flow-general・🌊

[grand zealot]

Lines like "A snuzzled whinglegosh shizzles through a vlorpy nichtwahlen"

Something like that, yes 🙂

[west rampart]

a man on the plane

[wise cypress]

Is there an option to change the default video model in Google Labs flow to veo3

I hate when I forget it's accidentally on the wrong model

[frigid pier]

bro. Veo 3 is insane.. great imagination

trad vfx industry is in serious situation :/

[craggy tiger]

Hello guys is it possible to create longer videos then a few seconds? If so, how can i manage to do that, thanks! 🙂

[frigid pier]

u gotta extend the video

u can use the feature in flow app.

[fair bridge]

I've burned about 8 Veo3 clips I can't use with these random gibberish subtitles despite all sorts of prompting against it, and one of my failed generations didn't refund. Super excited but my first night on Ultra is more frustration than anything

Not sure what else to do when 80% of the generations with speech put mangled text on the screen

Maybe I will just wait a couple weeks and finger cross it improves

I love it, it's amazing, just expensive to burn credits

[frigid pier]

maybe zoom in the video in your video editing tool?

[fair bridge]

Definitely a great idea - someone else mentioned last night about having the camera set to a wider field of view to have room to crop out the subtitles, wil try that. I am super excited about veo3 I need to remember its still experimental lol

[frigid pier]

yea right. but i am worried about after 3 months u know the sub price will jump 2x

gpt told me i need to stop sub it after 3 months if i am not making any profits from ai videos lol

[fair bridge]

This is just the first public experiment, it will get way better I have no doubt

Like using Runway since Gen-1, I was so amazed and mixed it with D-ID for lip synch shots I was so blown away, all these things will get ironed out

Gen 1 was so buggy lol

[frigid pier]

would u pay the doubled price tho

[fair bridge]

Not sure...

[frigid pier]

imagine it if u sub it like 4 months then 1k usd is gone :0

[fair bridge]

Right now no

[frigid pier]

1k for 4 months is really too much imo

[fair bridge]

I took a lot of debate for me to get it, I cancelled some other subs

[frigid pier]

i didn't sub ultra plan yet considring it tho.

and yes I canceled other subs too

I cancelled gpt runway midjourney

started realizing this is becoming a finance blackhole :0

[fair bridge]

my 10 in the google ai plan I got made me want to get ultra, only one or two had subtitles

[frigid pier]

oh

[fair bridge]

I think that when the kinks are ironed out this will replace lots for me and the output will be better

[frigid pier]

yea i believe so

kling released their new video model but it is not getting attention like veo3

[fair bridge]

Like I cancelled Runway unlimited and my Freepik

[frigid pier]

so google entered into some weird stage for now.

[frigid pier]

yea i love Runway unlimited tho

even tho it's relax mode it's still fast imo

[fair bridge]

The leap with voice is huge from gen4 with workflows for lip synch to just instant with veo3

[frigid pier]

I hope that Google release a new plan for specifically Flow app
and offer release mode like Runway later

The current Ultra plan is a too risky cost for normies

[wicked ravine]

what’s the over under on the scene builder features being available for v3

[frigid pier]

i know right

[fair bridge]

I think that the model will get tightened up, maybe the cost stays the same or goes down a bit, we can already buy more credits so the pricing model probably gets adjusted

[wicked ravine]

like how long

[fair bridge]

I can see a lot of companies buying a more rock solid version of veo3

[wicked ravine]

i love veo 3 but it’s impossible to do anything with characters

[frigid pier]

bro we just gotta cry for that somehow so google listens to us like sweet parents xD

[wicked ravine]

https://youtu.be/quYn53JPXY0?si=NpbcWv3JToIrNXHi

[fair bridge]

haha they are here hanging out in the discord right

[frigid pier]

i mean it's still experiment. not a good idea to have a high expectation on it for now

[wicked ravine]

ironically only way to get characters is to use ip

[frigid pier]

for me, if i am going to use veo 3 then it's only for short clips

[wicked ravine]

https://youtube.com/shorts/7LQ7Fk8RUh8?si=TGQWYcscNMlGBLLE

it renders known characters very consistently

i love veo3 just wish they would open the gates for us

[frigid pier]

wth it can say motherf* word lol

[wicked ravine]

lol yeah it’s rated r model

[frigid pier]

I realized Veo 3 has some weird audio quality hmmm

i hear something always throug the sound

u know what i am talking about

[wicked ravine]

i just ran out of credits and dropped another round lol

https://youtu.be/-B4rOwXOuAY?si=rgo3shjRZsF0SasK

here i experimented blending veo with live action

i was pretty pleased

[fair bridge]

That had me laughing nice work @wicked ravine

[wicked ravine]

haha thanks!

haha

kling 2.1 is better than veo 2

[fair bridge]

Yeah I am new to veo I was using runway since before gen1 so kind of used to that whole thing

[wicked ravine]

i love runway

https://youtu.be/1bZ386MuaVc

did that in runway

[frigid pier]

runway is good yes.

[wicked ravine]

what kind of stuff you make?

[frigid pier]

it is. it may be very cheap for real product but yes

[wicked ravine]

that’s dope love the styling

[frigid pier]

bro gotta breathe lol

[fair bridge]

I just use it as an expensive hobby, I work in marketing so this all comes in handy but mainly it's just fun

[frigid pier]

I think Veo 3 is already used in many companies for their profits

[wicked ravine]

good to know

[frigid pier]

so they started making more profits becuz they do not make contracts with real video producers.

RIP

job take over is for real :/

[wicked ravine]

i work in video production hardly anyone on my staff uses any of it

it’s surprising

[frigid pier]

did u guyys see what Claude founder said few days ago?
he said the unemployed rate will skyrocket to 20% in 5 years
we are f.ed up

bro 20% is no joke. it means society collapses

the teenagers and 20s are f.ed up RIP

[fair bridge]

that's nice @wicked ravine what did you use for the lip synch in the cops on runway just standard synch or did you use the "record your own talking" feature, forget what that's called

[frigid pier]

I think they hate AI.

there are some people who just block insta accounts using ai

[wicked ravine]

i used hedra

[frigid pier]

there are some how many peeps who hate people making money with ai

[fair bridge]

Most of my coworkers don't touch LLMs still

[frigid pier]

they hate ai

that's it

[wicked ravine]

those are the 20%

[frigid pier]

bro lol

[fair bridge]

I started with just playing around when Alexa came out w ai stuff like trying to make voice games

[frigid pier]

ohhh right. what happened to Alexa? i dont hear anything from it recently

[fair bridge]

yeah I have no idea I feel like it is stuck in the pre-LLM voice generation of technology where you used dialog flow to program conversations based on keywords/intents lol

now you can just chat away with an llm

It would be great if google pushed the google home more, I think they scaled that back

[wicked ravine]

i just want image to video Veo 3

how long you think?

[fair bridge]

before apple flopped with their headset I might have disagreed but yeah it starts to feel like at least for now people dont want to chat with computers, even having had the chatgpt voice app on phone for... 1.5 years? I only use it as sort of a novelty

chat verbally out loud / voice

[frigid pier]

I kinda think this can take time

Google is about to release new Ai studio version. I am hyped

[fair bridge]

I think it will get there and become normal, but it's kind of like the ecom dot com bubble with pets dot com and everything, where it was like, well of course people will shop online, then no one wanted to... But now here we are today, just takes time

[frigid pier]

also figured out Google devs working really a lot lol.
Like omg

The competition is overwhelming u know

[fair bridge]

lol when I saw the first veo3 demos I was like man, the whole effort is over its magic, and started learning vibe coding as a new hobby

but I am super excited about veo3 just this subtitles thing is killin me

[frigid pier]

yesterday I saw a guy making over 100k with vibe codinig :0

[fair bridge]

What was he doing

[frigid pier]

he was having multiple clients

so he delivers web services to them

he was just doing copy and paste his boilerplate lmfaoo

smart move.

[fair bridge]

I made a terrible but passable landing page for a client just slapping things into Replit just to show people at work more what is coming, I was even surprised it worked because I did it live on a zoom my second day of playing

not good enough to use, like not good quality professionally but just an indicator

yeah thats the thing its fine for a static site or an app or something but not for a live site that needs continuous updates

I am going to make a trio of really bad web games with replit for fun

the next thing I want to learn is n8n

Do it!

[frigid pier]

anyone working in real video production industry?

[fair bridge]

I work around it, adjacent

[normal grotto]

FlowTV is ... oddly attractive to just leave on in the background

[frigid pier]

[cosmic lion]

Disappointed to discover that I can't use a reference video or even pre-recorded audio to guide sync and image generation in Veo 3. Am I missing something? Are ingrediants really only for stills?

[frigid pier]

whats Hedra for?

[wicked ravine]

what’s Sync

[frigid pier]

i think its for lip sync

[cosmic lion]

Seems to work in Runway and Kling. I have an Ultra account so...

[wicked ravine]

never heard of sync

[frigid pier]

wdym u mean u can use flow app?

[wicked ravine]

hedra i use a lot

[frigid pier]

there are too many ai tools now xD

whats hedra for

[wicked ravine]

lip sync

can sync do long audio like hedra?

[cosmic lion]

There are a lot of lip sync tools. My issue is that I want to control the perofomance of the characters by recording my own and using a AI voice. That's easy with eleven labs. Just can't believe VEO 3 forces you to only eneter text for voice creation. What I don't like the voice or the perfomance?

[frigid pier]

xD ok

I think Lip sync tool will disappear what do u think of it?
I think people want videos like Veo 3 with voice

[wicked ravine]

which one you use hedra only one j know that can do long audio

[fair bridge]

Lip synch in Runway is not awful but there are better things out there I've just seen papers for, what else is there

[wicked ravine]

yeah but only 30 seconds

[cosmic lion]

I had good okay luck with Runway and lip sync. But the physics aren't even close to Kling or VEO 3.

[wicked ravine]

https://youtube.com/shorts/44Db__a1M2Q?si=vBwYx2yz3aFCN9zu

i did with only hedra

[fair bridge]

Oh there was that Wall Street Journal demo showed a good example check this one https://www.wsj.com/video/series/joanna-stern-personal-technology/we-made-this-film-with-ai-its-wild-and-slightly-terrifying/D17B233B-1E06-400D-9095-B5247306DD38

[cosmic lion]

Well, my liittle idea is character driven so I need perfomance controls. Maybe we aren't there yet.

[fair bridge]

you mean like mocap? I think the one runway tool lets you move arms and captures face and things for lip synch then you just voice change yourself in elevenlabs

but its a static shot

One hilarious tool I found is wonder labs

[cosmic lion]

Seanovision. Funny viudeo - and super easy to sync a robot with no mouth! Still, they did well with consitency and story. thanks.

[fair bridge]

This is so old some random thing I made with it

[frigid pier]

u made it with trad 3d tools?

[fair bridge]

this thing wonder dynamics, I got the beta it came out maybe 2 years ago?

[fleet hatch]

I love Veo!

[fair bridge]

replaces actors with 3D models

[fleet hatch]

Flow*

[frigid pier]

RIP VFX

[frigid pier]

the last one i can feel it's ai

[cosmic lion]

Guess I'll use VEO 3 for the objects and cars, and Runway for the people who need to be in sync to my performances (re-style function).

[vast remnant]

are flow and veo th same thing

[frigid pier]

What was the prompt

[fleet hatch]

I thought I was going to see a Balrog haha

Currently using Veo3*

[vast remnant]

so it is the same thing?

