issue is not the technology, it's the policies.
This is getting off track, I don't want to create anything explicit. I'm just tired of my images not being accepted. What exactly do they mean by a 'Prominent' person anyhow? I create them in Flux, they're random people.

[charred sail]

the policies are there for a reason man

yeah but who judges it as explicit?

Kinda seems like you wanna live in a vacuum to be honest

I think you're expecting today what will be availabe in 5 years for 6 figures of licensing

[marble hamlet]

Why are you talking about explicit images, I'm not. You're still not understanding what's happening. It can happen with just a face.

[charred sail]

You just said you didn't want anything explicit

but who judges that? You say this pic isn't prominent or explicit. The AI is making the call

[marble hamlet]

What's a 'Prominent' face?

[charred sail]

whatever the AI thinks it is

[marble hamlet]

I'm done with this. There's no point.

[charred sail]

there really isn't

cuz you want what you want and you don't care to listen to why you cant

this conversation is a microcosm of everything wrong with the world right now btw

[maiden basalt]

Does anyone know how I can get Tony sopranos voice for characters?

[charred sail]

Veo doesn't do that

[maiden basalt]

what about my own voice?

[charred sail]

best you can do is set the character in Jersey and prompt something like, Doing his best tony soprano impersonation he says,

it's more about the setting than any kind of voice control

[charred sail]

You can add your own voice in post prod

[maiden basalt]

how?

[charred sail]

that's a bigger question. check some youtube videos on voiceovers

[maiden basalt]

gotcha

how are people doing joe rogan podcast using his voice over babies?

[charred sail]

i haven't seen them, but there's always a way if you're good with editing tools

[maiden basalt]

yeah thats what i figured

[charred sail]

and you get what you pay for with those

there's some free, some cheap.. some expensive

i use capcut, but they charge for anything too fancy

thinking about dropping it and just going back to adobe

[maiden basalt]

anytime i go somewhere they ask for money lol

ive gotten good with accents but nothing as authentic as actual voices

[charred sail]

yeah, that's about as good as it gets

[maiden basalt]

im determined to find a way

[lapis wraith]

i think they have a chain of thigns they use but im not sure

[fringe berry]

https://youtube.com/@aspholyteachings?si=HfbFdn1-PFOHXPaR

[fiery nymph]

Is there a guide that explains best practices for structuring scripts in multiple scenes for veo 3? I know each gen is 8 seconds and I want to use the “jump to” to create a longer video, unless there is a better approach. TIA!

[lapis wraith]

the best way ive figured is structure like - The feel you want or the environment,- the subject (1-2 to work iwth) - dialog

[fiery nymph]

And explicitly call out each scene in that format?

[lapis wraith]

that s what i was do. for my gta 6 karen dlc video

i had the same prompt setting the scene in majority of the clips

https://youtu.be/AmHvHxVDFKw?si=C-6iSPwPzaFZkMis

i made sure i put GTA 6 style game and when i watned gameplay i added over the shoulder ingame footage

i limited the dialoge to one character per scene majorit of the time

[charred sail]

Ask Gemini

Long story short.. structure doesn't matter to veo as much as context, but it helps people have better context when structured.

[livid egret]

This is the perfect use-case for mixing tools. Create your TS voice in another AI voice tool and then use that as the lip-matching audio in another AI video app. Veo3 doesn't do this yet, but others do.

[charred sail]

I haven't tried any automated lip syncing. Does it work well? Which program are you using?

[white forge]

Yeah, now it's fixed. Available to more country. Pro access on Flow. Thanks <@&1173646484311384114>

[lusty relic]

LETS GO I CAN USE FLOWWWWWWW

[quiet stream]

looks very cool! Did you clone ron swanson's voice also lol?

[lapis wraith]

yeh i took some clips into eleven labs

[quiet stream]

oh jsut elevenlabs? they improved since last time i tried cloning there then, how many minutes of footage did you use?

[lapis wraith]

ive wasted alot

i just got the membership to see how it worked and sounded so i didnt really pay attention but it wasnt many

you cant clone trump though

that i tried

i use fakeyou alot

i only neede. aminnute of footage

[quiet stream]

for him and other really known people there are some other sites that already do a pretty good job i think, but yeah if you like a less known voice elevenlabs seems the way to go

[quiet stream]

damn, just a minute and it was enough, nice. Btw was this v3 or v2 on elevenlabs?

[lapis wraith]

yeh i think you only need like 15 seconds

ill say v3

it was oddly too good

ive been using his voice for majority of my videos with narration

https://www.youtube.com/watch?v=t9WprV5nnsg

[vernal kernel]

How many characters can flow generally keep track of? If I like describe 6 different animals, should I expect it to nail that or am I asking for trouble?

[quiet stream]

big trouble

[lapis wraith]

2 but youll get each other finishing

each others sandwhiches

i usuakk stick with one per shot

usually* sorry

[vernal kernel]

Yeah, tried one prompt in quality mode, and it was a poopshow, so maybe not gonna waste any more credits on that. 😄

[lapis wraith]

i wouldnt waste in quality mode

id rather roll the dice 4 times then lose on 1

[vernal kernel]

Yeah, it was a dumb move by me for sure 😄

[lapis wraith]

how do we see all of the veeo 3 videos generated together

is there a gllery thats available

i know flow

[near gate]

Hi guys !

[frigid pier]

is Veo 3 fast okay for audio??

[near gate]

Can anyone help me with an uploading issue ?

[quiet stream]

in my experience yeah

[near gate]

I can’t upload my videos to any platform

[lapis wraith]

are you using an app like cap cut

or just directly sourced from gemeni

[white forge]

Let's deep on MJ + Flow I2V. Let's Start the party! Let's Flooding the Social Media Fellas! 😜🥳

[quiet stream]

midjourney image + image to video with flow? what's I2v mean?
Also why is everyone using midjourney, is it the best image generator atm?

[white forge]

"sippin coffee" Yes, my dear.

visit our insta : @vxdepth

[quiet stream]

lmao these are great man, do oyu already have yt channel with these are still generating for enough material? i would watch it
Also is that the girl from love death robot👀

[lapis wraith]

i make custom cards so i do alot of thiking ridiculous

no thats just a typical blonde hair karen

[white forge]

Nice one kard! 😍🤣

[lapis wraith]

i did a tone of these for a vdeo game

gta karen BE B entitrled

entitled DL

i did a whole mock game demo

fight mechanics, driving , fighting mechanics, stealth obnjective and calling police and safe space ability

[quiet stream]

like as a personal fun project just to test veo3?

[lapis wraith]

well i make custom cards with art ideas so this is kidna like a funy thing and testing

i did a home choking charixard csard i did a clip with audio but now i could actualy animate it

or a fear and loathing in lasvegas parody with fearow and slowking of pokemon

theres alot of fun to be had here. the whole batman parody video series i did was actually based around a bit of truth

somre artists sent me a cease adn desist so i starte firing these videos out directed at them

imagine thinking you showed some youjng artist whos boss, and they post this

https://www.youtube.com/watch?v=pb2YQq0k5rE&feature=youtu.be

i did a video around the yoshi being a pojkemon using midjourney and i dont like midjourneys video gen as much as the others

however it did give me hte idea of using 1 promt to create 6 photos in one and use them in a video like that

this one i couldnt get the video to keep consistent

[errant inlet]

https://x.com/RastinaNFT?t=85L25JomkohVnaa1mM1GmA&s=09

Rastina (@RastinaNFT) on X

Rastina is a collection of 10k 3D voxel art avatar on Blockchain

Join and see/listen to the tale of Rastina
👉🏻https://t.co/NAL6TlM8LB

X

[muted hinge]

will any event like genjam ever happen online? It sounds very exciting but NY is too far 😅

[white forge]

Wish you all, Happy Weekend! 🥰

[sand bloom]

Anyone know where I can create a 9:16 ratio generated AI video?

[charred sail]

You can do it. It's not easy to get it exactly how you want and continuity is a real pain.

all those guys were prompted

continuity with 1 is difficult enough..

[muted hinge]

i dont think you can. you will have to edit in a software

[sand bloom]

chatgpt is telling me I can add prompt and tell flow to strictly make video in 9:16 format. 50/50 chance it works

im trying that right now

[charred sail]

chatgpt is wrong

[sand bloom]

yeah i assumed lol

[muted hinge]

why would you ask chat gpt that?

[sand bloom]

Why wouldnt I ask chat gpt anything my mind desires?

[charred sail]

ask gemini 😛

[sand bloom]

ok

[lapis wraith]

my chat gpt told me that torah was in facft "hator" spalled backwards

but i still consider it a god send

i prefer chat gpt for ideas and gemeni for facts

chatgpot almost fot me introuble for defamation

[muted hinge]

because half of it is not true, or unpolished?

[muted hinge]

you should think of ideas yourself though. better in the long run

[sand bloom]

yeah thats gpt for ya

[lapis wraith]

but its got my sense of humor so its come through alot

or it knows my

[muted hinge]

chat gpt is a great tool. I use it every day. but don't relly on it too much

[lapis wraith]

my work is all basically parody so its comes in handy but still i asked it ealirer if thiswas infact defamation

"its riding a tight line:"

[muted hinge]

parody? do you post online?

[lapis wraith]

yeh

[muted hinge]

are you monetized?

[lapis wraith]

not on youtube or anything

i sell cards

[muted hinge]

hm?

[lapis wraith]

as in how much?

[muted hinge]

wdym by selling cards

[lapis wraith]

like tradiing cards

[muted hinge]

ahh

[lapis wraith]

custom

[muted hinge]

I see

but you said your work is parody, wdym by that?

[lapis wraith]

my art work and i also parody another artist

theres an artist named piggybanx and i crearted piggyspanx. and he sent me a ceasse and deisist. so I went all in on it parody aspect

[muted hinge]

ohhh cool

[solid ice]

A parody is a humorous or mocking imitation of a work, style, or person, often exaggerating their characteristics for comic effect.

[lapis wraith]

its super petty

[solid ice]

Weird Al is a parody god

[lapis wraith]

yehs

when i. do a card , i get a video, i use suno and other software and create s a song unique to the card

i go all in

ii undestand why people sell things with politicians on them now

[muted hinge]

can you show an example?

[lapis wraith]

yeh let me see

theres my walter white pokemon card

mefflyn base losely on gastly pokemon cause the "meff" he makes

the piggyspanx video i posted ealrie was actualy for a card for batman but with politicians

and the card for the yoshi vido

im debating on the the wallter white inside flow as anime

[muted hinge]

ohh thats sick

[upbeat shuttle]

Gosh. I really wish I could make videos longer than 8 seconds. I would pay more for it in fact. "Extend scene" and "start from frame" have pretty bad artifacts so aren't useful for me.

[mellow grove]

Hey guys, me and a friend are doing google veo 3 videos for content. Anyone who is also doing the same thing can add me as we are trying to make a group chat between veo 3 content creators for tips etc. Anyone whos interested, add me!

[white forge]

Sure, where is it? count me in!

[lapis wraith]

im in

[red dirge]

im in!

[restive horizon]

IWhat exactly does the Jump to function do? Does it simply create an independent new clip, or is there any interaction with the previous scene?

[rose valve]

I’m not sure if this is the right place to ask this but I need help. I signed up for the 30 day free trial for flow. I didn’t realize they did a student free extended trial (assuming that still exist). Does anyone know how I can get that extended? I keep trying to do it the way YouTube told me 😂 but it’s saying I’m not eligible. I live in the USA. Any suggestions?