[frigid pier]

This could be a good scene for Scifi or fantasy

[fleet hatch]

More "Flow" with the UI

[pseudo gorge]

Realistic 8-second nighttime footage, as if recorded by a wildlife trail camera. The environment is a dense forest with fog, moonlight piercing through trees. A mysterious large animal slowly walks into frame. The creature is the size of a liger. It has the eyes of an owl, ears of a lynx, a feline body with panther and jaguar markings, a cow-like snout, lion-like claws, and a fluffy rabbit tail. It moves cautiously, its glowing owl eyes scanning the dark forest. The scene should look 100% photorealistic, cinematic lighting, shaky handheld night vision effect like a documentary or hidden camera footage. Atmospheric forest sounds and distant owl calls in the background. Animal briefly looks into the camera before vanishing into the fog.

[vast remnant]

ngl idk what half of u guys r saying i need u guys to dumb everything down

[frigid pier]

?

[vast remnant]

i’m jst tryna make tiktoks blow up and get some money 💔

[fair bridge]

I've been getting so many failed generations tonight

[frigid pier]

how u make money on tiktok tho

[fair bridge]

I wanted to make one video that's maybe 9 clips and still trying to get the first one lol

[vast remnant]

get monetized

[frigid pier]

damn veo 3 is cool

[frigid pier]

but i think u gotta reach some requirements before that?

[vast remnant]

10k followers average 1m views a
month not that hard if it pops off quick n all the ai vids are

[fair bridge]

Should make a virtual influencer or something maybe

[vast remnant]

this isn’t dumbed down enough ngl bro

[frigid pier]

i think some folks making money with it

[vast remnant]

ill give someome $3 if they let me use there acc 💔

this free trail is horrible

[fleet hatch]

@edgy holly - A little matrix ahha

[frigid pier]

i kinda feel like all the super stars may be replaced by AI stars.

so we may have a super star looking like alien

[frigid pier]

i could feel it's MATRIX

[vast remnant]

can u make a monkey doing asmr with another animal rq

i jst wanna see how it would look

[fleet hatch]

@vast remnant - Perhaps soon, lol

[vast remnant]

i mean like a monkey using another animal for asmr 😭

[fleet hatch]

And that is why i said Perhaps

Which Probably will not happen, Lol!

[rocky matrix]

https://youtu.be/dypBboiyslI?si=aWXa4hhEuG5yoEJ4

Official music video entirely made with text-2-video in Veo 3

[frigid pier]

Cell?

[fleet hatch]

Perhaps 😮

[arctic surge]

I am looking for a tool that can upscale 720p images generated by Whisk Animate, Veo 2, and Veo 3 to FHD or 4K after downloading them to a local PC.

[fleet hatch]

Handbrake

(IMO)

[obsidian latch]

I subscribe to the Ultra plan. Until about three days ago, I was able to generate videos with audio and lip synchronization. Today, I generated 20 videos and all of them came out without audio in the dialogues, producing mute videos. Is Google's server having any issues? Is there any news about this?

[simple lintel]

when did watermark became the thing?

[brittle talon]

I found that using flow didn't generate audio on any videos but using veo3 through the Gemini web page worked

[oblique summit]

https://youtu.be/DK5-VJtQ_GA - a short film created by Veo2 fast. Had to use it due to cost vs quality (easier to get better quality at lower cost)

[oblique summit]

This is so good how'd you get character consistency? I found it so hard

[violet mantle]

I'm back

@supple hull @zealous kindle 😈😜

[foggy heron]

Why isn’t flow available in my country?

[gleaming hatch]

https://myaccount.google.com/inactive?utm_source=sign_in_no_continue

[coarse citrus]

123

[lofty carbon]

i keep on getting "failed generation" on a very specific prompt. Any ideas why? The promt is the following:

A sleek, vintage black open-top limousine slowly cruises along a wide city street resembling a 1960s urban plaza. Sparse crowds stand quietly behind metal barricades, mostly watching with calm curiosity. The passenger in the backseat, a well-dressed man with slicked-back hair and a dark suit, smiles gently and waves at the onlookers. Old-fashioned street lamps line the road. The sky is clear with soft daylight. Ambient sounds include distant murmurs, gentle footsteps, soft rustling of flags, and a faint, subtle orchestral undertone to hint at quiet anticipation. The limousine has 3 rows, and 2 seats per row. The limousine is black. Each seat is occupied. The far back row is the passenger noted above, on the right side. Next to him is his wife. In the middle row is another man and wife, also dressed in suits. In the front row are two men, dressed in black suits. One of them is the driver, the other is sitting next to him.

i don't see anything that is negative or could be violating policies? i've set it to veo3, and a single output. thanks.

[sacred dawn]

Good morning, is there any update on when flow will be available to Gemini Pro subscribers via wqrkspace?

[wintry hinge]

Hello everyone, has anyone managed to fix in the scene builder the issue where the audio disappears when transitioning from the first video to the next scene, while creating a subsequent video, so that the voice and appearance remain intact?

[elfin willow]

Hey guys, Is there any way to achieve character consistency purely using veo3 for a 45 second direc to camera visual (i.e. the person speaking a 30-45 second script in selfie mode). If we do it in 5-8 second segments, I've not yet found a way for veo3 to produce the a continuous video, with sound, that is consistent with the previous generation. The "extend" feature forces fallback to veo 1/2 and the voice then disappears

Is this output currently possible with veo3?

[minor geode]

when is flow coming to india, any uodates

updates?

[jade crest]

Let’s wait until it’s released on third‑party platforms, but it still feels slow.

[bitter spire]

seems like scene editor works better today👍

still not showing frames while scrolling through the clip

it would be great to a last frame while extending the clip

[frigid forge]

So I just tried to do a veo 3 experimental with audio but their is no audio in the clip I got enough credits and used the right model maybe some bug or bad prompt but like if it is a bug RIP to them 100 credits ?

[tropic kestrel]

Optimize your prompt bro

[frigid forge]

I think my prompt was fine i got gemini to refine it before i sent it to flow pretty sure it may be a bug I got subtitles instead of any audio

this is part 2 I took the last frame from part 1 and then used that prompt you see and these are the resluts no audio at all

[tropic kestrel]

Do you have Ingredients on? If not, it won't remember what you just said and ment in the previous video and that might be the error (I had this too)

Can u try without? Should work 100%

[frigid forge]

maybe that is the issue, am just starting out did not use ingrediants just frames to video with the last frame of my other veo 3 generated video that had the audio and was perfect

[tropic kestrel]

Nice

[frigid forge]

will not let me 🫠 it changes to highest quality and drops the experimental audio trying to run the same prompt with the frames to see if I get the same results

so here is the scoop from the docs : Speech is only available in Text to Video mode

So folks must be using another tool to add the audio

[tropic kestrel]

No it's not true bro

If you make your prompt very strong and good

You can add audio in a Image to Video, I can show you

[frigid forge]

ok thanks I will give that a go and it should work with the frame to video feature ?

[sly trail]

Why is Ingredients to Video an Ultra feature? It seems like such an essential feature to generate video.

[tropic kestrel]

It is bro 100%

[sly trail]

I get charging based on credits to curb usage...but not paywalling an essential feature. And most people can't afford $250/month for this.

[distant pagoda]

Why is keep asking me to is not available in my country, my country is Poland, please help me.

[gilded badger]

Hey I'm having a problem signing in with the correct google account. Everytime I click Sign In it just automatically loads it to a wrong google account. anyone else have this issue?

[minor rune]

Because it's not available in the EU yet

[gilded badger]

yeah the EU hates freedom of expression so idk if it'll ever clear tehre

[minor rune]

It will be like veo 2, imagen 3 are available

Just a matter of time

[jade crest]

Veo3 Ultra has no watermark. Will the version on third-party platforms have a watermark?

[full peak]

Anyone else finding all generations are only a robotic synthy monotone voice and lipsync doesn't work at all? It doesn't matter what my prompt is, it could be a person sitting in a chair saying "Hello." The voice is a monotone synth computer voice and the lipsync is totally off.

[lavish dew]

Yeah, a lot of hit or miss, mostly miss. Perhaps try clarifying the tonal quality/intention of the actor performance as a part of the prompt? I was able to get a slighlty better read on some trying to layer that in.

[harsh quarry]

hey im signed in to flow and i have registerd and have gemini pro but for some reason i cannot use flow as its saying i have 0 credits, when i press upgrade it says i already have it...

[heady fulcrum]

Any advice how to generate consistent character?

[oblique spindle]

hey guys, i want to use flow to generate videos with veo 3, but the veo3 option is not available for me even though i'm a google ai pro(2tb) subscriber. How to fix it?

i logged in with the same account, and the veo3 is available in my country

[full peak]

Anyone also dealing with subtitles that shouldn't be there and often totally incorrect or spelled wrong?

[feral iron]

I'm trying to upscale the videos, but I'm getting an error. Anybody else?

[heady fulcrum]

Oh, veo 3 has image to video?

[slim beacon]

When will flow be available in the Netherlands? (ping by response please)

[analog sedge]

What model do you operate under?

[spring wave]

Is this true for Voice overs as well or only SFX?

I have been waiting for this update since drop

[copper remnant]

I'm having trouble getting audio at all. Just subscribed to Ultra today. No audio either in flow or Gemini. Setting does include highest quality (VEO 3). Anyone has any thoughts, pelase?

[shell mason]

Can you share an example prompt

[warm lotus]

can i use vpn to use it?

where country is available ?

where are you from

ok

im using vpn on usa to use it but i have the google ai suscription but i cant use it

[sour flower]

My video has no audio, despite me prompting it to do so. Any reason why this may have happened?

Ok, I figured it out.. I wasn't using the latest model with experimental audio.

[final dust]

Hello, is there a way i can give other people like editors access to my flow subscription?

[tiny parcel]

I would love to test Veo 3, but i live in France, there is no way to make it work here?

[minor rune]

👋

[minor rune]

Check your DM

[vapid stratus]

Who else is getting bad quality output like this with Veo 3? For some reason the contrast cranks all the way up!

[vale escarp]

Why does "Upscaling" constantly fail to generate from 720p to 1080p?? I successfully created two videos, but then they started to fail upscaling... Has anyone else experienced this issue?

[minor rune]

It works for me, I did it 1 hour ago

[vapid stratus]

Yes, I gave up on upscaling. I just use Topaz

[minor rune]

Maybe I'm lucky lol

[vale escarp]

Which Topaz app?

How bout Flow now hanging at "Loading....." system overload....lol

[vapid stratus]

Topaz AI video or Starlight. Depends

[vale escarp]

Yeah, Starlight demands are crazy... SPECS...

also "Peak" hours....

[vapid stratus]

You guys see the difference right? there's something really weird about Veo 3 and these over contrasted, over saturated outputs. Looks like a HDR video on a 400 nit phone.

[ripe mirage]

or why do I always lose the audio on the upscale

[real field]

Because they just upscale the video, they don't restripe the upscaled video with audio which is a separate process. Unfortunately you'll have to do that in post for now.

[ripe mirage]

I figured as much but don't love it.

[ripe mirage]

I understand and I am a IT person by trade. I understand well that everything is going to get exponentially better soon.

[vapid stratus]

I've tried simplifying the prompt, but it doesn't work. This only happens on Image to Vidoe on Veo 3 btw

[ripe mirage]

You don't think they will queue up some inovations for the next splashy release party?

[copper remnant]

can you share what that means? I have the latest model / advanced model VEO 3 selected.

[vapid stratus]

I mean, as you can see in my 2 example outputs that had the same exact prompts, the results are different. So I'm not sure it's a prompt adjusting issue.

But i'll try out those things you googled, thanks.