[north cedar]

This is feasible

[clear sphinx]

Is it possible to generate a 3-second video with Veo 3 and Flow? I attempted to make the shot short, but the resulting video was 8 seconds long.

[charred sail]

No. They are all 8 seconds.

[clear sphinx]

Thank you!

[flint summit]

Sometimes it's an almost seamless transition. But sometimes it's completely unrelated.

[frigid pier]

Why do I have Veo watermark in the video?

I am pro subscriber

[clear sphinx]

We need ultra to remove watermark

[frigid pier]

oh nooooo

I will just have the watermark then...

[clear sphinx]

Me to lol

[charred sail]

I don't know why i always have to pick some new difficult thing to make.. I'm doing a smokey and the bandit style car chase 😛

veo 3 knows nothing of driving LOL

having a heck of a time getting all the characters in the right car doing anything.. i tried making them turn and the car bent LOL

[lapis wraith]

Have you tried a fight scene yet?

It’s like letting to blind folded people against each other

[sharp quest]

yep

[rich hedge]

Does it normally takes some time to have Flow active for my AI Pro sub? From Germany and just ordered Google One

[charred sail]

yeah, it's pretty rough

[lapis wraith]

Here’s one I’m finding hard. A tweaked out sloth drinking coffee at the same slow speed as usual for a sloth

[clear sphinx]

How do you manage the audio's starting point? I'd like to place audio at the beginning of the video, but it frequently appears in a later section in Veo 3.

[mellow grove]

Hey guys, me and a friend are doing google veo 3 videos for content. Anyone who is also doing the same thing can add me as we are trying to make a group chat between veo 3 content creators for tips etc. Anyone whos interested, add me!

[white forge]

That's good 🤣

Spiderman, with deadpool persona.

[twin mortar]

Has audio been fixed yet? I don’t want to waste any more credits until it’s been sorted.

[lapis wraith]

what wa wrrong with the audio

these json prompts are amazing

its giving me the same exact video with slight changes

thats pretty good stuff to know

{
"video_prompt": "Two realist woman walking dead zombies stand in a post-apocalyptic wasteland, casually talking next to a burned-out food truck. One zombie has glowing, clearer skin and is wearing a filthy hoodie that says 'I ❤️ Brains'. The other looks crusty and flaky. They are mid-conversation about switching to eating vegans instead of general humans. The tone is dry and sarcastic, like a deadpan sitcom.",
"length_seconds": 8,
"visual_style": "dark comedic realistic walking dead television show subtle glow effects, muted earthy colors with greenish skin tones",
"camera_direction": "starts with both zombies side-by-side mid shot, slight zoom on the glowing one as she makes the reveal, ends with the other zombie skeptically sniffing the air",
"audio": {
"voiceover": [
{≠
"character": "Zombie 1 (glowing)",
"text": "Yeah, I stopped eating random humans. Strictly vegans now. My skin? Glowing.",
"start_time": 0.5,
"end_time": 5.5,
"emotion": "chill, confident"
},
{
"character": "Zombie 2 (crusty)",
"text": "...You do look less... moldy.",
"start_time": 5.5,
"end_time": 7.5,
"emotion": "confused, curious"
}
],
"sound_effects": [
{
"name": "flies_buzzing",
"start_time": 0.0,
"end_time": 8.0,
"description": "ambient buzzing around the zombies"
},
{
"name": "stomach_growl",
"start_time": 6.0,
"description": "hungry undead rumble"
}
]
},
"text_overlay": [
{
"text": "ZOMBIE TIP #32: EAT CLEAN.",
"start_time": 1.0,
"end_time": 4.0,
"position": "top",
"style": "green bold text with grainy glitch effect"
},
{
"text": "#PlantBasedBrains",
"start_time": 7.5,
"end_time": 8.0,
"position": "bottom-right",
"style": "white, all caps, bold"
}
]
}
≠

[finite moat]

Es kann bisschen dauern, wie hast du das genau gemacht mit Flow ?

[charred sail]

veo thinks this is a realistic cinematic right now..

I'm going to try the same thing on quality and see if they're still cartoons

lol it was even worse

That stuff is so bad I flagged it as offensive

jk.. 😛

[tawny kettle]

#flow-general・🌊 Close-up of the credit card being tapped on a sleek payment terminal. Upon tapping, golden glowing particles rise into the air, transforming into a stylized Etihad Airways airplane taking off into the sky. Ends with a clean, cinematic fade-out. No text.

[glad arrow]

I am really fed up with my prompt flagged as breaking the rules while all I want is just a scene where a kid waking up her mother, while the other people could post their AI generated of woman in bikini.

[charred sail]

don't say kid

[glad arrow]

fr? Thats the problem here?
Aight, I am gonna test it out, ty mate

[winged plinth]

Hi, you’ll need to use the website to generate your video
https://labs.google/flow/about

[charred sail]

yeah they're overly safe with kid stuff.

[lime star]

Do you have a prompt generator? I would like to create videos like this

[quiet stream]

In a post-apocalyptic wasteland under muted, earthy colors, two realistic walking dead style female zombies have a deadpan conversation next to a burned-out food truck. One zombie has glowing greenish skin and wears a filthy t-shirt that says 'I' followed by a red heart symbol and the word 'Zombies'. The other female zombie looks crusty and flaky. The camera does a slight zoom-in on the glowing zombie. The ambient sound is of flies buzzing. The glowing zombie says in a confident, chill voice: Yeah.. I stopped eating random humans. Strictly vegans now. My skin gloowing. The crusty zombie, sniffing the air, says in a confused voice: You do look less moldy.. No subtitles. No subtitles.

Tried with my custom prompt (fcked up the dialogue, forgot to add correct punctuation) but yeah your json looks better

[weak stream]

hey, how can i download the video i have just made in Flow? can't understand it in the interface

[glad arrow]

fr? just hover the video and see on the edge top right...

hover : move ur mouse into the video

[weak stream]

oh, thanks. i see it now. i was in Screenbuilder page and thought that i can download the compilation of videos at once.

i have another. once we get all these short videos. do we need to compile them by ourselves in the separate software like CapCut or can it be done in Flow?

[glad arrow]

not really sure about that but why dont u take a try the "add to scene" feature on the top left video... I am also new here, but seems like that feature could combine multiple video and maybe u can download all in one into combined 1 video.

[quiet stream]

i think you are both talking about the same thing (scenebuilder = add to scene), and yeah i dont think there is a way to download each combined scene together yet (which is weird)

[next karma]

Question: If I have google pro and I buy 20000 credits. Am I still limited to 3 generations per day? or should I upgrade to ultra to unlock the limit?

[sinful sigil]

Hey there, I would like to test the new lipsync feature. Where can I try it?

[frigid pier]

many times the voice sounds like robotic like monotone

how do I achieve realistic voice tone?

[crude scroll]

Had to do some composting in after effects but this was fun making.

Do YOU Like Sea Food? #animation #cartoon #memes #horror #funny
https://youtube.com/shorts/HVc6qDPIlF8?feature=share

[lapis wraith]

Where’s the shirts that was on the previous three

[lapis wraith]

Just use gemeni and tell it you want. Prompt in json format

That’s just to keep things consistent with json

[lapis wraith]

I just learned bout the json stuff a bit ago. It seems to allow minor Changed while keeping a lot of the same ideas

[spring mango]

Can't top up Google Flow credits on India's Google AI Pro subscription plan

[fleet quarry]

Have anyone here tried different models? Opinions?

[fair gull]

Is anyone else running into Failed Generation, try again later errors?

[glad arrow]

3 generation per day? fr? Just use flow bro

Dont use it from Gemini

[next karma]

I have ultra now, so I am using unlimited. I was just thinking which one would be better value for me, and I read somewhere about the 3 generations limit. So is it better to stick to ultra, or go pro and add credits?

[twin mortar]

Is this product going backwards in it’s development? Loads of people now reporting broken or no audio and continual captions.

[quiet stream]

I dont know if you can even buy credits with pro tbh, or if u can generate more than 3 per day

[marble hamlet]

Alot of things are not being done properly, not because they can't, but because they won't. Understand the difference. If it was left up to people like you, we'd never make any progress.
You think you're superior to others here. You just accepts things, even when those things are wrong.
I work 7 days a week as a careworker. The last thing I need is to come on here and have some jerk talking down to me.

[woeful fossil]

You can buy credits for use in flow as a Pro user, you get 1k per month with your sub.

You get 3 Veo generations / day with Gemini and you get 10 free in Google Vids.

[lapis wraith]

To bad nobody has like 10 google accounts

Cause that’s crazy

Wink

[quiet stream]

But does the free account strat actually use the same veo3? cos i swear, before buying ultra my videos looked garbage even though i used veo3 (i didnt know how to prompt back then to be fair)

this is an hypotethical quesiton in minecraft ofc

[lapis wraith]

Yeh

It’d veo3

I’ll show you the diff of 3 vs 2

Same prompt

[quiet stream]

Did someone you know in minecraft used this strat as an hypothethical scenario in a video game? and was it actually the same quality?

[lapis wraith]

What do you mean

[woeful fossil]

Thats definitely a way to do it, there are some drawbacks. Flow is the only one that has 1080 upscaling though, and Vids has a 1k character limit.

Also the way they handle your prompt is different, so if you use very rudimentary prompts you'll get varied results in each

[quiet stream]

damn, yeah that was what i was wondering, must be a drawback somewhere

[lapis wraith]

Vids?

1k character limit on videos

Or is there another generator or model you guys are talking

[woeful fossil]

Google Vids gives you 10 free Veo 3 gens per day with an AI sub, but the prompt window is only 1k characters.

[lapis wraith]

What the hell is vids

I didn’t know about this

[woeful fossil]

https://workspace.google.com/products/vids/

[lapis wraith]

Oh wow

Didn’t even n know about that

[woeful fossil]

Yeah its kinda secret sauce. So youre welcome I guess 😄

Was going to make a yt Vid about it but couldn't be bothered.

[lapis wraith]

Just make a automati podcast with notebook lm

[mellow axle]

something happened to daily free generations on Gemini? anyone knows?

[quiet stream]

it appears that something went wrong

but yeah on an other note i have no audio today

[jagged jackal]

So glad to see Veo 3 finally allowing i2v and credit packs without the super expensive plan. I re-subscribed.

[jade crest]

🤣

[sick fossil]

Same issue. Not sure what's going on.

[jade crest]

Are there any watermark-free video-generation tools that can rival Veo 3?🤣

[stark mica]

I got a lot of errors that i cant use photorealistic images in frames to video, even when the images are generated by Imagen4. Is there some way to do this? Otherwise i feel stuk to animals and objects

[gaunt magnet]

for god sake just fix the extended feature for veo3

[old umbra]

i have a question. The 3 daily videos limit from veo 3 in gemini, its the same for ultra plan? or if i pay ultra plan i can make more than 3 videos daily? i know flow doesnt limit the videos, but in flow it is not working the frames to video when i try to prompt a spanish dialogue, while in gemini it does work

[lapis wraith]

I meant “brainstorming ideas”. Just saw this

They limit your in the regular gemeni print thing either way. I almost lost it when I paid the ultra and it said comeback tomorrow

[glad arrow]

what does the 'Flag Output' uses for?

[sly sonnet]

Hey All, new to the server just got started trying out Flow. Looks great. MY org has a Pro subscription but I can't seem to generate vids (no credits error, I assume we don't get those with a Pro subscription?) and I can't seem to use the "generate from ingredients feature' either. I'm logged in on my org email and have confirmed that my email is given the correct permissions. Do we just not get credits for Flow with Pro? Sorry if I'm asking a frequent Q I'm fine with a redirect if this is common and has a solution somewhere.

[vague geyser]

image to video works for veo3 now? I have ultra plan

[sly sonnet]

Maybe it's just us being out of credits then, because I do have access to the features and they seem to work (excludingthe no credits error lol). I guess we just have to upgrade to Enterprise to get the generate from ingredients feature? That seems to be the best way around video gen's biggest hurdle of continuity of characters etc

[sharp quest]

I want to create my own prompt template, but I’m not sure whether I should use ChatGPT or Gemini

I want to create a series of sequential scene prompts consisting of 5 or 10 scenes. I don't have any issues with character consistency, voice consistency, or lip-sync. What I’m mainly focused on is the setting and the objects within the environment.

So, do I need to describe the lighting, setting, and objects myself in every scene? Or can ChatGPT or Gemini handle the environment for each scene in a consistent, sequential way for me?

[sly sonnet]

I'd say try both and see which gives you a better prompt in the end tbh

You'll probably want to generate still frames of each scene and specify the thingsl ike lighting etc when generating those

you can add effects in your prompt for the video or in a video editor

Shotcut is free but limited to what it can do, I've seen some impressive post edits from others with adobe after effects

Digging in deeper I think the issue is its not letting me access our org's subscription as we should have 1k credits unused. Do I need to link it on my end or is that something my boss should do?

[fathom shadow]

Anyone having issues getting into projects right now?

[sharp quest]

I only subscribed to the Google One Pro one-month trial to use Google Veo 3, and I entered my card details. Right now, I can use the 1000 credits. But it seems like Veo 3 isn’t available in some countries at the moment… I’m not entirely sure though.

By the way, if you generate videos with Gemini Veo 3, you get 5 attempts per day… The 1000 credits are used in Veo 3 Flow. The real work happens in Flow.

[main saddle]

fwding!

[sharp quest]

For example, in this video, the scenes look like they were all shot in the same location—only the camera angles are different. This video was shared before the most recent update. I'm trying to figure out how he created sequential prompts to make it. I’ve tried a few times myself but failed. He also has a few courses, and his most popular one is this: https://bigfootboyz.gumroad.com/l/aivlogs. I’m not sure if it’s worth buying.

https://www.youtube.com/watch?v=dBc2nOvfqPc

[fathom shadow]

Anyone? It's just stuck at loading for me. Now I logged out and it won't even let me log back in??

[cobalt junco]

I'll try I haven't been in there in a while

[cobalt junco]

just went into all my projects on my phone with np

[sly sonnet]

Probably used midjourney or another paid sub where you can do eidts/revisions on generated images

Not sure about Goggle's features gonna learn about those this weekend lol

*Google

[charred sail]

Right now there's like 4 people in Austin pulling their hair out after 3 months of 60 hour weeks going, "We're f'ing tryinig, man!"

[lapis wraith]

if you dont care for audio i would

but if you want audio thi seems liek the way to go

lol thats a damn good gif

[ivory talon]

Figured it out a work around - alter the face slightly in photoshop. Slightly. It's working for me. "To alter a face shape in Photoshop, the Liquify filter's Face-Aware Liquify feature is a powerful tool. This feature allows you to adjust various facial features, such as eye size, nose shape, and jawline, using sliders or by directly manipulating anchor points on the face. "

[sleek oxide]

Lol

[spice phoenix]

https://youtu.be/t7fthKOjaI4?si=FR523zy7It1uhBDi

Trying to make a series just upgraded to ultra but I don't find the experience better. I'm hay trouble with continuity with prompts. Any suggestions would be greatly appreciated!

[north cedar]

How to implement the prompt words for the surround shot?

[elfin ore]

Hi. Why is it that flow says it is available in other countries but when I switch off my vpn it doesn't work

[slim hedge]

Here's a tip for European users of flow (Like myself) who get the 'photorealistic image' garbage rule. Generate a photorealistic anything in Midjourney, make the image 'body only' do not generate an image that shows the face, cut the face, have the face obscured by the top of the image, however you want to do it. So now you have your headless face obscured image, upload that as your start image in flow and in your prompt get flow to pan up to the face. Veo-3 will pan up to the face and create whatever face you wanted in the first place in the video gen process. Flows barriers to European users doesn't key from photorealistic images at all, it keys from a face only. If it sees a face in the start image it will refuse to use it, so work smart and get around the problem👌🏻

[orchid geyser]

I have some suggestions for drastically improving Flow: We should already work on the functionality of the characters first and first, i.e. Veo 3 is able to retain the characters we want to see appear throughout our film and to be able to use several of them just by mentioning the name in different scenes

[lusty relic]

wait.why only europeans?

do europeans have a bad thing with faces? lol

[elfin ore]

Why say it works in all those other countries when it doesn't? Clear communication is crucial.

Atleast just say it only works in the us.

[orchid geyser]

Google also needs to seriously review the credits because even taking the ultra version we only have 12500 credits and using veo 3 quality if all our prompts work properly (which is never the case) we can only produce about 10 minutes of videos and that's a long way off for the ambition to want to make a movie or a series with episodes

[elfin ore]

<@&1173646484311384114>

[quiet stream]

Different copyright/privacy laws etc
In the total opposite, in china you can create copyrighted material like star wars they dont really care

[lusty relic]

👍 thx for explaining

[jaunty stratus]

guys

is veo 3 still monetizied

[quiet stream]

I dont know how many channels using veo3 are monetized or not but there is a fake news going around that youtube is demonetizing all AI which is factually wrong

It will only target low effort automated channels (example: slideshow with boring ai voiceover, or youtube shorts where you just cut planets and all the videos are the same etc)

[crisp basalt]

Did you report it in feedback channel

[sharp quest]

I think I finally created the Bigfoot model I had in mind.

[spice phoenix]

I am not in China and can make star wars prompts

[prisma flint]

Hey guys! I literally started using Veo3 as of today. Absolutely love it! I am trying to generate video content that also includes a chart. Now I have Google Pro subscription. Veo3 Fast at the moment does not allow you to include any attachments. It is pure text to video. Has anyone in the group tried to build out AI video content that has data driven analytics? I have troves of data and charts from an opinion research firm and I want to put that to good use. Any help would be appreciated folks!

[lapis wraith]

there ae gonna be so many good gifs that this things makes

[livid canopy]

anyone else having the issue where veo 3 quality stops generating at like 40% and then spits out some low quality video?

[thorny dagger]

please do video veo 3. A sunny day in front of a traditional village office building labeled "Kantor Desa Sindangheula". A male village official (perangkat desa), wearing a batik shirt and peci, stands proudly while holding a flyer with both hands. The flyer is clearly visible and reads "SEMARAK HUT KE-46 DESA SINDANGHEULA". He speaks confidently in Indonesian and Sundanese, saying: "AYO, segera daftarkan dan ikuti lomba-lomba Semarak HUT Desa Sindangheula yang Ke 46!" The scene shifts to dozens of villagers—men, women, and children—eagerly rushing towards him. The official is slightly surprised and raises the flyer high above his head to protect it as people around him try to grab it. Lively background atmosphere with festive music, flags, and banners, showing a sense of joyful celebration in a Sundanese village.

[kindred needle]

https://www.tiktok.com/@nonsensedaydreaming/video/7520017757589605662

[rich hatch]

A salesman with violet polo shirt with TONIK word logo in left chest, with violet ID lace with black pants and black shoes.

approaching a woman with a problem with her broken washing machine.
and ask her to try our product
TONIK SHOP INSTALLMENT LOAN

#flow-general・🌊 A salesman with violet polo shirt with TONIK word logo in left chest, with violet ID lace with black pants and black shoes.

approaching a woman with a problem with her broken washing machine.
and ask her to try our product
TONIK SHOP INSTALLMENT LOAN

[unique forge]

Hi all, can we use this channel to get help on veo 3? i'm currently in the process of making a short series and have my first minute 30 of clips created. had tons of issues and stitched them together using flow. I'm having a hard time trying to get different scenes and explain the scenes and shots

[quiet stream]

Not sure what the question but the best would be to put the clips together in a video editor program

[midnight cobalt]

I am in Europe and trying out Veo3. I'm working with photorealistic fictional characters (images of my group's D&D characters generated in other AI image generators), but obviously Flow cannot tell that. So I hit the "no photorealistic images" block.

For now I can explore other uses for Flow and Veo, but I would like to know if there's any indication on whether this policy is temporary precaution or something that is long term.

Is there anywhere I can go to discover the reasoning behind this (and other) policies? There's a big difference between it being a precaution-based interpretation of EU law that may get reviewed soon, or something decided and locked in based on a specific AI content bill that may be fixed for years.

I would like to understand better what my options are for thinking longer term about my projects. I could for instance start creating a bunch of illustration styled images of the same characters, but that is a significant time and effort investment compared to e.g. waiting for a few months for a possible policy change.

[quiet stream]

Any suggestion to upscale images (but not just the size, one that can also "enchance" it, make it more detailled etc)?

[slim hedge]

No, we have a set of privacy regulations called GDPR, which are a good thing but every once in while they block a society from making any progress at all

[slim hedge]

Use the method I outlined about obscuring a face in the start image and describing the face in the prompt to let veo gen the face for you, or use a good VPN to mask your location. I'm in the same category as I live in the EU too, although it isn't the image thats the issue, just the face, if you use a photorealistic start image with the face hidden it will let you use it

[midnight cobalt]

Won't work for me. I want specific already-generated characters.

Understanding where the policy is coming from, and anything resembling a roadmap is what I am looking for. Once I know that, I can decide whether any specific work-around is my best bet.

My alternatives would include generating voice separately and using lipsync tools in a two-pass approach in a different generator for example. Or dropping photo-real in favour of more animation style in Veo3.

It all depends on understanding what my options will be over the coming weeks.

[jaunty stratus]

@hybrid helm <@&1366831844502274180> VEO 3 Videos are not monitized on YouTube anymore?

Do you know this

[woeful fossil]

Not really a question for Mods.

[unreal viper]

please don't ping mods if there is no moderation issues. thanks

[jaunty stratus]

How to reach the Google Staff i cant mention them

[jaunty stratus]

@main saddle @fierce valve @ripe path

[woeful fossil]

This would be a question for Youtube support, theres different branches of the company.

[jaunty stratus]

Okay

[spark delta]

Next time please ping <@&1173646484311384114> !

[snow iron]

Hey! I just signed up for a pro account to play with flow and VEO, I cannot find where/how upload a reference Image, can somebody point me in the right direction?

[charred sail]

frame to videoo.. upload

[woeful fossil]

You may just want to create very detailed character descriptions and use Text to Video to get what you want. You could use your images as references to remake them in whisk and try that?

I used whisk to recently generate a D&D character and then used Frame to Video to create this which worked well. But I'm also in the US. so I can't speculate on regional differences.

[snow iron]

Thanks @charred sail , I'd tried that and it kind of works, but it always starts with the image in the video, I want to use an image of my character

[charred sail]

what video? there's no video until you prompt it

[snow iron]

As in - I upload for frame to video, do the prompt and generate, first frames of the video ( .5 seconds ish) are my images

[charred sail]

yep

frames help with continuity but they have their limits

I don't know what you're trying to do, but I can tell you it's going to be 100x more difficult than you think..

LOL

you've decided to create art with a 2 year old savant 😛

[charred sail]

I've been trying to do a dukes of hazard style car chase hahaha.. why...i don't know.