[ripe mirage]

What I really really will love is when we can refine the scenes in a collaborative way. Like I can sit and have a conversation with Gemini and suggest a number of changes. Like render me a rough cut of a whole long scene then go back through it like "At 15 seconds you have the bunny jump up and to the left but we need to really focus in on that and see the teeth of the wolves snapping at it. Show that he dodges the teeth and jumps off the wolf." type just talking through everything. That will be amazing.

I would say both.

and much more

[stray kernel]

For the basic flow account is there a storage limit on videos I create? I made about 15 2 days ago and see like 6 in my library now. Also I’m assuming I can’t reuse them in the middle of a project only launch new projects from them? Thanks in advance for any help!

[copper remnant]

simple one, for example, like this -- birds chirping and man saying hello good morning.

[grand smelt]

HI GUYS

[stray kernel]

Ty

[frigid pier]

I gave the same prompt to Veo 3 and Imagen 4.
The results are sooo different. Veo 3 has much higher quality and more cinematic or realistic

wish Imagen 4 was like Veo 3 :/

not sure why they are so different tho. I thought they would learn through the same dataset

[mystic forge]

I can't even get a person to sit down in the couch using frame to video. It just keep failing every time no matter what I put in 😩

[crisp basalt]

its in your library

[mental crane]

Hi I bought Google ai pro for one year but I don't got any credits?

[robust zenith]

Do credits get carried over month to month?

[vale escarp]

Geeze....Still LOADING..... like it's been 2 hours....same screen LOADING......

[tall kestrel]

Are there any motion designers in the group?

[mental crane]

No

[onyx fiber]

is the $125 plan on google AI like 125 for 3 months (125 each month) or 125 for 3 months (125 for the full 3 months) and then it jumps to the 200's

[muted cape]

125 per month. then it costs more per month after 3 months. (that's 375 total)

[dusk bear]

Can I be reimbursed credits for failed generations? Some of them immediately failed and used up credits.

[muted cape]

yes, they usually refund automatically. there's a transaction/usage log

[bitter ice]

is there any chance to use flow as free ?

[median needle]

Guys, how come I have a Google AI Pro plan and have 0 Ai credits in Flow?

[frigid pier]

god dang. just subscribed Ultra plan

hope I can do something with it ><

[glossy pelican]

Can someone help me get access to flow?
Im from germany.

[frigid pier]

what can i help u

Why does it say Veo 3 consume only 100 Credits?
because it's 150 credits per video in document.

[jade crest]

The price has been reduced, but the document hasn’t been rewritten yet.

[frigid pier]

ah thanks. interesting! and it's good price reduced!
thanks google team!

[frigid pier]

they must be very busy

[tiny patio]

"AI credits can be used across Flow and Whisk." Did you use them all up in Whisk?

[hot sage]

Why I bought Google AI pro but my Flow is not updated?

[hot sage]

same issue

[next rock]

Any news on when Flow when be available in Scandinavia?

[vale escarp]

Anyone get FLOW up and running???? still "LOADING..." over here......

[median needle]

no, but used Veo 3 in Gemini. May be that's the issue?

[vale escarp]

So FLOW is busted already?

[tiny patio]

Mine opens fine but i've gotta wait til the 7th for credit top up 😦

[civic bison]

Hi guys, does anyone know of a text-to-audio tool that delivers realistic results with emotional depth comparable to what we can see in Veo3? I mean things like shouting, a voice breaking from crying, whispering, etc. Thanks for your feedback!

[river rune]

How do you stop getting captions? Like its so random and annoying, im wasting generations over these broken captions

[celest plover]

Hi All, i just subscribed to Google AI Tools. I'm trying to use Flow but I am stuck at the " Say hello to VEO 3...." page and I cant click on anything.

[frigid pier]

bros

it's scary to consume credits lol

[median needle]

Guys, did anyone manage to generate and then edit a full YT video as if it was created by an influencer? Do you have any tips on how to build the process?

For example, how do I make sure the voice stays consistent and how to generate the 8sec videos to match them

[frigid pier]

Is Frame to Video availabel for Veo 3 now?

I just tried it but not sure if it works?

[stiff harness]

have you got sound as well?

[frigid pier]

I got sound yes

but the animation is weird

[stiff harness]

You don't 🙂

[median needle]

I saw people having it consistent 😅

[frigid pier]

how can we prevent subtitles it's annoying

[hearty drift]

🤓 Hey everyone, how’s it going? Can someone help me out, please? I’m trying to make this character speak a line in Brazilian Portuguese, but in my tests, I’ve had different results: one came out with audio in Japanese and subtitles, another came out without audio but with Portuguese subtitles, and another without audio and with Spanish subtitles. Can anyone shed some light on what might be wrong with my prompt? I’d like to generate it without subtitles and have the line spoken in Portuguese.

PROMPT:
A push-in slow zoom shot begins mid-distance, slowly advancing toward a stylized demonic cactus character standing in a terra-cotta pot, set on scorched black volcanic terrain. The ground is cracked with glowing lava veins, and embers drift lazily in the hot air. The atmosphere is smoky, dense with ash, and the horizon fades into a foggy, bluish haze.
The cactus has a dark green, sharply ridged surface with charred edges and sinister textures. It is completely engulfed in photorealistic fire — roaring orange flames lick upward, dancing violently around its two matte black curved horns.
Its glowing orange eyes lock directly onto the camera, flickering with malice. It makes a wicked, mischievous expression, eyebrows slanted in a devilish arch, spiny mouth curled into a smug smirk. Smoke coils into elegant spirals above its head.
As the camera gets closer, the character stares straight into the lens and says in Brazilian Portuguese, with a gravelly, teasing voice:
"Bora queimar um verdinho?"

Immediately, it erupts into a deep, villainous laugh, echoing across the volcanic plain. The flames surge higher around it as it laughs.

Lighting is intense and infernal — the fire casts strong, dynamic shadows, while the lava below adds subtle orange highlights to the pot and ground. A rim light outlines the cactus and horns, giving it a glowing silhouette against the dusky background.

[frigid pier]

bro also put this if u wanna remove subtitle.

The characters never look into the camera. Do not include any captions, subtitles, or on-screen text.

[tiny patio]

And don't use " " for any speaking. Try something like The woman walks to the counter and says Can I have a large Cappuccino. Although I've only had 10 attempts that has a better chance of no subtitles.

[frigid pier]

does it work always

[frigid pier]

I also found that Veo 3 is exceptional for selfie style videos but if u try the same prompt and just change the shot style then the quality goes down hard lol

[modern sail]

For Meta stock investors , fun prompt had Tonto making the horse sound
https://www.zavamed.com/?srsltid=AfmBOoqiWhaz1FQXTK8heTGrGAS0al_iMoftheYajFsVaRw555JXTshPmpt Veo 3

[umbral root]

Guys, am i the only one not to have sound with Veo3 using Jump tp feature. The 1st scen is ok, and whe ni do extend or jump to, the character doesn t speak anymore. Just move his mouth but no sound

[frigid pier]

bro wthhh i am amazed by this veo 3 sometimes. whoa

[frigid pier]

I this this is a known issue

[broken cairn]

how do you get audio in flow?

Right now it's just vid

[frigid pier]

just make a video then ai understand what kind of sound it should put there

[frigid pier]

did u choose "highest quality"?

[hearty drift]

thanks bro! but, why never look into the camera? i need this to all prompts? even if i want the character looking to the camera?

[frigid pier]

nnah then dont use it

btw i just tested that prompt and it doesnt work many times. so dont use it maybe

it still includes the text so i guess Veo 3 ai is not understanding this prompt properly

[hearty drift]

I read in some places that using negatives like “don’t do this” or “avoid that” doesn’t work very well with AIs. Could that be why my prompt isn’t working?

[frigid pier]

yea this is my idea too.

[frigid pier]

I think these generative AIs just put anything if they were forced to learn it before

so i cannot really say these are intelligent tbh. it's just weird.
it's not even smarter than kids.
cuz kids wouldn't put subtitles if someone said no.
maybe AI is too hyped

already consumed 2k credits omg

have only 10k credits for now

[frigid pier]

it seems like the next year is the end of the trad film industry.
Veo 3 is already showing insane quality :0

[warm burrow]

how do i fix flow not detecting my subscription?

[frigid pier]

whats your plan

[warm burrow]

the month

[frigid pier]

i mean pro or ultra

[warm burrow]

free trial, but i paid

[warm burrow]

pro

[frigid pier]

hmm ok

whtat happens when u access flow?

[warm burrow]

^^

i am subscribed tho

[frigid pier]

maybe they do not allow pro plan users anymore? idk

[warm burrow]

it says tho

"Flow is available exclusively for Google AI Pro and Ultra subscribers. Ultra subscribers also get first access to our latest models and features like Ingredients to Video. Learn more here."

[spice jewel]

Task type (Text to Video, Frames to Video, Ingredients to Video) vs. Model selection

So many times so far I’ve wasted credits because I was hoping to get video with audio but got video without audio, either because the generation went through Veo 2 (instead of Veo 3), or because the given task type didn’t support the given Model selection.

Remember the model choice!

First, the app keeps forgetting my last model choice, it silently reverts to "Fast". But I get Veo 2 generations anywhere, elsewhere. I’m not interested in wasting credits on unintended Veo 2 generations!

Model selector should be explicit

If a given task type does not support AUDIO output, then the Model selector should have 4 choices:

• Fast (Veo 2)
• Normal (Veo 2)
• Quality (Veo 3)
• Quality with Audio (Veo 3)

If I choose a task type that does not support audio, then the last choice shouldn’t be selectable. Really, make it explicit!

Stop with the "experimental, preview, beta" lingo

Some management within Google seems to pressure designers to put the "Preview", "Beta", "Experimental" exculpatory crap. Please talk to your management and ask them not to pressure you, and REMOVE ALL THAT. Yes, we all know that it’s ALL "experimental / preview / beta". Just add a general disclaimer at the bottom and be done with it. It’s not necessary to have audio output labeled "experimental". I mean, that suggests that the video output isn’t experimental.

Stop with cheap marketing lingo

Also, it’s not really needed to "sell" the quality of Veo 3 to us. "Quality" and "Highest Quality"? That sounds like some 2nd-grade marketing talk. It’s unworthy of Google. Seriously, "Normal" and "Quality" would suffice 🙂

Settings button

The Settings icon that opens the Outputs + Models dropdown really should become some status indicator. Right now I have to click it EVERY TIME to make sure that the right amount of outputs (mostly 1) and the right Model (ALWAYS Veo 3 for me) is selected.

[spice jewel]

I’ve posted these as separate items in https://discord.com/channels/1172568727942860810/1319424187730755636

[twilit linden]

does anyone else have issue getting sound from flow generated videos? I am on ultra subscription. The videos generate fine but none of them have sound even if I download the videos and play them locally

[grizzled prairie]

Same issue. A lot of wasted credits 💔 I upgraded to Ultra thinking that a plan upgrade will fix it. But nope!

[twilit linden]

yeah I'm pretty dissapointed in my ultra plan atm... flow is almost useless.. the video generation through veo3 from https://gemini.google.com/ works perfectly fine... generating videos with the same prompt in https://labs.google/fx/tools/flow no sound

[grizzled prairie]

Ah you're in a better position. Mine still won't generate audio in Gemini. Really frustrating.

[humble tusk]

anyone have 100% success using 2 camera angles in one scene? i've gotten it like 1 out of 20 tries

my prompt would go something like "over the shoulder view of computer monitor then cut to close up of man's face"

[twilit linden]

oh actually I just fixed my problem, if anyone else is having sound issues with flow you may be using fast generation which uses veo2. You need to change the settings (kind of hidden) for the text-video prompt. note it'll cost more credits (veo2 is 10 credits... veo 3 is 100)

[humble tusk]

oh it has to be on veo 3? i swear i remember using an ingrediant and thats veo 2

i really dont use veo 3 because it cant do ingrediants.