Veo thinks cars are bendy...

[unique forge]

the start of something

[charred sail]

It does okay on this one til the end.. They werre supposed to turn left..

i can cut that and make it work.. but it drives me crazy!

just turn f'ing left

[woeful fossil]

😄 cars are definitely interesting. I have some great bloopers from making my Mustang short film.

[charred sail]

yep

[woeful fossil]

That zoom into the cab and back is pretty sweet though. What was your prompting direction for that?

[charred sail]

Let me look. I think i did originally, but then that was a frame to video of the original

So the way i got that transition was started with a frame from that angle and then the prompt was all inside the cab

[midnight cobalt]

You may just want to create very detailed character descriptions and use Text to Video to get what you want. You could use your images as references to remake them in whisk and try that?

These might be OK work-arounds, although I have some doubts I might give them a try and see whether it could be made to work. My suspicion is that "detailed character descriptions" would suffice for non-photo-real and very different characters (e.g. all the bigfoot videos, which are actually not consistent but you tend not to notice because of how identifiable the character is through being non-human). But if I wanted those, I can just do them anyway with an image. I just tried with a more cartoony version of one of the D&D characters, it was recognisably them and worked fine.

Back to first question, although I appreciate attempts to help with alternatives, I am not really looking for work-arounds at this stage of investigating Flow.

I'd like to know if anyone has the answer to my initial question - is the reasoning for the UK/EU policy on photoreal videos explained anywhere beyond it being an enforced policy in my region?

To decide if it's even worth playing with work-arounds or compromises, I'd like to know first where the policy is coming from (specifically the no real-looking faces allowed in UK and Europe) and whether this is something temporary or a long-term issue that I need to account for.

My goal isn't to critique the policy (I can see where it might be coming from if UK/EU have applicable laws), but to understand how the service will work for me in the next year.

[charred sail]

I get it, but does it matter where it comes from? Yeah as an American I should be able to make parodies of Trump. It's legal. But google will never allow that, right? Doesn't matter why. Isn't going to change.

I'm sure in a few months there will be a chinese knock off that lets us LOL

[midnight cobalt]

I get it, but does it matter where it comes from?

Yes, because:

Isn't going to change.

It's not at all clear that's the case. The error message even implies that it could change at any time: We do not allow . . . in your territory at this time.

So I am looking for clarification of those ambiguous words

[charred sail]

I mean.. they're ambiguous for a reason

they're playing it safe

[midnight cobalt]

Kling 2.1 is close.

The audio is not a patch on Veo 3, but rest of it is kind of on par, some things better, some worse.

[charred sail]

they piss off the wrong politician and it's over

[charred sail]

I say, enjoy this stuff while you can. This technology will eventually be off limits to unlicensed professionals. I give it one more election cycle. It's scary what a bad person can do with it. I'm thinking about starting a cult myself 😛

jk about the cult.. who needs the pressure

Did you notice it lets you do star wars stuff?

How is that possible?

you'd think disney would be all over them

[midnight cobalt]

That would be a shame.

I hope providers can find a way for well-behaved creators to have access without requiring them to be large corporations. Whilst of course still having ways to block bad actors. It is a powerful technology though, I agree, so the task is hard.

I just want to add some extras to a table-top RPG, and use my own imagination as opposed to buying pre-made things by those corporations.

[charred sail]

ah the plight of the regular joe..

[midnight cobalt]

Actually that's not something I'm interested in. Fan art, parodies, memes etc. It can be a big deal for other people's enjoyment of this though

[charred sail]

i dunno.. an imperial cruiser showing up in a d&d campaign sounds kinda fun

used to have a dm that would do that stuff

what's the save for a force choke?

[midnight cobalt]

WoTC have been doing a lot of that cross-over stuff in Magic, and there are some commenters thinking this is where D&D is headed too, as a kind of IP licensing platform to sell us just that sort of thing.

[charred sail]

isn't that nuts.. all you need is 2 or 3 core manuels.. use your own imagination

gotta go afk

[midnight cobalt]

There's room for that licensed IP. Pay money, get the thing for low effort.

But hope there's also room for small-scale original creators homebrewing all sorts of content

[deep jewel]

hey all how to go from 16:9 ratio to 9:16, what tool to use to make it inexpensive, simple and fast

[oblique gust]

Have you guys able to find a way to stop subtitle or caption on the video?

[verbal portal]

hello, i love creative with flow , I m ondering if the bug with animation and speech can be repaired or its a policy of the company> I need to create dialgues fo rmy animated movie here obviously many characters are little animals like bunnies, mice etc. and veo 3 wouldnt create voices as it says that "voice generation with minors is not supported" but they are animated characters for a kids movie

or is there any way to overcome this restriction with a prompt? or to wirte to support directly about this issue.?

[charred sail]

I think kdenlive might do it for free, capcut does it if you go premium

[deep jewel]

i do it with https://dream-machine.lumalabs.ai actually

[charred sail]

Try this. Technical (Negative Prompt): music, subtitles, captions, watermark, text, words on screen, logo, branding

[deep jewel]

do u have an exemple with cap cut pro, lumalabs do it great but expensive for a lot of video

[charred sail]

don't describe them as young or little

or kids

little might be ok

try youthful?

LOL i saw someone do a bigfoot vid where the baby was just a miniature spanish guy with a mustache

[versed shale]

How do people have the jamaican voice?

[quiet stream]

I wasted around 1k tokens just to make the character turn left, it always turned right no matter what I tried, veo3 does absolutely not understand left or right as we understand, it seems to have a different logic

[oblique gust]

it worked.

[charred sail]

yeah, gotta play editing tricks.. it understands left or right of the screen

[oblique gust]

try using json style. It happened to me when i used regular prompt style. But i switched to Json and it worked.

[oblique gust]

it seems like json style worked for me almost every single scenes.

[charred sail]

Will chatgpt do that? LOL

[quiet stream]

Maybe my Json are not structured correctly but whenever I tried it brought some other minor issues compared to my custom AI that writes it for me in plain english

[oblique gust]

yeah i've been doing it.

[verbal portal]

Can someone please help me animate with voices the scene? "cute pixar 3 animated scene, the buss starts to ride slowly towards and Two happy bunnies look at each other, one says " let's go! " The other giggles : "finally, going home!" turn and walk joyfully toward the open door with green seats visible throught he opening of a cheerful school bus, one of them eating a red apple at the same time as he walks in cheerful step , the other buuny smiles and giigles kindly as he leads the way joyfully. The bus blinks its big expressive eyes warmly, smiling and watching the bunnies with delight as they approach. The bunnies giggle and bounce with playful energy, climbing the stairs eagerly one after another. . Once they enter, they disappear inside the bus, leaving the door gently closing behind them. and the bus starts to ride forward The atmosphere is bright, welcoming, and full of childhood wonder, bathed in soft sunlight with colorful houses, flowers, and greenery surrounding the scene. The camera softly glides alongside the bunnies, tracking their movement toward the bus door and capturing the bus’s joyful gaze. Cute funny Pixar 3D professional scene

[charred sail]

I'm going to try it thanks

[oblique gust]

with your negative prompt. It worked well for me now too.

[charred sail]

I think that's something prompt alchemist does.. https://promptalchemist.com/app

[oblique gust]

I think it is 90% worked though. I just tested 5 different scenes with that negative prompt. It worked in 3 scenes and 2 other scenes they have sub on. So i think it is Veo 3 side.

[charred sail]

Hard to say.. could be too much dialog

[oblique gust]

also in json file, if you have a character talking using "dialogue": change that to "voice":
I think it works better.

[spice phoenix]

Let me know what you think - https://youtu.be/Cz7GS--l0Ck

[charred sail]

I do like that format

let's see what happens 😛

[oblique gust]

try it. I just tried 3 scenes just now and no sub so far.

[charred sail]

nope still didn't turn

"action": "Gripping steering wheel, driving on right side of road, turns left onto dirt road to evade police"

probably need to work on that prompt in that format

[oblique gust]

This is an example of one the command i asked camera to do.

"camera_action": "Bigfoot briefly tilts the selfie camera to show the receptionist behind the desk"

It did exactly what i asked it to do in Veo 3

[snow iron]

kind of got it working! the inkie robot that walks in is the character i uploaded. I just trimmed the video start when generated 🙂 for the second scene i used a still of the first scene with all of the characters in it as the frame upload, and removed it afterwards. BOSH

[quiet stream]

Do you notice a difference when using frame/image to video vs just text the video?
Because i notice this weird "black filter" in your video same as mine when I use frame to video

[viral thicket]

is there a way i can share my account with my team?

[snow iron]

Hey @quiet stream tbh I've only just started and i've only tried frame to video

[snow iron]

one thing i will say, i think i get what you're saying, trying it on veo 2 ws a lot more crisp, e.g this

[viral thicket]

is there a way i can make a ticket for support

[quiet stream]

yeah I noticed only yesterday, frame to video sometimes has weird artefacts (veo3 fast)
eg: first is text, second is frame

[charred sail]

:p

[cursive root]

Hi, why does i2v only work with English audio? I can't make people speak in other languages. Will it be implemented later?

[charred sail]

voice control is more about context

have it imagine the character in a certain setting where they'd speak with whichever accent you want them to

[cursive root]

It says "only English supported" whereas t2v works with a lot of different languages

[charred sail]

ah well yeah i was speacking of accents

It's always something.. now I can't get a ramp to go up.. lol it goes down for some reason

[cursive root]

Yep accents are fine... Weird thing is I managed to do it once, i2v with someone speaking french, but that happened once only

[charred sail]

says not supported ... doesn't say won't randomly happen LOL

That's the 2 year old part of the savant i was speaking of earlier

omg i just about got it but it messed up the freakin flags LOL

if it didn't screw up the canadien flag

there's gonna be a stop mid jump "This is about the time, the squatch boys knew they were in trouble!"

a la dukes of hazzard 😛

[eternal furnace]

does anybody know how to access this tool

[charred sail]

check here https://labs.google/experiments

[eternal furnace]

I tried flow but i dont see where you actually do this

[charred sail]

The json format hasn't got the flags right once

[charred sail]

Yeah flow can't

probably one of those other apps

[quiet stream]

lmao, crazy watching veo3 build the videos, it probably decided to put the ramp like that for whatever reason, but did not change the action later

[charred sail]

i added a ramp with an upward trajectory

but it refuses to take the green out of the canadien flag

it's launching right in every attempt now though

[dry frost]

Progress!

[charred sail]

smh

just a bit outside

[snow iron]

it's comboing it with the irish flag!!

[charred sail]

nah it's a mexican flag and a canadian flag.. the sasquatch are from those countries

they were protesting ice

now they run

I'm afriad it wins this one.. I'm not wasting anymore quarters on the canadien flag 😛

I got the jump footage i needed

I even tried adding red and white (no green) lol

[tardy ridge]

It’s a conspiracy 🤣I bought more credits VEO 3 worsks again

[muted fulcrum]

is this an "accident" to you all?

[muted fulcrum]

how come?

[finite moat]

Anyone else have this problem? Can someone help?

[woeful fossil]

What was your prompt?

[finite moat]

I don’t think it’s because of the prompt

8-second, realistic vertical video (9:16). The camera starts with a static medium shot and, during the call-to-action—when the alien laughs and speaks directly to the camera—quickly pushes in on the alien. In front of the iconic Cologne Cathedral on a lively afternoon, a 16-year-old athletic boy with short brown hair, black hoodie, ripped jeans, dark backpack, and AirPods stands next to a slim, 1.7m tall gray alien with large black eyes, no nose, small mouth, wearing a pink beanie, yellow backpack, silver necklace with Earth pendant, and modern streetwear. The alien holds a distinctive, colorful UFO-shaped microphone (neon colors, glowing rings) as a signature accessory. The boy speaks in a youthful, energetic voice with a clear Cologne accent. The alien, curious and friendly, speaks in a youthful, slightly high-pitched, surprisingly good Berlin accent. Five meters behind, a group of teenage boys in casual streetwear (friends, same age) are watching and filming with their phones. The friends remain quiet and watch until after the boy answers; then they start laughing and pointing at him. All figures are in sharp focus. Dialogue in the foreground (no subtitles): Alien: „Ey Kleiner, Messi oder Ronaldo?“ Boy: „Digga, ganz klar: Ronaldo. Messi ist nur für TikTok-Edits gut, mehr nicht.“ After the boy’s answer, the friends laugh and point at him. Alien (laughs, quick push-in, turns to camera): „craazy. Folgt mir für mehr Alien-Interviews!“ Sound: Clear dialogue, background laughter from the friends only after the boy’s answer. Style: realistic. Camera: medium shot, static, quick push-in on the alien during the call-to-action. Duration: 8 seconds.

[finite moat]

👆🏽

[finite moat]

Haha nice one.
So Would it be enough if I just write “age 18 to 19” in the prompt?

[woeful fossil]

That should do it yeah, I usually go 20 just to be safe. 👍

Also maybe change "boy" to young man or something as well.

I typically give the characters names in my prompts so that it refers to them by name rather than the pronouns.

[finite moat]

Thank you Prompt Alchemist 🤙🏽

[thorny elbow]

Need some love ❤️ & likes on this fun anime I made in 4 hours 😃

https://x.com/ai_car_pool/status/1944177117974933666?s=46

AI Car Pool (@AI_Car_Pool)

👁️ I woke up this morning 🛏️
an had an idea 💡it was to make a story about yummy SHIN ramen 🍜 @NongshimAmerica and @7eleven 🏬

Where do you buy your ramen?🍜
(I made myself into an anime dude and created this in 4 hours) 💻 and
now I want 🍜 ramen!

Like 👍 & Follow 🚶

X · Jul 12, 2025, 11:28 PM

[charred sail]

added some music https://www.tiktok.com/@sid_jenkins_atx/video/7526369732870540558?is_from_webapp=1&sender_device=pc&web_id=7524987652798760503

I think it makes it better

don't sleep on musicfx if you have it

[bronze fiber]

Is pro offer more credits

[muted fulcrum]

no 9:16 output yet?

[buoyant dock]

Hey I know I’m not in here as of much. I’m in so many different Discord groups and channels but I wanted to share a few of my Lemon8 🍋 posts (if I can or if it’s allowed)

https://v.lemon8-app.com/s/OgRhYesfs

[buoyant dock]

Nice!!!!!

[lapis wraith]

@woeful fossil

[buoyant dock]

These are so good!

I definitely have used Flow

[ruby olive]

https://youtu.be/PkjbTSAHW8I

For those who wonder how I did Vader's voice, you can use voice changer from 11labs on Veo3 clips 👍

[woeful fossil]

As I said, it doesn't support audio for prompts with kids. That doesn't mean it will never produce it. There have been hundreds of people in this discord frustrated with audio generation failing because their prompts included verbiage that led the model think it was about children.

I tend to give advice that cautions people from trying things the model isn't suited for, and getting frustrated/wasting credits.

[lapis wraith]

i made her say that

or i prompted

but she might be a little person

[charred sail]

Absolutely brilliant. What did you use to create the music?

[ruby olive]

Suno

[frigid pier]

I am trying to make two characters speak to each other

but two characters speak the same words at the same time etc.

any good prompft format for this?

[woeful fossil]

https://promptalchemist.com/share/6RybUtlcaRbQqGe28Ccg

[frigid pier]

cool man

[frigid pier]

perfect

[frigid pier]

should I pay for that prompt service?

[woeful fossil]

Hahaha I mean, I'm a little biased.. I made it to be a valuable resource for people in here. If you like it I'm happy!

[frigid pier]

I mean that's a good idea.

[woeful fossil]

I can say we're over 150 total users now with 50 active daily. So it seems people are liking it ❤️ !

[frigid pier]

bro you are already rich :0

[woeful fossil]

Nah nah, definitely not that many paid users. I have a pretty robust free tier. I'd rather get people using it first.

[frigid pier]

cool. Do you plan to add more features there?

[woeful fossil]

Thats the idea, I just added Scene tagging today. Works just like the Characters Manager So now you can create a scene style and reuse it easily.

[frigid pier]

that's great man. I like the UI too.
will look forward to see how it goes 🙂

[woeful fossil]

I appreciate that, if you have any questions or anything feel free to dm me.

[frigid pier]

okay man thanks 🙂

[toxic totem]

hey guys .. why this happens when i upload same image?

[eternal furnace]

Because I can’t find it anywhere

[toxic totem]

@woeful fossil Hey G ..

[woeful fossil]

Hey, whats the image you're trying to use?

[toxic totem]

my own AI Avatar ..

[toxic totem]

Anyone have faced similar issues??

[frigid pier]

I am getting this error too

I think the AI misunderstands the face

this is annoying really

[toxic totem]

I think they have intent did this .. but yes ..

[livid lynx]

I am also getting this error. How to avoid it

[frigid pier]

I got a really good shot and fk i have to use this!

but this warning keeps coming! and i cannot make video!

[toxic totem]

just working on it .. no ideas figured yet

[frigid pier]

maybe this is something Ai going wrong at this moment

maybe google team touched something now

because all of us are having this problem now all of sudden

[toxic totem]

I guess they have kept it intentionally ..

[frigid pier]

I paid and I cannot use it properly wth

This is not okay

[foggy heron]

i'm starting to become addicted to this ai

[frigid pier]

what videos do u make?

[foggy heron]

i make video involving my oc

[frigid pier]

what is oc

[foggy heron]

my personal oc whos is based on me

[charred sail]

lil sasquatch freestyle

[frigid pier]

how did u make the voice tone?

my characters sound so boring

[charred sail]

I had nothing to do with it LOL

i said she was dancing through the forest like a gangster rapper

[frigid pier]

hmm I see

I think VEO3 characters often have monotone

their voice is very flat

not natural at all

[charred sail]

give it context

[frigid pier]

okay ser

[tender hornet]

Hi All, nice to meet you

Wanted your advise based on your experience, is it better to Prompt the Text to Video in Veo3 in Flow, with plain text or xml code, for better results

[muted hinge]

hey. so its been a month already and my credits shoud have refreshed. but they didn't. I need credits. where are they?

[frigid pier]

I am not getting audio from Veo 3 fast anymore.

Is anybody having the same issue?

gimme back my credit

[lunar coral]

how to keep the voice consistent in three scene

[frigid pier]

impossible

[spice moss]

How to generate a video within 8 seconds and play the described content quickly? Many of the described problems did not achieve the final effect.

[lunar coral]

share your prompt

[charred sail]

A realistic wide angle drone shot of a sasquatch dancing in a forest clearing, softly lit by dappled sunlight filtering through tall pine trees. The sasquatch stands upright, playful and free. The camera captures fine details in her face and fur, with slight head movement as she dances with rhythm through the forest like a gangster rapper shooting a rap video.

[dense gulch]

Audio is not Worliny in both model: fast and quality from yesterday

[crisp basalt]

Gemini in Chrome + Flow ❤️

[twin mortar]

Ha ha good luck with that.

[lunar coral]

tried this..share your feedback

[twin mortar]

What are the usernames for the Google bods who post in here?

[frigid pier]

Where is he standing..?

[lunar coral]

outside building

[frigid pier]

it would be more engaging if he was walking on the street.

[lunar coral]

now he is sitting in office chair

share feedback now

what about this fear

[twin mortar]

@main saddle @limber zodiac @hardy halo Seeng as audio is still broken on Veo (no accents, tinny sound, no audio) and we’re getting constant captions with our generations (when we actually manage to get generations), what exactly is the point of us paying for Veo either Pro or Ultra when our credits are constantly wasted on useless generations?

I’m not wanting a stock answer about audio being a beta or an experimental feature - audio and lip-sync are the things that made Veo stand out from the competition. And until a week ago they seemed to be working fine.

So what exactly is Veo bringing that we couldn’t get for a lot less by using, for example, Midjourney?

Also, would you yourselves be happy paying top dollar for a service, any service, that has as many bugs as Veo?

[sharp quest]

https://bigfootboyz.gumroad.com/l/aivlogs
Has anyone purchased this bigfoot course?

[sweet dew]

no but i seen it on tiktok follow bigfoot and many others lol the two old grandmas that are in the plane are hilarious

[charred sail]

I tried to get this lip synced, but I could not find a program that could pull it off.. if anyone knows of one let me know.. https://www.tiktok.com/@sid_jenkins_atx/video/7526541223356665143

[sacred dawn]

when is the image feature (real people) coming to the UK Ultra workspace users?

[frigid pier]

why is it so unnatural?

Google Veo characters acting skill is really meh

[clear sphinx]

Hey everyone, has anyone found a reliable prompt to force Veo 3 to generate fullscreen video without the cinematic black bars?

I'm getting inconsistent results – sometimes it's fullscreen, sometimes it's cropped. It's happening more and more lately. Any tips or workarounds would be awesome!

[frigid pier]

image to video

[clear sphinx]

Thank you for the tip! But can't we do that with just a prompt?

[warm parcel]

hey guys, i just got ai pro to make ai videos with sounds but i got this message and it does not make sense for me. am i the only one getting this message?

[frigid pier]

u cant

[clear sphinx]

Veo 3 sometimes makes an error. It fails to generate audio.

[warm parcel]

`thanks for respond and also i do not even have option to change it to other versions like veo2, veo3 fast etc

[clear sphinx]

tab the setting icon next to Veo 3- fast letter.

[warm parcel]

yes sirr thank you

[bronze fiber]

I did

[mellow pendant]

anyone having no audio generated issues with veo3 quality?

just hit 4 in a row

[woeful fossil]

Hey, I've tested both and found that plain text works just fine. It also makes it easier for most people to make small adjustments to prompts when they dont get results they expect.

[mellow pendant]

prompt replies is next level lmao

[polar atlas]

@woeful fossilcan you make them 9:16 i can't find the option?

I am trying to make memes or anything in that type, i just get super ultra realistic 16:9 videos

[woeful fossil]

Unfortunately its not available as an option yet with Veo 3. For now you'll have to crop it down to 9:16 or do some of the fill effects available in some editing tools.

[near gate]

veo 3 should have options for custom duration instead of just 8 seconds that just gets me upset

[frigid pier]

lmfaooooo now u talk using Veo characters xD hilarirous

[candid creek]

Are there any freelancers here that can help me generate a few simple video ads for whom I have the general ideas, but not the prompts and time to make them be usable?

[vapid frigate]

A professional industrial rope access team working on a modern building facade. Daylight. They install safety nets, clean stone walls, and repair rooftops. The team uses climbing gear and rope access techniques. Add captions in French: "Experts des travaux en hauteur", "Sécurité et efficacité". Cinematic style, 16:9 or 9:16 format.

[steep wren]

Any ideas how to stop the camera from zooming in on Veo 3 with first frame included?

[faint wyvern]

yes , me.. why ?