[grizzled prairie]

This still won't work for me unfortunately 🥹

[twilit linden]

odd, other sound works for you?

[ocean flume]

Guys, can someone help me, I paid this but this appears 🤡🤡

[violet hull]

hello , can someone help me create an introduction for my podcast with veo 3 ?

[twilit linden]

which subscription did you pay for

[twilit linden]

what kind of introduction are you looking for? If you just want an 8 second clip I can get you something from gemini

[upper niche]

Does anyone know if and when Flow will be available in EU (NL Specifically)?

Thanks!

[violet hull]

I just send you a message

[main saddle]

while we don't have immediate news to share, we'll post in #flow-announcements・📣 when new countries are added! for a full list of countries, see our Flow FAQ doc: https://labs.google/fx/tools/flow/faq

[turbid ravine]

Hey folks, I was wondering. In most cases when I create a prompt, the system is not following the prompt, when it comes to sounds. Instead there is no sound at all

Here is the prompt I was aiming for. It has nothing to do with North Korea. Also if I replace it with south Korea, still no sound. What am I doing wrong?

North Korean news broadcast: female anchor in traditional hanbok, solemn expression, announcing in Korean: "The biggest fan of our great leader celebrates his birthday." Cut to Pjöngyan Elementary School: uniformed children presenting handmade treats to a visitor, symbolizing friendship. Bright lighting, patriotic background imagery, consistent with North Korean state TV.

Sound: Anchor's voice is clear and authoritative. Transition to soft, uplifting instrumental music with subtle sounds of children in the school scene.

[livid egret]

This isn't helping me either - I'm having a LOT of trouble getting sound output even at highest quality

Sound output seems to be VERY broken today - I had it working most of the day yesterday and now pretty much no generation at high quality will output sound. I'm marking them all as inaccurate and hoping I get my credits back

I feel like credits should work much differently in this early period - we only get so many, and we're basically paying to be alpha testers in many ways

The tool is INCREDIBLE when it works, but when it outputs no sound, it's wasted time and credits for everyone

[craggy jewel]

What could be best for audio

[ebon robin]

Is it possible to update a prompt rather than reusing a prompt? I like the output I got, but have a slight tweak I need to make to it. And when I do the reuse option, the output is very different

Got it. I was using "ingredients to video." Thanks!

[craggy jewel]

I am new to it... please help how can i generate video here

[livid egret]

You can't do it here - you have to use gemini or flow to generate your video from prompts

[craggy jewel]

Ohh I already have gemini subscription but in my country(India) only veo 2 is lauched. I thought i could use veo 3 from here ..Thanks boss for saving my time

[umbral root]

Yes Be sure to select veo3 all the time, because it might switch to veo2 easily, but anyway, if i use jump to feature to continue my scene the news scene has no sound while the 1st one is ok.

[azure nymph]

yes

i just sent an example on #flow-videos・📹

[kindred terrace]

Is scenebuilder down in Flow.? I cannot access any of my videos. It was working fine this morning.

[spice jewel]

What exactly is the difference between "in production" and "public beta" when it comes to Flow?

[glad cedar]

Is upscaling down for anyone right now? It hasn't been working for the last 30 minutes for me.

[crisp basalt]

Remember back when VideoFX had those "early-access," or "preview," or straight-up "work in progress" tags slapped on it? 😅
😆

[spice jewel]

Well, "Flow" is on "labs.google", and all sections there are called "experiments". And within it is Flow, and within it "Highest Quality" video generation with "experimental audio". What's the point of labeling audio support as experimental, when the whole tool is an "experiment" within the "labs"?

[crisp basalt]

what

adam i think you replying to wrong person

[lavish dew]

Are there any updates for "Couldn't generate video. Try again later." for 1080p downloads?

[azure nymph]

I tried to generate a video and then an error occurred and the video was not generated...but my 100 credits was wasted...

This is very annoying

[livid egret]

It's the most annoying thing about the current state of the app. I get that we're getting a discount, but a discount isn't worth much if we're generating junk.

[crisp basalt]

It’s in your library if anything

[spice jewel]

You mean, for cutting-edge AI video generation models? Interesting. I wasn't aware that this is an "established" category.

[crisp basalt]

Yes that is true

https://tenor.com/view/chaetzukimbap-nereris-chaeyoung-twice-gif-26480086

[spice jewel]

Right, but "Flow" is an "experiment" within "Google Labs". So audio is "experimental" within an experiment, and Veo 3 is generally speaking not experimental, and neither is Veo 2.

Maybe.

But: is it relevant to the user?

[golden cobalt]

Hey everyone, random but probably more people asking the same question, I have google pro but I cant use flow?

Any fix for this?

I'm a Google AI subscriber but it still is telling me I am not one?

[spice jewel]

Within the context of a user interface of a "Google Labs experiment", is it relevant to the user that the "audio" is "experimental" while the other things are not? Does the user of the Google Labs experiment namd Flow care about the "development cycle"? I highly doubt it. The "experimental" tag may be relevant for the Google Vertex AI, and experimental features should be indicated as such there. But in Flow? It’s just visual clutter.

The fact that it's true does not make it relevant to the point of promoting it to a 1st-level on-screen information.

You're off on a tangent all by yourself.

I could say the same about yourself, but blindly diminishing one’s social standing is a loser’s eristic trick.

"Stop with the "experimental, preview, beta" lingo, Some management within Google seems to pressure designers to put the "Preview", "Beta", "Experimental" exculpatory crap. Please talk to your management and ask them not to pressure you, and REMOVE ALL THAT."

Partially disagree ^here^. Veo 2 is in production, but they should use industry standard terminology, not new internal labals. Veo 3 should be 'Public Beta', not experimental, which doesn't really mean anything.

You replied to my message, which was just a new message here. You made the point that Veo 2 is "in production" and Veo 3 is "public beta", but currently neither is labeled as such but instead "audio" is labeled "experimental".

My point was that for a user of Flow, the development cycle of any particular feature is irrelevant, because the whole Flow app is a Google Labs "experiment". So "experimental audio" is just visual clutter that increases the cognitive complexity. It doesn't contribute valuable detail to the user in this context. Audio in Veo 3 isn't deterministic, Veo 3 isn't deterministic, and Veo 2 isn't either.

[livid egret]

I feel like I'm spending credits even reading all this Adam

maybe chill out and stop overanalyzing nomenclature

@spice jewel most of us are here to troubleshoot why the model is doing oddball things and help each other get better at results - you're just honk-honking around about the word experimental, it's just noise

it's been a loooooong time since I've encountered the self-important linguistic blowhard character in a chat room, so I guess thank you for reminding me you all still exist

[spice jewel]

Sorry, I didn’t realize I’ve dropped by an exclusive club of super-pros. Hey, you’re in a Google Discord and you have seniority. Have fun, then 🙂

[spice jewel]

It’s not been long since I’ve encountered people who in the second message say to you "maybe chill out and stop". Yeah, there’s plenty of them around. Cheers mate.

[crisp basalt]

Ooh, now the upscaled video just auto-downloads instead of showing a download button.

[long steeple]

How do I check how many credis I have? And how do I know how much the VEO 3 generations cost?

[spice jewel]

All right, I clearly see that you have very important things to do here. Anyhow, my colleagues at Google just sent me the contact to the Flow PM so I’ll write to the team directly about the UX.

Have a good evening.

[livid egret]

I'm sure they'll be overjoyed to get that ten page email

[spice jewel]

If your emails have "pages" then sure you’re printing them out. 🙂 Sorry, didn’t realize this was that kind of forum. 😄

[crisp basalt]

can you ask your colleague at google when are they releasing pixel 20 ultra

can u also ask your colleague to give out free ultra plan

can u also ask your colleague, when is veo 5 coming out

[spice jewel]

After Veo 4.

[crisp basalt]

i see

[livid egret]

""Is" is only actually "is" when the circuitous route that brings me to understand the - and by the I mean YOUR - meaning of "is experimental" when it is either neither an experiment nor a hypothesis, therefore..."

[lime hull]

A young Algerian man sitting on a folding chair on the beach of Oran, Algeria, staring indifferently at the horizon. In front of him, a beautiful young woman in stylish summer clothes is talking to him with enthusiasm, but he is uninterested. The scene is set during sunset with warm Mediterranean colors, gentle waves, and the city of Oran visible in the background — a relaxed Algerian beach atmosphere with subtle urban elements."

He speaks men and woman in Modern Standard Arabic Algeria(الفصحى) saying:
: "طريقي طويله بصح تكمل بالهنا"
"واه نيشان طريقك طويله بصح تكمل باللي قلنا"
جبديها جبديها هههههه‍

[oblique summit]

https://youtu.be/HMFhm4rvcEc?si=wlr9Vllir8-ZfopT - can't believe we got AI GTA 6 before GTA 6

[ripe mirage]

Is it not possible to use veo3 and the ingredients. I'm trying to do a talk show format and it's just going horrible. 1k credits down and nothing much usable. I need the consistency of the characters for the scene and other things to work. Questioning why I upgraded to Ultra. If I wanted to go entirely off text prompting I would have stuck with pro.

[humble tusk]

no not possible. veo 3 is useless.

i use veo 2

[wooden helm]

@main saddle how many flow generations we get every month, in pro plan?

Can we check our leftover credits?

Thanks 🙏

[waxen tulip]

hey

Does anyone know where there is a promt tutorial for flow ?

[ripe mirage]

which seems pretty useless if you want to have a actual dialog between two people.

[humble tusk]

yea, i work around its limitation. my recent video is a voice over narrative in the style of a BBC documentary.

[surreal rain]

anyone knows when flow is available in europe?

[shell mason]

@surreal rain they haven't published any date for this yet

[mortal stone]

Hi guys ! anyone know why in heaven "extend" option isn't available for VEO3 generation in Pro version ? was the case before wasn't it ?

[mortal stone]

is "extend" option with veo3 available in Ultra ?

[ripe mirage]

yes but no audio

[lavish dew]

That has no 1080 option, only 720, from what I see.

[crisp basalt]

Oh u don’t have it? Too bad . jk jk

I should’ve clarified what I meant

[digital yoke]

Any time frame when it possible be available in Sweden?

I paying for pro but dont really have any feature that should be included? Can someone look in to this?

[sick coral]

Flow wont allow me to scene jump. Drs office in scene1, jump to scene2 in park... stays in the office with the actions from teh park. Ive waste 200 credits this morning trying to do this. Or is there a way to carry the chars to a new sceen?

[dusky stone]

So when will it be available in Spain ?

[digital yoke]

Pay 26 euro for something that is not included for the Swedish market :/ please Google show some love

[azure nymph]

I use all my attempts with the VEO 3 in the Gemini app…and now I have to wait almost 2 weeks to use it again lmao

That’s crazy 😂

Is it normal?

[earnest oak]

Hey guys, I really need some help around here

[frigid pier]

for what bro

[earnest oak]

Do you have any access to Veo 3? I paid for Gemini and I dont get why cant I access it, or how 🫠
And it's really important because my grade in some university project relies on it

[narrow knoll]

There is a free trial you can do for veo 3

[earnest oak]

how and where?

[narrow knoll]

Search up google veo 3

[jade saddle]

so how do we get sound for it

I just paid for it tried to create a video and ofc no sound

[verbal raven]

did they fix no audio in scene builder for subsequent scenes yet?

[high jolt]

Is not yet in spain?