[faint wyvern]

@woeful fossil Hey bro, what’s up? Funny you’re not saying anything about them posting in two places. You only had eyes for me, huh? Got a nice little warning thanks to you. Cheers.

[woeful fossil]

Lmao, I'm not a mod, you posted in every channel at once. I didn't even ping mods on you. They should get warnings too but that aint up to me.

[jagged sinew]

trying to make 3 wizards brothers and death story with veo3 (harrypotter), but when i do extend it only going in veo2 with no sound anyone know how to keep in veo3

[lost plank]

Hi guys, I am Alex and I just joined. Started using flow today, as it just opened its doors to Europe (I am from Lisbon, Portugal). I have a question. It seems that (maybe because google is entering Europe carefully) almost every realistic character I try to use in image to video is rejected , with the message that, for now, I cant upload realistic people. Is there any info on the real reasons? I talked with American friends and all of them say that is not a problem for them . This is a bit disappointing for me. Also, if Europe is having a more heavy censorship, shouldn’t the prices be lower while this happens? It seems extremely unfair. I would really love to understand it. Loving flow and veo 3, but this is a problem for serious projects that need specific character consistency. If there is a mod that maybe could help me with this matter, would love to chat. Thanks guys, nice meeting you all.

[charred sail]

That's kind of funny if you're into history and philosophy. Had you posted this on july 4th.. post of the day for sure LOL

[radiant mantle]

Any free worth veo 3 bundle

[lost plank]

Sorry, but I didn’t got what you mean. I am not criticizing anything, just trying to understand if this is related to the more strict Europe rules. Because the feedback I have is that in US this is not a problem.

[charred sail]

Google isn't censoring you. Your government is.

II mean they are too LOL

there's a lot at play

but that one is definitely your government

They still won't let up upload pics of trump and make parody

so... it is google to some extent

[lost plank]

Yeah, i understand what you are saying. I wasnt trying to say that google is censoring anyone. And my english is not the best, so i hope i am not choosing the wrong words. Just wanted to understand if this is really a thing related to Europe strict rules. I believe it is. And , on my case, I don’t do any famous or real people videos. All I wanted was to have my fictional created characters in my AI movies. Thanks for you reply. 😉

[humble mirage]

Cant have your charact with curse words crazy

[pale terrace]

is this where i can ask for help?

I used this prompt:1. EXT. RICHMOND, VIRGINIA – DAY – 1951
Thomas Jr (a young black tennis player), wiry and focused, practices tennis against a battered wall. His father, Thomas SR., stands nearby, watchful and proud.
Thomas SR.
Keep your head up, son. Remember—discipline.
Thomas nods, returning ball after ball, sweat beading on his brow.

I got this clip

a few obvious issues:

tennis racket disappears, practice against a wall is backwards, racial disparity of father and son

used veo fast

[dry frost]

Add a description for the father. Practicing toward a battered wall might work better

[hazy shoal]

trying to make an AI Video in planet MARS
https://www.youtube.com/watch?v=Z7IveIalTlA

[winged girder]

https://youtu.be/UAAozGcWyOo

Hi everyone, today I made a Veo 3 vlog of my cat. What do you think about it? Any suggestions to improve it and make it even better?

[pale terrace]

used veo 3, changed prompt to read 'practice toward a wall' and got this:

[winged girder]

https://youtu.be/UAAozGcWyOo

Hi everyone, today I made a Veo 3 vlog of my cat. What do you think about it? Any suggestions to improve it and make it even better?

[pale terrace]

so after 300 credits I got this:

where did google find this kind of tennis play?

[tiny canyon]

Guys I have a problem about extending the clip.

Currently I'm on PRO free trial plan and when I make one Veo-3 clip, app says to me I can't extend a video, only using Veo-2 which doesn't have sound.

Anyone with solution because I need videos with a 45s-1m duration.

[pale terrace]

make 5 10 second clips and splice them

[tiny canyon]

@limber zodiac

[tiny canyon]

I don't understand?

[pale terrace]

i also cannot extend a 10s clip. i had to make multiple small clips and dissolve one into the nextw

but the output is so weird it's useless (

[tiny canyon]

I understand but the thing is, after extending to the second clip Flow doesn't allow me to use Veo-3 which Is one with soundtrack

[pale terrace]

unless your intent is to make weirdness

i use veo 3 to make 10s clips. but even then the output is unrealistic (see above)

[humble mirage]

Why is religious video getting denied from making as well

[tiny canyon]

Does anyone know the solution

@pale terrace

Do you know what I'm talking about?

[queen juniper]

same for me... used to working before...

[queen juniper]

anyone has a workaround for this?

[stable sky]

Try to be super specific "the man is 3 meters away the wall and facing the wall while is hitting the ball with his racket in the wall direction, the ball rebounds on the wall and floor and he hits it again, the camera is behind the man..." (something like that)

[kindred needle]

Do you guys have any ideas on how to generate vertical 9:16 videos, like a week ago i was able to get videos that were like sideways

[stable sky]

I don't think this is a bug, but, no matter how insistent we are on the prompt indicating that we want Portuguese language from Portugal or European Portuguese as some AI models call it, the results is almost always in Brazilian Portuguese. Is a very hard credit burner. I have to make a lot of gens to obtain a useful one.

[sharp quest]

I watched the video of Bigfoot and Yeti flying F16, it was very funny 😄

[sweet dew]

he great lol i am still learning all the stuff but i love it so far...

[quiet stream]

Huh, since when veo3 does 10sec clips? isnt it 8 still?

[sharp quest]

All I want to learn is to be able to make consecutive prompts as if any character were in the same place for several scenes but with different camera angles.

[kindred needle]

are you guys real

[finite moat]

Hello everyone,
how can I create videos in 9:16 format or edit them afterward to post on social media?

[sharp quest]

create 9:16 image and put into veo3...it works for me 👍

[sharp quest]

create 9:16 image and put into veo3...it works for me 👍

[jade crest]

Veo 3’s text-to-video capability is far stronger than its image-to-video capability.

[sweet dew]

@jade crest I wish when you extend it would stay on 3 and audio lol 😆

[alpine cedar]

Boss

[frank finch]

How do i buy credits on flow?

[past linden]

New to this how are people getting such crisp videos I tried my hand at it today to work on some basic animation for a project and some scenes came out great some others I couldn't get right. Just sucks bc you have to guess if it's gonna be good or not bc of the credits. Idk how people are getting the smooth yeti clips made

[woeful fossil]

The best thing you can do is use highly detailed prompts. The more you explain to the model the more it will understand.

You also need to be aware of some of Veo's limitations which kind of just comes with practice.

[past linden]

That's fair I just saw some of the other prompts makes sense now somewhat. Do you recommend not using my own art and just trying to create something from scratch in the text to video @woeful fossil I'm just trying to keep consistentcy to create frames I can put together.

[woeful fossil]

It depends on what you need. I have had a lot of success with consistency using just text-video.

[past linden]

So as long as I build it in the same project it should remain consistent then ?

[woeful fossil]

No. The prompting is much more important. Create detailed character blocks and change things as you need. But describe every aspect of your subjects and keep the details the same.

[past linden]

Ok that makes a lot more sense then I will give that a go then I really appreciate the input.

[charred sail]

This site helps.. https://promptalchemist.com/app

[charred sail]

What's your thoughts on json for prompting? I haven't delved too deeply, but i hear it helps with dialog

[past linden]

@charred sail lol thank you

[dry frost]

I'm just started to dabble with dialogue and I'm also using json. I think it's pretty good at setting pacing but i need more time to find out the proper set up.

[charred sail]

Does it land where its supposed to with multiple people?

[dry frost]

I'm focusing on single subject material at the moment so I couldn't tell you

[charred sail]

I think i have a video i can test it with.. it's been driving me nuts

[dry frost]

The bird lady I posted in #flow-videos・📹 was scripted in json. I just need to figure out words to properly inflect and also time each line.

[woeful fossil]

Ay that looks familiar 😉

I find that Json prompting has a similar rate of inconsistencies and it makes it harder for people to make modifications to their own prompts. Plain text is more user friendly which I definitely wanted in my app.

[charred sail]

I tried getting this lip synced. Couldn't do it. Any ideas?

I was hoping bigfoot face would be easier to sync but nope

[echo flame]

Beautiful African woman

[woeful fossil]

Lip syncing them to the song? Yeah... not a real simple way to do that at all with Veo at least.

[charred sail]

Yeah, i figured one of the editing tools would do it, but they all kinda suck at animals

[steel scaffold]

Am I missing something or are you not able to extend a video to have an additional scene with audio? Whenever I try to extend an existing clip it says veo 3 isn't supported and switches me to veo 2, but there's no audio.

[woeful fossil]

Extend, no. Use jump to.

[steel scaffold]

What's the difference?

Does it act as a sort of "cut" transition?

[woeful fossil]

Well, the scenebuilder was allowing me to scrub to the end of the clip and click Jump-to and then it automatically took the final frame and added it as a starting frame for the Jump to generation and it used Veo 3. A pseudo version of extend.

[placid sluice]

not sure if this has already been brought up...but what happened to the quality of flow generations? they were amazing and now it looks like a broken version of Toy Story

[lapis wraith]

lol

you an seee a guy actually holding the drone and filming

[steel yacht]

guys i need your help pleaseeeee lol how do i extend a video in flow what does that workflow look like

[nimble scarab]

That is fun

[steel yacht]

guys how do i delete from the timeline. i need to delete all those clips

[nimble scarab]

There should be a delete key in the editing software, and that area will be removed.

[steel yacht]

dont see one anywhere can you please do a screenshot?

[nimble scarab]

I'm new to these editing software. What are their names?

[woeful fossil]

Click the little minus symbol.

[steel yacht]

yeah i finally got it to work ,m there were some orphan files. thx

so i can create a video from image and add audio with veo3 but i cant extend it? it says that is not support it and to go to veo but then there is no audio so how do i get audo, is that being done outside google with like elevenlabs?

[bright agate]

Anyone know, How to create 9:16 text-to-videos in flow. (like a prompting style or something?).

[stiff harness]

You can't

[wintry oasis]

Hi, I have google creds in my google console, is there any way to utilize these google creds to access google flow & veo 3 ?

[nimble scarab]

Flow uses a fixed aspect ratio and doesn’t support creating 9:16 vertical videos through prompts.

[muted bluff]

Do you know when Flow will be available in Germany?

[civic bison]

Someone with this issue? Thanks

[sharp quest]

you can

but only frame to video for now..put into premiere pro thats it

[lost plank]

Yes, just talked about it yesterday in here, Still no mod answer on this subject. I believe (and this is just a tough) it has to do with Google being specially careful with strict European rules. Not sure, but that is what think its most probable, as i know that is not a problem in US. You are in Europe, right? I got the same problem, so i believe is affecting everyone in Europe.

[civic bison]

damn!

[bright agate]

Yes, that's working. However, I needed a text-to-video prompting method. Finally, I tried these wordings and got results like a 70% of success rate :

1. "The composition is made to look like it's filmed through a vertical phone screen. This video is designed to be cropped or rotated later for vertical viewing on social media."
2. "The scene is designed to be rotated 90 degrees for vertical viewing on TikTok or Instagram Reels, creating the feel of a portrait video."

I think this is a little good if someone wants a portrait size.

This method has a 40% success rate. These methods are little bit good than cropping 16:9 into 9:16. Because it will cut your video's good parts.

[faint wyvern]

You asked if anyone bought this course and I told you I actually paid for it. And now you’re messaging me in private asking for it for free? Seriously, man… I pay money so you can get it free? Nice one. Really loving this about you.

[sharp quest]

So, what did I write to you privately yesterday?

[sharp quest]

Hey! That course you got — over here, the price feels like it's made of gold. When I saw you had it, I thought maybe there's a hero out there. If you're open to sharing, I'd be super grateful, but if not, I totally understand. I've been trying to build my own prompt template for over a month now, and I've burned through nearly 2000 credits with no real success… I'm honestly just exhausted at this point. With the exchange rate where I live, resources like this are a real luxury. Thanks so much for even reading this — and if you decide to respond, I'd truly appreciate it.
Also, I’ve got some experience with Veo 3 myself — if you ever get stuck on something or need help, I’d be more than happy to support however I can.

[sharp quest]

Actually, the only thing I was stuck on was the environmental details. I didn't know you were so depressed..People ask me things or want things too..but I don't try to label them like that and embarrass them.

[steep wren]

ANy ideas how to stop the zoom in on Veo 3 Frames to Video?

It constantly zooms in on characters

[vernal kernel]

Is this a common issue, that when downloading the clip from flow to computer it loses audio?

[vernal kernel]

Never tried frame to video, sorry :/

[versed shale]

HOW TO GENERATE JAMAICAN VOICE IN VEO3????????????????????????????????????

HOW TO GENERATE JAMAICAN VOICE IN VEO3????????????????????????????????????

[jagged jackal]

Anyone here know if Veo3 on an API will work in a blocked region?

IE: if Veo3 blocks me cause of my region. Would using another services useing Veo3 API that is not blocked in said region allow access?

[nimble scarab]

Can we share videos here, guys?

[silk ember]

does anybody know why it says flow TV when I click create with flow?

[nimble scarab]

[silk ember]

this is what it shows me

[dry frost]

You are most likely using a currently unsupported browser and/or mobile device.

[woeful fossil]

Hey sorry about that. Try now. It should be fixed.

[sacred dawn]

yes all working thank you

looks amazing so far

[placid hare]

Hi everyone! I don't know if this is the correct channel, but I have a question. I'm trying to create a video in VEO3 Fast using the "frame to video" option, but it says that "This feature isn't supported on Veo 3, switch to Veo 2". My goal is to combine my avatar with a product that I can talk about. Does anyone know how to solve this?

Creating on Veo 2 would limit my video because it doesnt have audio

[twin mortar]

@main saddle any update on the numerous technical issues that are plaguing Veo?

[red drift]

dropped a new video this morning, already passed 10k views.

[dreamy lance]

How do you guys create funny scripts? Whats your approach? I am aiming to create as funny content as the famous big foot videos but with different characters and a different setting...

[abstract hedge]

@sharp quest Hello 👋, please I'd be very grateful if you share the knowledge you have on Veo 3 because I'm kinda stuck in the process of making more videos despite the fact that I've bought like 1,000 credits..
I'm really confused along the line , please can you share more knowledge of what you have please.
I'll be expecting your response please 🙏🥺.

[ancient mulch]

What did you use for lip sync?

[woeful fossil]

The easiest way is to just be funny.

[sharp quest]

first bigfoot vlog try video

second bigfoot vlog try video

[tired juniper]

hey! does anybody know if it's possible to connect Flow to a GCS bucket like it's done in the Veo API? i'd like to store generated videos in the bucket

[dry frost]

There's always just uploading the files directly, would get more control on how it's organized I'd imagine.

[warped arrow]

is it me, or Veo3 is getting worse when 2 people are talking to each other, like lines/dialogs not matching the name of the speaker. If there are 3 lines alternating between 2 people, the AI doesn't know who will be saying the 3rd line.

[prisma flint]

Hey anyone at <@&1173646484311384114> can you help me please?

Hey @woeful fossil can you help me on whether Google Veo3 has the potential to include charts and stuff in your AI video generated content?

[charred sail]

That's because it doesn't pay attention to format. It's going off of context. The scene needs to be very clear on who should be talkiing.

[orchid atlas]

hey how can I clear memory on veo 3?

[verbal flame]

@vernal kernel

[woeful fossil]

Its going to be best for you to add those in separately. Ai struggles with rendering accurate text and figures. If you need it to be a useful chart youre not going to want to use Veo. If you just want it for illustrative purposes you can always try.

[verbal flame]

Você viu uma história do macaco comendo banana em inglês? Posso te ajudar com isso! Aqui estão algumas possibilidades:

---

🐒 História curta em inglês: "The Monkey and the Banana"

Title: The Monkey and the Banana

Once upon a time, there was a little monkey who lived in a big tree in the jungle.
One morning, the monkey woke up very hungry.
He looked around and saw a big yellow banana hanging from a branch.

"Yum! My favorite!" said the monkey.

He jumped from branch to branch, got to the banana, and ate it quickly.
"Delicious!" he said, licking his

[sharp quest]

https://www.tiktok.com/@big.foot.life/video/7527038186653027592

[vernal kernel]

The 1000 credits run out so fast 😢
Has anyone kept count how much credits they typically burn on a project?

[bronze fiber]

You got pro

?

[vernal kernel]

Yes, I am just a dumb bus driver, so can't really afford to throw all that much to hobbies 😅

[bronze fiber]

Do you do veo 3 fast 20 credits

Per

[vernal kernel]

Yeah, that's right. I did try twice the quality, but have to say, that I was not impressed 😄

[bronze fiber]

Do 20 per also does it give you a option to buy more credits I could swear they updated it where you can

If you wanted to

[vernal kernel]

On the flow site, on the top right is the balance and "buy more" and that just takes to "upgrade to AI Ultra plan", so I think in the pro plan, it is not possible to buy more credits.

[bronze fiber]

Yeah

[steep wren]

What is the best way to create constant cahracter iamges? I figured out it's ChatGPT.

[vernal kernel]

Seen people using midjourney for that

[tepid lotus]

Create realistic bigfoot talking vlogging

[woeful fossil]

The Pro plan definitely does allow you to buy more credits.

[bronze fiber]

Does it I

You sure

[ancient mulch]

Depending on the project really, short video clips are around 1k. 15k if I'm careful with a larger project like a music video. Less is more sometimes with these prompts. Veo 3 fast loves short and simple prompts, that's where it's strong. Sometimes you have to work with your word play.

[dry frost]

Understanding structure and verbiage in the prompts is a big learning point.

[vernal kernel]

Okay, 15k does feel like a lot, for a 3-4 minute music video maybe, but might be what it is. I certainly have wasted some credits due to messing up my wording in the prompt already, translating things quick in my head from finnish to english and getting confusing results and after considering a while why it happened can be painfully obvious in hindsight! 😄

[vernal kernel]

I went to read the support page on google, and it turns out you can buy AI credits with the pro sub in countries that support BOTH Flow and Whisk. And as it so happens, Whisk is not available in my country (Funland 🇫🇮 )

[woeful fossil]

Welp. I do wish verbiage on regional differences was more clear. Definitely leading to a lot of confusion.

[vernal kernel]

Yeah it's a swamp 🤷‍♂️

[bronze fiber]

Ok good

[ancient mulch]

What I have noticed is that even though I tell it to have a delayed sound for an explosion, it doesn't...

[sweet dew]

@ancient mulch those are nice 👌

[dry frost]

Establishing a (proper) break is my current testing with dialogue, so I might get back to you on this.

[ancient mulch]

I would appreciate that, because I've tried a few different things that didn't work consistently.

[bronze fiber]

What the key to getting lots of likes

[placid sluice]

the Veo logo didn't used to appear on my videos

oh, it's only ultra members that don't get the logo. ugh

[random yoke]

Made a discord bot that does my work for me. Still a major work in progress lmao 💯.

[ancient mulch]

You can use 3rd party apps to remove it, or you can always crop the video a bit. 😄

[bronze fiber]

lol

[ancient mulch]

Who makes these?

Loved it! 😆

[sharp quest]

me 😛 thanks

[ancient mulch]

This one was my favorite

[ancient mulch]

Cool! My cousin sent me this one and I was dying of laughter. Lol

[sharp quest]

hahahahahaha

[sharp quest]

holy 😛

[ancient mulch]

Love it 🤣

[dry frost]

Wearing the suit rather well.

[inland arch]

🤣🤣🤣award winning video

[nimble scarab]

Haha, I'm flattered

[ancient mulch]

Gold 🥇

[muted tinsel]

is uadio glitchy rn?

[mystic crypt]

Hey @unborn bane how did you do the parrot video that clean u posted ont he 12th i wanna say?

@woeful fossil Hello!

IF i don't know how to prompt, would ur tool help me for sure?

[steel yacht]

hi guys, is there a way to get around this message that i keep getting every time i try to upload an image created in midjourney that does not look like any celebrity, but i keep getting stopped because it says is "Our policies prohibit the uploading of prominent people."

[woeful fossil]

Yeah, I mean the whole idea is you describe something very basic and it will elaborate and format it in a way that Veo understands well. The better the prompt you put in, the more specific your outputs will be. It is free to try though so give it a go and lmk how you like it!

[earnest kraken]

anyone know how I get these vids to be more satisfying Im trying to make it better but its a little buggy still. Ive been trying to fix my prompt for 2 days but i cant seem to figure it out.

[woeful fossil]

Whats your current prompt? Your premise is obviously so specific, and it seems to be hitting the beats, but failing slightly in the persistence of the cars. I think that's just a byproduct of having so many and I'm not sure you'll be able to significantly reduce the errors with prompting.

[earnest kraken]

got it that does make a lot of sense, I am making videos so ive been trying to use chat and Gemini to edit the prompt when something went wrong

Here is my promt
Create an ultra-satisfying, hyper-realistic ASMR macro video filmed in one continuous, unbroken, real-time shot — with absolutely NO cuts, transitions, fades, glitches, skips, or time jumps at any point. This video must show every single part of the process visually, in real time. Use a fixed, slightly pulled-back and slightly elevated camera angle, like a professional cooking video. Show the entire black luxury plate with gold etching, the full slice of warm golden-brown toast, and a bit of the surrounding surface. The camera must not move or zoom. A black-gloved left hand is already placed at the toast's edge, gently pressing the crust with the index finger and thumb only. This hand must stay completely still and not obstruct the plate or toast. From the right side, a black-gloved right hand pours 6 to 7 miniature yellow Porsche 911 GT2 RS cars with matte black rims from a dark bowl directly onto the center of the toast. The pour must: Be smooth, realistic, and glitch-free Be fully completed within 3 seconds Show each car falling naturally and satisfyingly Ensure the cars land evenly spaced and fully on the toast — no cars may fall off Immediately after the pour, with no delay and no camera cut, the right hand enters holding a completely clean ornate black and gold butter knife. The knife must: Enter the frame clean, with no butter or residue Begin smearing every single car in 1–2 slow, broad, real-time strokes Visibly crush, melt, and spread all cars underneath the knife — clearly on-screen Each car must: Melt and transform into a smooth, matte yellow butter-like spread The color of the butter must exactly match the yellow paint of the Porsche cars The transformation must happen only through visible knife motion No fading, no skipping, and no jump to a finished spread You must not skip or transition to a fully spread toast. You must show the entire smearing process, fully and satisfyingly. Do not hide the transformation off-screen, or replace it with a finished image. By the end of the 5-second smearing phase, the toast must be: Evenly and completely covered in matte yellow car-butter With no visible cars remaining With the entire melt-and-spread process done by the knife and clearly shown

[woeful fossil]

Yeah you've got things pretty tightly dialed in. I do like structuring things and organizing some of the direction notes so theyre not just strewn around in the middle, but at this point... I'd say you're being limited by the model itself.