[earnest oak]

When I do it leads me back to Gemini

[narrow knoll]

Maybe it’s not available in your country

[tiny patio]

FAQ should have current list of countries. #flow-general・🌊 message

[azure nymph]

Where? Gemini app/web? Flow?

[frigid pier]

hello guys

[jade saddle]

flow is good but can it get it right I just bought it went through like 40 promots already and it fails somewher in the video

so annoying

[frigid pier]

it seems like Flow is only good with character without no objects characters holding or stuff and static setup

but yes still amazing

[marble hill]

Has anyone had any luck Flow accepting your uploaded image. It keeps saying that it might violate their policies. An example is this image

[frigid pier]

I think maybe the ai thought it's blood from human body

or like gore scene

[marble hill]

It is literally not accepting anything I upload. Just normal images of normal humans walking etc. All receive the same message that it might violate

[cyan hamlet]

zu teuer

[frigid pier]

upload another example here let me see

[frigid pier]

maybe try it with the image that is just pure normal shot

like walking on the street in a city

[warm flint]

Veo 4 when

[marble hill]

I guess we cannot be too artistic with image to video

[frigid pier]

bro u joking lmfaoo

[marble hill]

I think it is a restriction as I am based in the UK. When I use a VPN and set USA it accepts the image

[frigid pier]

that's very inteesting

[jade crest]

Upscaling with Flow can duplicate watermarks.😭

[teal glen]

im pretty sure people already do this in ComfyUI but he is claiming it was made using Veo3 which i doubt it

[warped arrow]

are they fixing the degradation when extending a scene/clip?

[faint wyvern]

Veo3 masterpiece — truly flawless, as always. But come on, be honest… where are the glorious fails? Veo3’s real talent is in those disaster clips no one ever posts. Gotta love how social media is just a highlight reel of perfection, right?

[faint wyvern]

Perfect — just make sure to label the clips properly, so we don’t confuse them with your ‘masterpieces’ again.

[mystic forge]

Yea veo 3 flow no accepting no images in the frame to video.

[kindred terrace]

Hey Google, what’s with the un requested gibberish subtitles in Veo3? I’ve tried many forms of negative prompt but they just randomly appear. It ruins an otherwise useful render.

[warped arrow]

do you ever put double quotes? I think putting double quotes on your prompt will make flow assume you want that as caption

[kindred terrace]

I do sometimes but that isn’t the caption. The unwanted captions are usually nonsense; like mangled English.

[sick coral]

I want to keep the same characters, but move them into a new location. Thats how "jump" is explained. But instead of allowing me to introduce new elements, include a new backdrop, while keeping the characters, it doesnt "jump", it extends...

[onyx moon]

Extend doesn't even use the Veo3 model. It switches to Veo2 immediately. Why is that?

[fresh lagoon]

How would it be possible for me to use Veo3 if i'm in a country that it's not released yet?

[onyx moon]

It doesn't even use Veo3 model for extending or jumping. It defaults to Veo2. I'm an ultra subscriber.

[warped arrow]

they probably still doesn't know how to copy and continue the audio from one scene to another.... I think most of the videos you see in youtube with a continuous audio are dubbed by a different tool

[pallid kite]

https://www.youtube.com/watch?v=DY5vnaCx_KE&ab_channel=uisato

[elder adder]

I thought it was a plan issue! I'm pro and I was thinking to upgrade to Ultra to be able to do this, so it still switches back to Veo2!! thanks for the tip

[long steeple]

when a generation gets stuck at 99 and doesn't generate do i get my credits back

[dense gulch]

does anybody recognised that now on all Veo3 generated materials there is very small VEO sign in right down corner?

[abstract sage]

To generate a video just put the prompt here???

[onyx moon]

Where might these docs be location? Link please. This is the extent of the documentation I could find which isn't helpful at all. https://cloud.google.com/vertex-ai/generative-ai/docs/models/veo/3-0-generate-preview

[spice jewel]

I crop them out, use Ray 2 or Ray 2 Flash at https://dream-machine.lumalabs.ai/edit/new to reframe (inpaint) what I’ve cropped. And because that process reduces the quality a bit, I composite the the original (but cropped) video on top of the Ray output, with a 32px-tall smooth transition.

[kindred terrace]

But currently only possible with Veo2 yeah?

[spice jewel]

As I’ve suggested in the other thread:

• Don’t use Scenebuilder (jump or extend) at this point. You won’t get audio, and most likely the outputs are Veo 2, not Veo3.
• Write your prompt. If you have a character, give that character a name. Make the prompt very specific and detailed.
• Make a new Flow project and generate a clip there.
• Drop the clip into Gemini 2.5 Pro or Gemini 2.5 Flash, and ask it to describe it in detail.
• Incorporate any details into the next prompt. Veo can consume very detailed and elaborate prompts, and if you’re careful, you’ll get consistency TO SOME EXTENT.
• You can always use the "3 striking features" trick for characters: think of 3 characteristics that are very unique and poignant, non-generic, for the character.

One thing that I’ve noticed but I cannot fully confirm is that within one session (one chat sesson on gemini.google.com or in one Flow project), if you create new clips but you reference similar location descriptions, similar character descriptions, and also the character NAME that you’ve given — there is increased chance to get consistency.

I’ve never found Veo 2 outputs useful for anything. Veo 3 outputs however are instantly useful. I most certainly wouldn’t be able to do something like #flow-videos・📹 message in Veo 2 with “lip sync AI”.

[barren verge]

Probably been asked this question 1000 times but is there an issue with scene builder in Flow? It was there yesterday but i cant find the option today.

[somber moth]

hey, im new to this and trying to make videos, would u know if thers a way to make the vertical rather than horizontal?

[minor rune]

Anyone seeing the increased credits as announced on the amount of pro or ultra subscription credits?

[shy depot]

A 1000 for me.

[minor rune]

It was that already, they said they will increase, nothing changed 😕

Damn every time he says something it's not what it means lol

[barren verge]

The option to build the scene in the top left corner is not there at all

[crisp basalt]

are u on small screen or something

[barren verge]

Im on mobile, could that be the reason?

[crisp basalt]

yeah thats the reason why

[barren verge]

I remember the other day i used it both on mac and on iphone but maybe im wrong

[crisp basalt]

is it possible for u to zoom out

[barren verge]

Thats it, without croping

[crisp basalt]

yeah thats still not enough

do it in landscape orientation

[barren verge]

I dont think this is of much help tbh😂😂

[crisp basalt]

well

too bad , no choice but to not use mobile phone then

[barren verge]

Yep, thank you for the info

[golden cobalt]

Guys, am I the only one who can not use Flow while owning Gemini Pro?

[pulsar blade]

Me too

[golden cobalt]

idk seems weird, tried everything but still no fix

flow seems to be way more detailed than on gemini using veo 3

Like you can actually use the same character on flow etc and also use images

[pulsar blade]

Yeah I tried everything too, came here expecting help on that

[full flicker]

Please keep the chat in English. Thank you!

[naive pine]

i see we still default to veo 2 😥

[versed hound]

How do you download your video or share it ?

[white forge]

How simply this prompt

[mystic forge]

Wow what's up with the sound on veo 3 flow. It's worst. Now they have voices on there with just subtitles an when the music plays it have the person lip singing what's playing the music. Then when it's no music it don't no sound at all and voice. It's getting crazy 🤦‍♂️

[honest copper]

whoevers filling instagram up with the yeti and bigfoot videos, yall are great

[frigid pier]

It's kind of annoying that Veo 3 kepp generating Video game scenes even tho I didn't say it

[honest copper]

it's in "outdoor boys" format, one ig is outdoor_yetis, another ig is deepjungleai

[true socket]

How to use Flow in India?

[real obsidian]

I've seen so many youtube channels doing this 💀 didn't know big foot was so popular

[honest copper]

Idk who's creating it but it's filling reels lmao

[jade crest]

I noticed that Ultra also has a watermark.

[olive kayak]

I'm trying to figure out how to make a video like his, where the character holds a selfie stick and the POV looks natural. Also, why does my lip sync always feel off? Any tips?

Here’s a video I made

😭

[analog sedge]

Just ask gemini

[olive kayak]

I’ve solved this problem.

[wispy dagger]

Hey all, forgive me if this is the wrong place to ask but I'm struggling to wrap my head around the differences between Flow and Video creation via the Google Vertex AI.

• Both allow the use of Veo3
• Both allow creation of video from prompt or image
• However one is UI based and the is more API/console/SDK

Where it gets confusing for me is the feature mismatch?

Vertex allows creation of 9:16 aspect ratio videos with Veo2 while Flow does not.
Vertex doesn't seem to have ingredient based creation?
I also am not sure how to enable continuity using Vertex?

Am I correct in understanding that there isn't feature parity? i.e. I wouldn't be able to get the same functionality out of Vertex as out of Flow?

[frigid pier]

Veo 3 is the best for selfie style videos.

[frigid pier]

what the hell! Veo2 frame to video now has sound too!!! YAY

Veo 2 Frame to Video is insane!! :0

[halcyon trellis]

Really?

[inland rampart]

I'm getting this problem for some reason even though I bought a Google One account

[errant inlet]

How can it be used? Please tell me I really need it for my work.

[bitter spire]

why flow keeps eating my credits while failing to generate a video?

[bitter spire]

thats a big LIE

[frigid pier]

it's real bro

[frigid pier]

idk just put the image and make the video. then it auto generates audio

[bitter spire]

its not possible technically

[frigid pier]

but it worked for me!

[bitter spire]

prove it.show the video. and a model u used for it

[fathom hatch]

In wht all countries Flow is available?

[long sedge]

A highly realistic video shows international football star Cristiano Ronaldo standing in front of the historic Citadel of Aleppo in Syria. The scene is vivid and lifelike, with detailed lighting, authentic crowd interaction, and natural movement. The camera opens with a wide panoramic shot of the citadel, then slowly zooms in on Ronaldo, who is dressed casually in modern clothes.

Ronaldo is seen speaking fluently in the local Aleppine dialect, smiling and engaging warmly with a group of local people. He uses common phrases like:
– “Shlonak ya zalame?” (“How are you, man?”)
– “Allah ykhalleena hal balad al tayyeb!” (“May God preserve this good country!”)
– “Ya ‘ayni ‘al-akel al-Halabi, ma fi mitlo!” (“Oh man, Aleppine food is unbeatable!”)

His facial expressions are natural and relaxed as he laughs and chats, showing genuine emotion. Behind him, the citadel appears in full historic detail — weathered stones, arched doorways, and bustling street life around. Ambient sounds include street vendors, distant children playing, and locals chatting in the background, adding a rich layer of authenticity.

Key attributes:

Style: Ultra-realistic, cinematic

Location: Citadel of Aleppo, Syria

Language: Syrian Arabic (Aleppine dialect)

Mood: Friendly, casual, culturally rich

Details: Natural facial expressions, accurate clothing, live ambient sounds, interaction with locals

[real obsidian]

fake news? I just tried this and doesn't work

[mortal stone]

troll spotted

[gaunt mesa]

@golden cobalt Be aware you cannot use the same character (Ingredients) on the Pro account. You need Ultra to use Ingredients

[turbid quartz]

GUYS I AM STUCK INTO A NEW STATE OF CREATION PAGE ON FLOW

WHY?

[shell mason]

I'm an ultra subscriber and since today I'm getting watermarks on most of my videos, what's going on?

[fathom hatch]

you need more AI credits to complete this request. flow

What does this mean? I am already on Gemini PRO Plan

[gaunt mesa]

@fathom hatch You only get 1000 credits per billing cycle. You most likely used them. It's 100 credits per output for Veo 2 Quality and Veo 3

That's only 10 videos

Veo 2 Fast will get you 100 videos per billing cycle

[fathom hatch]

Thanks @gaunt mesa i want to create a 30-45 second launch video with audio via Veo3 how can I do that?