[round chasm]

hi guys anyway to extend scene in veo 3 with audio

[quiet stream]

is midjourney the best realistic image generator currently for image to video with veo3?

also heard about flux context, seems good apparently

[buoyant topaz]

Hello! Can someone help me understand how to keep character consistency between scenes in Veo3?
I'm trying to make sure the same character looks identical across different shots. Any tips?

[red drift]

closing in on 600k views for my video, first mega viral post.

[modern wave]

Is it not possible to generate videos in 9:16 format

[red drift]

Not yet no

[modern wave]

ok thx

[frigid pier]

lol new ASMR dropped

[steep wren]

Consistent characters, flow, accent, everything. Friggin insane, very happy about the results. I know what to improve for next time too!

All done with Veo 3 and Gemini. Nothing else is in play here.

https://twitter.com/Quantinium_fdn/status/1945147421698265434

Quantinium | The Decentralized Wi-Fi Network (@Quantinium_fdn)

. @Quantinium_fdn x Tesnet x @avax , powered by @AvaCloud 🤝

X · Jul 15, 2025, 3:44 PM

[ancient mulch]

It's not perfect but you know... 😆

[frigid pier]

too weak supaman

[ancient mulch]

After playing around with it a lot I think this is probably the best I could do with it

Different cutscenes are difficult when you only have 8 seconds

[solemn stirrup]

how do i get my credits back because i created a video with a 22 year old but apparently 22 year olds are minors now and doesn't make sound

do i have to make them like 80 or something

lets waste all my credits to see what veo3 thinks an adult is

ok apparently a 30 year old is now a minor

[dry frost]

Managed to space it out. Still need to work on inflection.

Adding the term "short silent pause" in an Action block is how I got it to work

[solemn stirrup]

oh woww exciting! more credits wasted, and apparently a 40 year old is now a minor. jesus, where are the announcements of all these new laws coming into place?

let's see if veo3 thinks a 50 year old is a minor

no lets just skip all the way to 80

[woeful fossil]

What is your prompt lmao?

[solemn stirrup]

my original prompt was: someone brushing a korean 22 year old males hair, and his hair is made of actual acid, the acid is his hair, it doesn’t fall on the floor, because the acid is attached to his head and the person brushing. the boy giggles as he is watching a tiktok on his phone scrolling you can hear music from his phone

[woeful fossil]

Dont use the word boy.

[solemn stirrup]

but that's just another word for male

so i wasted all those credits over a word when it didn't even add any children into it

[woeful fossil]

It means a male adolescent or child... of course it flagged. Try this.

https://promptalchemist.com/share/ftX8dAlSDU1s2SY7pGvA

[solemn stirrup]

so there's no way to get the credits back even though it was veo3s fault?

[woeful fossil]

No. The model still generated what it could within it's policies. You only get credits back for generations it produces nothing for.

[solemn stirrup]

oh well, i don't blame them, it keeps the money rolling in

[woeful fossil]

Lesson learnt, now create some cool stuff.

[ancient mulch]

That's great!

[ruby crag]

how did you get thhis ? can i have prompt please , thanks

[rain sky]

Guys is there anyway we can use flow for free ?

[ruby crag]

man flows exp 😦 and thhe credits finish so fast even for the prompts that you dont even want when it give u something different from what u tell it too

[night shell]

hey guys!

im having trouble generating audio with veo3

few weeks ago all prompts were generating audio but since last week 1 of 10 videos generate audio

any idea what could be the issue?

[bronze fiber]

Scams

[rain sky]

Thats sad to hear maybe you should highlight "about audio" in veo3 this should work

[vernal kernel]

Good evening guys, what are we working on today?

[rustic sand]

in what setting can we use ingredients to video?

[sharp quest]

tiktok link https://www.tiktok.com/@big.foot.life/video/7527367236193979655

[quiet stream]

i didnt know you could post landscape videos on tiktok, is that new?

[sharp quest]

yesssss

[formal vessel]

Create video

[earnest kraken]

dang that sucks. So is there any way I can adjust the prompt to make it smoother and more satisfying?

[earnest kraken]

holy crap dude that’s so sick what video did you make?

[odd basalt]

How do I prompt flow to stop adding subtitles to my videos?

[oblique gust]

Did youtube just announce that AI videos or images are no longer count and you can't make money from it anymore?

"YouTube clarified that creators using AI tools or reposting content from other platforms will still be allowed to monetize, provided their uploads feature clear additions such as original commentary, educational input, or entertainment value. The platform added that it encourages the use of AI tools like Dreamscreen and auto-dubbing to enhance storytelling, as long as creators follow transparency and disclosure norms.
“YouTube welcomes creators using AI tools to enhance storytelling, and channels that use AI in their content remain eligible for monetization,” the company said."

So i guess if you make orignal contents out of Veo 3. I guess you will be fine?

[woeful fossil]

Basically just don't make low effort slop and you should be fine.

[sharp quest]

Will we continue to make money from videos featuring fictional characters like Bigfoot?

[oblique gust]

No idea lol

[mental mulch]

Google Gemini's update messed up the timing of when I can generate Veo 3 again :/

It just says Something went wrong...I can tell its exceeded but now it doesnt have a timeframe

It says something went wrong, thats it

[jagged delta]

Hi

[vernal kernel]

been 3am for me

[ancient mulch]

Man.... This is looking too good, I'm convinced you're dressed up in the suit with your friend and just put the Veo watermark to mess with us. 🤣

[sharp quest]

Oh no... the truth is out 😅

[charred sail]

LOL i see you ran into the sideways slide too..

I tried makiing them snowboard... they kept going uphill or sideways. 1 out of 10 would go forward

I might have found a way to make it move the way you want.. I found using the word trajectory helps. Like a forward trajectory or upward/downward.

worked with my ramp jump

[sharp quest]

@charred sail Thanks buddy...I noted this information..I'll show you my next video.

[sharp quest]

My only problem right now is that after entering the prompt for the first scene in chatgpt, chatgpt renders the following scenes incomplete or incorrect.

[edgy vale]

that's amazing what prompt did u use

[wise tulip]

Intriguing Characters: The man's hidden face and the woman's spellbound reaction create immediate mystery and draw the viewer in. It suggests the perfume has a powerful, almost supernatural effect without being overt.

Focus on the Product: The Aura Vault bottle is central to the action, and the slow-motion spray highlights the product itself.

Subtle Storytelling: The narrative of attraction without direct interaction is sophisticated and effective. The man never turning adds to his mystique and emphasizes the perfume's allure.

Clear Messaging: "Unlock Your Aura" is a fantastic tagline that perfectly complements the visuals and the implied effect of the perfume. It's aspirational and fits the high-end aesthetic.

Sensory Details: The "mist floats through the air, catching the light" and "scent lingers" appeal to senses beyond just sight, which is crucial for a perfume ad.```

[fluid crystal]

what do you guys think?

[sinful bramble]

perseolis

#flow-general・🌊 Persepolis

[sinful bramble]

Create clip from Persepolis

[mystic crypt]

<@&1366831844502274180> SPAM

thanks

[dreamy shore]

Is it possible to generate 9:16 vids in veo3?

[pure lodge]

I think so

[sacred isle]

Hi guys! Do employees of the Flow team look at the posts in the bug section and process them? Because whats the point of a bug section if they dont, having a talking therapy for people to shake off their frustration? I want to give those guys my money but they dont let me. Does anyone have problems to add AI credits too - maybe especially from Europe?

[twin mortar]

A little acknowledgment and a list of bugs currently being dealt with would be useful.

I’m paying for Ultra and renewed this month’s subscription on 9th - I’ve yet to use more than 100 of the credits because the results are continually spotty with either shocking audio and/or captions.

I’m trying to sell AI products to clients and currently I can’t guarantee them a level of quality because of the random bugs.

[vernal kernel]

If you have Pro plan, you can only buy credits in countries that have BOTH Flow and Whisk available.

[twin mortar]

I can however 100% guarantee that Google will take payment for next month’s credits regardless of whether they’ve fixed the bugs or not.

[vernal kernel]

Always use "no subtitles"

[twin mortar]

Premium prices does not equal premium service, it seems.

[twin mortar]

Mate I’ve tried everything. No subtitles, no captions, subtitles: off, captions:off and every variation in every possible place on the prompt. Still get them.

[vernal kernel]

Welp, it is what it is. Musky boy, is starting to train his own video model with 100k gpu's in a few weeks, so we might be getting a model even 5-10x better than Veo, if the quality goes linear with the amount of compute...

[charred sail]

I've found it works better like this Exclude: music, subtitles, captions, watermark, text, words on screen, logos, and branding.

[jagged delta]

Hi, tell me how to generate a video, I don't understand

[vernal kernel]

I'll try this next runs👍
Though I have not had much trouble with that issue.

[charred sail]

ever notice how it says beta audio? rethink your expectations

[charred sail]

Yeah, I think it has a lot to do with how much you're trying to pack into that 8 seconds.

over time you just figure out how much it can do in 8 seconds and you see it less.. my theory

[sacred isle]

Ah, thanks! Does it mean I could add with Ultra?

[vernal kernel]

I don't remember straight about that, so better not promise you anything, as I only read the Pro part in Google's support page😂

So, if I read that right, even in the Ultra plan, you can't buy additional credits if both Whisk and Flow aren't available. Correct?

[twin mortar]

Yeah I get that. But the audio was working fine last month. This month though… 🤷🏻‍♂️

[sacred isle]

Yeah seems so. I wonder if they could seperate Flow and Whisk AI credits to enable people from countries without Whisk to purchase credits for Flow.

[vernal kernel]

Surely they could, but maybe the focus is atm working around the EU regulatory hurdles what not and lawyers reading the fine print to make sure they can just put it all out there for people to use? Idk, just my guess, but who knows.

[vernal kernel]

I don't know what kind of audio you are looking to achieve, but there are also workarounds: you can produce the audio track independently with other AI tools and use lip sync AIs, and so on.

[lilac violet]

Hey Everyone, does anyone know how to get results with 1:1 Similarity to a real person in Veo3/Flow? I tried a bunch but all results fastly change the look of the face, so that the original Person is not 100% recognisable

When i try to get a Realistic Picture of a real person into the Frames to Video Mode i get told its not allowed. is there any work around to that?

[upbeat rivet]

Can anyone tell me how to make a series of videos of the same person talking about the same thing? Because, I prompted the first video, and it turned out great. I hope I can create more videos of the same stuff.

[ancient mulch]

Probably my fault for all the videos I made of my father in-law crying and pointing a Glock 40 at the camera. He thought it was funny but I guess the guidelines didn't. 😂

[oblique gust]

this is a shot that i did for my series 😄 Dude is a full youtuber now lol

[ancient mulch]

😂

[oblique gust]

I do have a guide to use for Veo 3. If anyone wanna give it a try then pm me. Free. I'm still working on perfecting the guide

[median hearth]

anyone else having issues with audio on fast model? been using it all morning no problems, but none of the new clips generate audio.

[oblique gust]

veo 3 fast?

[median hearth]

ya

[flint summit]

Exactly. Keep it simple and you'll rarely see captions.

[oblique gust]

are you using text to video or frame to video?

[median hearth]

text to video. its been working fine all day for me. just the last 5-10 mins of prompting has been the issue

super odd was just wondering if anyone else experiencing it or just me.