[real obsidian]

yeah this seems like a bug, saw people mentioning it on x

[shell mason]

I'm also getting more and more videos that have no sound whatsoever

[gaunt mesa]

@fathom hatch You just need to select VEO 3 in Flow when you generate the clips that you need. The problem is, you only have 10 outputs per billing cycle. You cannot just buy extra credits on the Pro plan. You can only top off credits on the Ultra Plan. So your three options are: 1) Make the 30-45 second video just using 10 clips that you get from VEO 3 for your billing month. 2) Generate 10 this billing month and then wait until next month to generate more. Hopefully out of 20 clips, you have your 30-45 second video, although it will take a while to accumulate. Or 3) Upgrade to Ultra. This gives you 125 Veo 3 clips per month, AND if you need more than that, you can buy additional credits

[frigid pier]

bro this Veo 3 is crazy. I'm amazed whenever i try it. i believe Veo 4 will replace actors and stuff omg

[kindred terrace]

Not at all happy with Veo today. So many inaccurate renders. I think I have flagged them all. Not just stylistic issues but physical nonsense. Objects suddenly appearing or morphing.

Why am I getting billed for failed attempts?!

[kindred terrace]

Same. I was creating corporate scene in a Doha office and it kept putting an ‘Al Jazeera’ style watermark in the top right corner.
Unacceptable Google.

[shell mason]

that is really strange. I just got a Veo watermark at the bottom right. And gibberish subtitles, despite all prompting efforts

[icy maple]

Yea that seems to be a problem rn the subtitles at least

[subtle needle]

There is no sound in the videos I create in Flow. Is there anyone else having the same problem?

[real obsidian]

this just happened to me and I realized they randomly changed me from veo3 back to veo2

[jade crest]

I produced six videos at once for 600 credits, but after finishing I discovered that this VEO3 was severely distorted.

[forest raptor]

can't tell if y'all are rich or just have poor spending habits

[subtle needle]

No, I checked multiple times to make sure it wasn't.

[proven musk]

Hello! So like if I am on budget, what ai video gen tool should I use?

[real obsidian]

a little bit of both

[real obsidian]

runway i think is the cheapest

[forest raptor]

Vidu

[proven musk]

Do you think the google ai suite is worth it? I am looking into different subscription options

I have never heard of Vidu I am looking into runway as well

[forest raptor]

Runaway is very meh

Hailuo is good too

[proven musk]

Interesting, Vidu has me interested I just looked it up

[forest raptor]

Vidu is good if you want reference-to-video

[jade crest]

I found that VEO2 Fast is better than VEO2 Quality at controlling object deformation.

[proven musk]

That's cool, what do you guys use your ai videos for?

Are ya'll just trying things out for fun or are you implementing them into work?

[jade crest]

Runway reference-to-image is currently the best in the world.

[forest raptor]

nah

Vidu is best

[forest raptor]

I just express my creativity. It's pretty much a hobby. It's like telling stories to myself XP

[jaunty plume]

Question I might be really slow but I was breaking down the pricing of using veo 3 because I have seen so many of these ai generated vids on Instagram now, and I’d love to just make my own random 30 second stories or soemthing, but 125 for a month for 125 5-10 second vid generations seems hard to justify,

Say I wanted to make 30 second vids if I didn’t mess up 1 prompt I’d get roughly 25-40 (depending on clip length) 30 second videos. For 125 bucks.

I hope my math is way off because I’d love to start using this tool

[sharp pecan]

I totally agree with Google, it has to help us

[real obsidian]

$1 = 100 credits = 8 seconds. 30 second videos are $3.75. veo3 isn't for if you want to be cheap

[frigid forge]

I feel like I'm getting cheated here quite a few veo3 outputs without any audio with top notch polished AI refined prompts. I want ma credits back GOOGLE 🤣

[marsh olive]

Me waiting for it to come to Europe

[onyx moon]

Did you ever find a solution to this? I've run into the same problem and I had to export each clip individually and stitch them in CapCut. Super annoying.

[misty patio]

Someone knows when FLOW will be available in Europe? 🙂

[marsh olive]

Sadly no

[misty patio]

Pity, so i need to call my friends from USA to get access 😂

[spice junco]

I've been getting the same thing. Will we be getting credits back on those generations?

[frigid forge]

That's what's the new normal is ...

[open knoll]

Hello guys!! Woow official Google has a Discord server 👀

[onyx moon]

Does anyone know why after creating multiple clips in Scene builder using Extend - we cannot then export the multi-scene as a single clip?

[frigid sapphire]

Guys, I need your help. I'm making videos for Instagram, and I need Veo 3 to generate a vertical video for me.
And I'm using this prompt at the beginning (A vertical cinematic video (9:16)), and it generated perfect vertical videos, but it also generates horizontal videos. Is there anything I can add to it to generate 100% vertical videos?

[velvet flume]

Hi guys does anyone ever have any issues using the ingredients or frames part. Whenever I upload an image it says it might violate their policies.

[frigid sapphire]

the vertical videos it generates are videos laid down like this

but at the same time it generates videos like this

[velvet flume]

That's my msg now gonna be lost lol

78k people and no one able to offer help 😦

[frigid sapphire]

what is your difficulty?

[velvet flume]

This

For your one you need something like capcut and to rotate it on there and then export id recommend

[frigid sapphire]

sorry, I don't have Google AI Ultra, but try to get another framing of the photo, if it's an image generation, try to generate something similar

[frigid sapphire]

I know that, I'm generating the vertical video lying down and rotating it in the editing program, the problem is that it doesn't generate the videos lying down all the time, it also generates horizontal videos instead of vertical ones.

[frigid sapphire]

Was I able to help you with anything?

[velvet flume]

Unfortunately no coz I've tried multiple images.

[velvet flume]

Yeah that seems odd hopefully someone will help us both

[dire maple]

I've tried upload the same image using Iphone and Mac, got this messagem with the Iphone and worked fine with mac, same image.

[jovial vessel]

you can't directly make vertical clips from veo models. It's built for horizontal views.

[rapid crow]

All good. This is an old message

[halcyon vector]

any news on ingredients features for veo 3?

[velvet flume]

Mad

But in scene builder if you use veo 3 you can't use that on extend so the following clip won't have any sound

[crisp basalt]

Was there a new update that I missed

Also fal.ai veo 3 pricyyyy

[wise cypress]

If I want to add credits are there any credit add on promos available

Ultra

I've made about 6 or 7 minutes of content got 2000 creds left lol

[crisp basalt]

True

[raw swan]

if anyone knows, what is the pricing for veo 3 per second video in CAD? it would be helpful.

[wise cypress]

Any apps for that for veo3 yet

Selling gens cheaper than ultra add on creds

What do you estimate at per Gen currently Google is a dollar per gen

Currently each Google veo 3 generation costs 100 credits which is USD equivalent of $1. What do you estimate the discount would be with a third party. @edgy holly

[barren wagon]

why does my video keep getting removed in the #flow-videos・📹 ? theres legit humans with blood here and mine was just a muppet with unrealistic blood - i added a trigger warning and a spoiler. i dont understand. i also read the rules. i wonder why in particular a super unrealistic video gets flagged when the rest of the realistic content floats under the radar. if my video gets removed than any content with blood should be removed as well.

[barren wagon]

eh ill give it that

oh i was talking about when i upload a video to the gallery on discord it gets removed and i got warned

[wise cypress]

Discord mods what do u expect

[hearty token]

Want a cool way to do it and also maybe upscale it? Dont know if itl work on veo because my access is gone meow, but im sure its worth a go

[vocal stag]

there is no sound on flow

and he still got subtiles

[red oracle]

Over 1000 credits used (Veo 3 - Highest Quality in Flow) and no sound 🙃

Has anyone got audio generation to work with frame to video? I’m thinking it only works with text to video now

[dapper wagon]

I've gotten a few to work with frame to video but a lot of fails

[wooden orbit]

I am so tired of this. Just generated 16 Veo 3 videos and NONE of them have audio. Total waste of money. They need to return credits for this

[fallow elk]

Guys, why i can't use flow AI, i already have membership (edu) for it

[sour furnace]

Hey does anyone know why when in screen builder I don't have any audio output when I "extend," or "jump to?" Only the first generated video outputs any type of audio, (This also occurs even when I use the same prompt to extend or jump to for reference.)

[fierce nimbus]

Extend uses Veo 2. Jump To should have sound as it uses Veo 3.

[warped arrow]

Have you tried this? I tried it before, but it failed for me. But that was like 2 weeks ago, so maybe they fixed it now.

I mean the video was created but no audio despite using Veo 3

[winged plinth]

Nick is google staff. Trust his words.

[daring vortex]

hi

[mystic forge]

Wow the veo 3 flow voice not working for sound. It plays music background but voices don't play

Yea no audio still for me too. That was the best feature. Why they stop it

[storm horizon]

Hey guys. I got my Gemini AI Pro plan and I was able to generate some images using the Google Gemini App. However I tried to go into flow and It says I have 0 credits. Do you have any idea why and how to fix this issue?

[ripe mirage]

#1376636428674666640 message

I think you are probably not doing text only...

[brisk idol]

its not available in india ? when will access possible?

[gritty verge]

just some feedback for image to video and switching to Ultra plan to try out Ingredients: turns out ingredients only supports veo2 and the results are blurry and don't follow the prompt much. the file output is also generally under 6kbps which is very low quality and "upscaling" to 1080p just oversharpens the video. not impressed at all. I know txt to vid is the main strength but damn

[mortal stone]

Thx for the tip @fierce nimbus ! Is extend option planned for VEO3 someday ?

[silk forum]

Hello everyone, I want to ask a question. I bought your Gemini ultra packages line but the features are not working. I enter the flow page and I encounter a problem that it does not work in your region. When I go to the Gemini Ai page, it shows that I still have a pro membership but I bought the premium membership.

[remote umbra]

Does anyone know if flow currently available in UK? I've signed up to AI pro but the "Create with Flow" CTA is still greyed out on here: https://labs.google/flow/about

[jovial vessel]

should be available, i'm from asia and using it without problem.

[shell mason]

Did you verify your age?

[tribal crest]

I lose audio every time I use jump to in veo 3

[crisp basalt]

You didn’t read the faq?

[gritty verge]

no I don't read faqs before trying out generators especially since I've used all the other ones without needing to. it should tell you before upgrading that it requires Ultra and that it only supports veo2 but it doesn't mention that important part

the UI having veo3 as a toggle for ingredients makes it even worse since it's not a real option right now

[prime delta]

i have a question, i tried to prompt a scene with a dialugue on it but it says, "Only English prompts are supported at this time."

[verbal void]

I just burned the last of.my credits with silent gens that needed speech that was included in the prompt in quotes

[fading dirge]

dang ive waisted 3 hundred credits on bad outputs, is veo 3 alright?

[red oracle]

You need the ultra plan to use it in flow

[mortal garnet]

Create a 1-minute AI video in Moroccan Arabic (Darija).

🎙️ Voice-over:

• Male voice
• Tone: sarcastic with a hint of sadness
• Sounds like a tired but clever citizen

🎥 Visual Style:

• A poor Moroccan city with broken streets, trash, construction left unfinished
• Busy cafes with people watching silently
• A few shots of empty offices and politicians in suits ignoring people
• Sad music in the background

📝 Script (in Moroccan Arabic):

"مرحبا بك فشايل زايدة…
المدينة اللي زايد فيها كلشي، إلا الخير!

زايدة فالثمن ديال الخبز،
زايدة فالحفر،
زايدة فالكريدي،
وزايدة حتى فالنفاق ديال المسؤولين.

فين ما مشيتي… حفرة كتشوفك قبل ما تشوفها!
الضو كيتقطع، الما كيمشي، والحيط كيسمع الشكاوي أكثر من القايد.

شايل زايدة؟
مدينة معمّرة بالأمل ديال البارح،
والمشاريع لي ما كتمشيش حتى فالحلم.

الناس؟ ساكتين… ولا لابسين السماعة باش ما يسمعوش الظلم.

المسؤول؟ كيشوف المدينة بحال طَرطة…
كيوزعها على صحابو، وباقي كيقول: 'راه مازال ما كملناش الإصلاح.'

ولكن بصراحة…
المدينة ما خاصهاش إصلاح،
المدينة خاصها الناس ترجع تعيش، ماشي غير تعيش باش تموت واقفة!

هنا فشايل زايدة…
زايدين غير فالصبر، والحساب… باقي."

[median hearth]

Is there any word on when we can extend clips using the veo3 audio enabled model?

[frigid pier]

I consumed more than 9k credits in just few days

unfortunately most of the videos weren't worth to use but also got surprised by the quality from some generations.

[jaunty moon]

anybody can help me for some reason i cannot access my flow... i am a google ai pro Subscriber, i am FL. for some reason when i try to enter flow its grayed it out. its not letting me in. anybody had this same issue? or has an idea what i can do?

[mortal stone]

@fierce nimbus VE03 : i can confirm that sound isn't generated when i use "jump to" from one perfectly-working audio scene to the next (and gibberish subtitles replacing dialogues are back too) if it helps

[random wave]

<@&1173646484311384114> I am from Tanzania how can i access Flow

[kindred terrace]

It is. I’ve been using it (with differing degrees of success) for a week now.

[remote umbra]

With AI Pro account or with Ultra? If AI Pro - forgive the stupid question - what link/button are you using to navigate to it?

[kindred terrace]

I have an Ultra.

[kindred terrace]

Although my son has used it via the (student) Pro account.

[ripe mirage]

It doesn’t appear to.

[kindred terrace]

Indeed. Doesn’t give me sound

[ripe mirage]

Only text to video has full audio and sfx

What I would really like to know from Google is if that is a manufactured limitation to keep it from swallowing an industry or if it is a technical limitation for the time being because the answer to that determines whether or not the subscription for ultra is continuing until next month.

You can have object consistency or audio but not both

[crisp basalt]

https://tenor.com/view/wtf-gif-18370552922777914358

[static mason]

New here and considering an investment in Flow ultra to see if i can somewhat quickly create a rich trailer for a set of training materials for early childhood teachers which my team and I plan to build. I'm a preschool educator self funding this experiment so this is a lot of bakesales to cover the cost, and I'm wondering if i can make sure it will succeed better than my failed attempts to do this in synthesia and google vid, and also to get really clear on how to do it well before i burn through my first months credits. I've scrolled up but is there a good source doc about recent advice ? I'll aslo post my desired product here and would be so grateful for feedback about how realistic it is that i'll get an output i can show to other teachers if the community can weigh in?

My flow idea: Short video showing best practices in action. camera studies a shelf with materials, pans back to show us two children selecting materials, and bringing materials to a small table to start working. The details need to be heavily influenced by me so they show best practices in action and then i want to overlay pop ups with a little more info of the pratices and a (fake for now) link to training materials and templates that would help a teacher enact the brief scene she saw unfold.

My concerns about flow: 1) EDUCATOR DETAILS PRESERVED? is it going to be hard to control the maerials enough to meet my detailed opinion about what are appropriate supplies for a 3 vs 4 year old, 2) WHOLESOME KID VIDEO OKAY? will i have trouble prompting video about young children even for this totally wholesome purpose,

3. CONTINUITY OF CHILDREN, TEACHER, SUPPLIES, and CLASSROOM CORNER? will the children and any teacher and classroom i render stay consistent or will we randomly cut to changing people places and things as we move through this simple overview of what owuld be 5 successful minutes in a real classroom. 4) POP UPS AND NARRATOR EADY TO ADD? Ideally i'd like to add the pop ups, a voice over, maybe an avatar narrator. but the main thing I want to do is generate something teachers can watch and feel inspired by before we ask them to read and reflect and work hard after a busy challenging day.

5. FUTURE VIDEO SHOWING UPSET CHILD? Ultimately I also want to render videos of challenging moments (child crying) from real events I have solved but with the richness of video without worrying about the privacy of an actual child acting upset on camera; will i be able to capture, even if looking from afar or behind, enough realism that teachers will relate to the generated video of a tough moment? 6) TIMELINE ASAP REALISTIC? I want to do all this in a few days next week to get a good enough prototype for my leadership team (who use few computer tools and no AI of any kind) so they can see what is possible and feel more inspired and less afraid about the possibiliites of me investing more time on this-/.....So will a big investmnt in Flow and Ultra help me pull this off fast?

[ripe mirage]

Hi Erica,

First, I’m a no one. Just playing with this for fun. I have a background in IT system administration and web development. Bottom line take me with a not a creative professional view.

I read your comment. I think it is probably doable with just pro. The benefit to Ultra is it adds ingredients for object consistency.

Given that it’s probably fine for you to create short clips and then just include them as cut scenes.

I would suggest you start there as you can get some credits for flow and use veo3 to generate clips in Gemini without having to spend anything.

Play around with the videos you create in a video editor add your narrator there.

Google’s Gemini AI studio will let you do voice narration and even multi voice output.

[austere hearth]

Will flow replace content creators and introduce only creative ideas?

[ripe mirage]

Not yet

[azure nymph]

Nop

I think it will take a long time for this to happen (maybe never)

Or soon

[austere hearth]

Not yet. But I'm afraid that it will happen in the future

[azure nymph]

We all are afraid about the future of AI

At the same time looking forward to what’s to come

[austere hearth]

Actually it's good that creativity and unique thinking will be valued but professionals will have to face challenges because of unlimited competitors

[austere hearth]

Yeah. AI has big potential and we'll have to take risk

[ripe mirage]

I think the big problem with AI is our world is structured in a way that we receive our self worth and financial livelihood from work. A work environment that has to change and adapt and we are woefully under prepared for the agenic future even while it is on us.

[azure nymph]

Yea

[stable tree]

Good afternoon, folks! Greetings from Brazil 🇧🇷

I’m trying to create a 60-second video with a nice flow.
My idea was to split it into 8-second segments and then join everything together.

The challenge I’m facing is generating multiple videos that feature the same AI agent (person) consistently across all segments.

Does anyone here know how to help with that?

[azure nymph]

I like this new era

[ripe mirage]

You can have consistency or you can have full audio. You can’t have both.

[azure nymph]

it’s hard

[austere hearth]

Exactly. Everything is happening so fast that we aren't able to adapt with this era yet. Many are losing jobs but we aren't able to create more job sectors using ai yet.
More jobs will be created, but it'll take time

[azure nymph]

There is a survey that says that 85 million jobs will end up because of AI and +-120 million will be generated because of AI

Something like this, im unsure

[austere hearth]

I hope so

[ripe mirage]

IMO if we don’t want massive social upheaval we will need to establish UBI to create a base that is enough to live on. Modestly. Then any work you do would add to that base.

UBI has had study after study that shows it creates freedom and people overwhelmingly use that agency to be better people. More creative, higher pursuits.

In the long run AI and thinking machines will be a tremendous boost for all of humanity. We just need to get there.

[austere hearth]

Actually our country has unemployment issues. Maximum youth have to work hard till 28/30 to get a job. Before that they have to rely on their parents. Which is really unexpectable

Many got their living from freelancing. Now it may get shut because of ai

[azure nymph]

Interesting

[ripe mirage]

Anthropic’s CEO is talking about what’s coming.

https://amp.cnn.com/cnn/2025/05/29/tech/ai-anthropic-ceo-dario-amodei-unemployment

20%+ unemployment worldwide before the end of the decade.

[azure nymph]

hmmm

What is the solution for this?

[hard panther]

i dont mind automating a couple healthcare jobs away if thatll lower health insurance bill here back to a justified level 🫣

[azure nymph]

I saw a case here in my country (Brazil), a company owner fired 16 employees and replaced him with only 1 AI that brings more return in less time.

The 16 employees cost R$ 128,000 per month, AI only R$ 4500 per month...

He made a video for instagram very happy with this 😂

https://youtube.com/shorts/q92HwaI3hUo?si=E3FolfvhBFSsmIhe

[hidden torrent]

Can flow produce video is 9:16 format?

[azure nymph]

hmm I don’t think so

[hidden torrent]

Ok, thanks!

[sharp shore]

/create prompt: a futuristic city at night, flying cars, neon lights

#flow-general・🌊 The girl is walking on the podtum, I'm filming her paparazzi

[verbal void]

just recently i think in the past 24 hours tools such as fal, openart and some others have added lower priced subs to get a few limited veo3 generations if you wanted to experiment without the big upfront investment if that helps

[boreal quail]

is there any way to access flow in europe?

[minor rune]

Maybe, check your DM

[red oracle]

Does Veo 3 Allow swear words in audio prompts?

[low nexus]

yo i thought you get it daily? why does it say 13 days?!?!?!

[tardy ridge]

With the release of the API and increased use generation quality has taken a big hit is google going to bring on more gpu? I burned thru 500 credits trying to get a shot that I would have gotten in 1 before the api

[fair bridge]

What musical styles & genres is Veo3 best at?

[ripe mirage]

CW: includes profanity.

Nobody ever went to the bathroom in the series. They crossed time and space but never waded into this crap.

[red oracle]

Fantastic lmao

[ripe mirage]

That in Gemini app? I no longer believe those warnings.

[low nexus]

gemini website

[azure nymph]

this warning appears in app and web for me

[ivory thicket]

why does this pop up? ive subscribed to the free trial

[frigid pier]

hi

[knotty lintel]

Hey there I have been trying to generate on Flow with Veo 3, everything I do never has audio and when I go back to the clip later on it says VEO 2 at the bottom...what am I missing here? I am on pro plan. Also anyone have a good way to prompt for first person POV?

[visual cove]

I am an Ultra user, using Flow, and yet, I still get a visible watermark, which isn’t supposed to show for Ultra users using Flow. Why is this happening, and how do I turn the watermark off?

Also, there is a 50% (for your first 3 months) promo advertised, but every time I would log in to subscribe, that promo completely disappeared, DESPITE ME BEING A NEW USER. (I even tried multiple Google accounts, I’ve never subscribed to Google AI on ANY OF THEM, but same issue).

I eventually caved and just paid the full price, as I was unwilling to completely forgo access to Flow, and the Pro plan (which by the way, was supposed to be FREE at first, but again, defaulted to $20) is insufficient for what I want to use it for.

Can anyone tell me WHY the promo kept mysteriously disappearing for me when I logged in to get it?

And now I’m stuck with a watermark that Google has specifically claimed shouldn’t be there as an Ultra user using Flow?

[real obsidian]

i guess parody is against the TOS? can't reference any public figures in prompts. rip

pretty lame, since the main use case of AI right now is parody content

[silver knot]

Hello, can anyone help me please? I subscribed to a free plan but AI credits weren't added to my account, what can I do to fix it?

[icy dew]

Hi all, is there a good place to test a prompt before using it in flow? I'd like to use the credits smartly 🥹

[frigid pier]

wow we have Veo 3 fast now!

oh wait it disappeared weird hmmm

[raw swan]

@frigid pier ????

[frigid pier]

?

[raw swan]

yes I just saw it

but it doesnt generate

[frigid pier]

yea I think they cancelled it

[raw swan]

no they didnt

it is even on their guide page

[frigid pier]

hm that's weird then?

[frigid pier]

I have 1.5k credits only

can generate only 15 videos

then have to wait til July wow

😦

[raw swan]

@frigid pier https://support.google.com/googleone/answer/16287445?hl=en

[frigid pier]

oh wow :0

it was real!

[raw swan]

yep

[frigid pier]

but cannot use it :/

[raw swan]

yeah me too

[frigid pier]

how many credits do u have

[raw swan]

I think when they officially announce it

maybe tomorrow

I have 180 credits left

i am on pro plan

[frigid pier]

oh i see. I am using Veo 3 now and I am not getting any audio now why???\

[crisp basalt]

is that

a fast mode i see

[prime delta]

why I got error "Only English prompts are supported at this time."

[crisp basalt]

😄

[prime delta]

but saw some content creators prompt Filipino language videos.

[crisp basalt]

do it in gemini website

flow can only do english language prompt

[prime delta]

Flow deosn't allow other languages to be speak on a prompt video?

[frigid pier]

This is so weird. I upscaled a Veo 3 video that didn't have audio
then it has audio in the upscaled version lmfaooo wth is this

the software has some unexpected process.

[azure nymph]

nice

veo 3 fast costs 20 credits

What is the difference between Fast and Quality?

[crisp basalt]

the waiting time is 1-2 minutes now for veo 3 gemini

[frigid pier]

so it's for preview ok.

[crisp basalt]

rather than 3-5 minutes

[frigid pier]

bro this Veo 3 is really for producers who make money from videos.
cuz omg I consumed 90% of 12k credits in just few days!!

[prime delta]

hahaha this is the output, I tried to use the Filipino for the lady to speak but it automatically changed to english. 😆
https://g.co/gemini/share/71f866085f5e

[frigid pier]

I feel like my wallet is draining.
if i am not making any profits from it

[crisp basalt]

iirc,, theres no way to ./.. prompt it to speak exactly what the dialogue u want it to say if its other languages

[azure nymph]

Gemini app just Veo 3

[crisp basalt]

the only way to do it is to prompt it like.... "the guy explains how to use veo 3 in tagalog"

[azure nymph]

Not fast

[crisp basalt]

its a slow roll out

[frigid pier]

slow?

[azure nymph]

web is already Veo 3 fast

Has my limit been reset? hmmm

[crisp basalt]

yeah it might be

mine got reset too, supposedly i had t owait until june 12

[azure nymph]

thats good

[crisp basalt]

i made a tagalog spoken video just now ,

A static, eye-level mockumentary shot under natural lighting, set in a slightly unkempt urban backyard. The frame centers on a middle-aged man with a serious, slightly bewildered expression, engaged in what appears to be a very important conversation with a completely uninterested pigeon that is diligently pecking at the grass near his feet. The visual style is naturalistic, like a TV documentary segment. Ambient city sounds are present, mixed with the soft shuffling and pecking sounds of the pigeon. We hear the man's voice, speaking in fluent Tagalog, his tone earnest and explanatory, creating a humorous contrast with the pigeon's complete lack of reaction. An understated, perhaps slightly offbeat, documentary-style musical cue underscores the scene.

this is veo 3 fast btw

made another one with veo 3 fast, with spoken dialogue in the prompt

[brisk panther]

That sucks. Regular Veo 3 used to be there.

[prime delta]

thanks! ahahaha this one is cool!

[crisp basalt]

was the dialogue spoken makes sense to you? it sounds legit but im not sure if it makes sense to you

ive seen some veo 3 that has spoken dialogue in my language , and so far theyre all sounds so natural

[prime delta]

yes it's tagalog. but I already find a solution on how to make it speak in tagalog script. thanks

[frigid pier]

I am kinda curious why uploading image is quite slow on Flow hmm

[torpid moon]

I have 1 month subscription but when I go to flow it says, I can generate veo3 videos with Gemini but wanna use the flow:
You must be a Google AI subscriber to generate. Feel free to look around in the meantime.

[frigid pier]

what's your location

[wise cypress]

Veo3 fast has anyone messed with it how are the gens

[glad cedar]

Has anyone tried the new Veo3 - Fast for 20 credits? It appears to be available on mobile only

[azure nymph]

i just tried and i didnt like tbh

[glad cedar]

Did they already remove veo3 fast? I can't see it anymore

[wise cypress]

Ya its gone

I was just about to try it was prompt building

I just don't throw away creds lol

Rest of the night I'm gonna do post prod until it's back up lol

[torpid moon]

I fixed

But It says Only English prompts are supported at this time.
I used Georgian Language to generate and it said it couldn't but clearly people have been doing it on TikTok and all over internet in different languages with subtitles. Why can't I make it

[frigid pier]

bro I am really sad. I have only 330 credits now.

can make 3 Veo3 videos xD

[wise cypress]

Save for veo3 fast

[frigid pier]

already ran out bro :/

[wise cypress]

Lol

[frigid pier]

wen veo3 fast tho

[wise cypress]

It was up for a little bit but they took it down

By the way dude if you have any Clips you like that have s***** captions on them just take them in a video editor and crop them down or put letter boxes instead so the captions are not seen

Bad captions I mean

[raw swan]

Just used Veo 3 - Fast and it is mind blowing

[sinful gorge]

I got a free trieal for a month but veo 3 still say to me that I have to buy a subscribe, why? I wanted to try veo 3, but even with a subscribition I can not do it.

[stable ocean]

Hi there!
Where can I post feedback about Veo 2 in Flow and it will actually be read by developers or product managers?

[bitter spire]

Veo3 fast?! Is it possible to use it on browser?

[raw swan]

Yes

[bitter spire]

Yes i can see it on a flow mobile version..but there was no veo3 fast model yesterday....but why i can't run it with frame to video ?

[raw swan]

I think it is text to video for now

[bitter spire]

That is not a good news..i really like to use for animations with my own images ,so much fun..but veo2 gives maybe 1 good result from 5..too much defusion,it rotates the image,and dont follow the prompt well...but i love it anyways..i edit those pieces afterwards

[raw swan]

yeah hope they add the frames to video to veo 3 - fast, soon

[brisk idol]

please launch flow in india asap . so without vpn we can also chek ot out

[fierce jetty]

I’ve generated a few videos. When I click on add to scene I keep getting “Error: something went wrong “ please help 🥺

[lament canopy]

Let me guess, you are on mobile^

I've only encountered this problem when on tablet / mobile -

[fierce jetty]

Correct, I thought maybe since it would generate it would work. Is there really no way to use it on mobile?

[lament canopy]

not that I know of so far, I've tried iphone and Ipad both with desktop mode activated tried on google chrome tried on opera, brave and even safari and always the same error

[fierce jetty]

I see, thanks for letting me know

[cinder geyser]

i can't use different language prompt..it was working but not anymore

[tiny patio]

I like the 20 credit cost but so far VEO 3 Fast is giving me computer game graphics, not realism.

[cinder geyser]

20 for low 50 credit for normal quality 100 credit for high

how to remove veo watermark on bottom?

[crisp basalt]

Subscribe to ULTRA PLAN

[desert leaf]

hey! im trying to use the veo3 option in (frames to video) and this message is appering and returning to veo2 option . someone knows how to fix this?

[cinder geyser]

frame to video with veo 3 fast doesnt work.

[real obsidian]

prompting skill issue

veo3 fast seems pretty legit, slightly lower visual quality but 1/5 cost

[tiny patio]

Ok, exact same prompt in VEO2 was not cartoony.

[cinder geyser]

does anyone know how to remove music from video i want enviroment sound more than music in background

[real obsidian]

what's the prompt

[sonic vigil]

Has anyone figured out how to ensure that you get audio when you select that model? I don’t wanna keep wasting credits.

[tiny patio]

Nope. It's pot luck.

[analog zealot]

in order to keep the character the same across veo3 clips do i need to use scenebuilder or can i use a character description and itll look the same?

@tiny patio i feel like veo3 fast is good to test audio issues without spending the full amount

basically reduce visual and keep audio the same

[tiny patio]

My last 2 had better video but no audio 😦

[analog zealot]

ive noticed changing the way you explain how they speak or how sounds are related to the video it helps. even then ive had muted videos

honestly not as many as when i first started

[tiny patio]

The exact same prompt gives some silent, some with audio. It's RNG.

[analog zealot]

can u send the prompt

[tiny patio]

"Set around the 1870's a boy covered in soot is sweeping out a dirty chimney. He turns to the camera and says: Gas and Electricity will take my job. Then the camera zooms in to the boy's face for a close up shot and the boy says: Stop Progress." First few in the same style had audio but bad video. Last few better video but no audio. Using " " tends to give terrible subtitles.

[analog zealot]

oh ok

yeah so use " " and tell it do not add subtitles at the end of everything

👍

oh wow it doesnt let u do audio, maybe cause its a kids voice. telling it that its a teenager instead of a kid fixes this

[crisp basalt]

exactly. using fast mode means user needs to iterate the prompt until it can generate what they really want

[analog zealot]

^

[crisp basalt]

them watermark is getting bigger now i noticed

it means i can flex my pro user benefit

🙂

[crisp basalt]

just curious have u tried with fast mode veo 3? i noticed it doesnt do subtitle anymore

[analog zealot]

that was fast and i guess i wasnt forceful enough with the no subtitle message.

[crisp basalt]

oh damn

wait let me iterate the prompt and try it myself

[analog zealot]

In a gritty 1870s Victorian home, a teenager sweeps out a filthy soot covered fireplace. He turns to the camera with a smudged face and tired eyes and says: “Gas and Electricity will take my job.” The camera slowly zooms in for a close-up as he leans forward and adds with quiet intensity: “Stop Progress.”

Realistic period setting, dim lighting, dust in the air, muted historical color grading.

no subtitles.

[crisp basalt]

ah i wrote "teen" and it failed to generate

thank u for sharing the prompt

[analog zealot]

sometimes u can redo the exact same prompt and itll work, if it initially fails

[crisp basalt]

oh u right lmao

the subtitle appeared correctly for me now

[analog zealot]

fast?

[crisp basalt]

yep

idk what was he doing in the beginning lmao

[analog zealot]

i have a feeling that pro and ultra have access to different models

[crisp basalt]

You are actualy correct tho

in the videomodel config list

theres's veo 3 model with the name veo_3_t2v_pro

[analog zealot]

where is this

[crisp basalt]

its in the network log, its not something every users can see

[analog zealot]

well that explains why ur video is weird

[crisp basalt]

finally no subtitle
Inside a derelict, low key lit 1870s Victorian house rendered in gritty realism. The scene shows a young man with a face streaked with soot, finishing the arduous task of sweeping a vast, blackened fireplace. The air is thick with settled dust, giving the scene a muted, historical color grading. The ambient sound is hushed, punctuated by the soft scrape of his broom and the faint crackle of cold embers. He abruptly stops sweeping, turning his gaze directly to the camera at eye-level. Gas and Electricity will take my job, he states plainly. Leaning in closer, his expression shifts to one of quiet intensity, and he adds, Stop Progress! A slow camera zoom centers on his eyes, emphasizing the desperation beneath the soot.

[analog zealot]

nice

i am so thankful for the library

[knotty lintel]

Where is the library? Also every generation I do right now is without audio anyone else?

[analog zealot]

anything that could seem bad, audio will probably be muted is my guess

make sure ur using veo 3 instead of veo 2

[knotty lintel]

You know whats funny, I do Veo 3 selection, and when the clip is created it says VEO 2 at the bottom. I am on pro