#flow-general・🌊
1 messages · Page 17 of 1
Im guessing Flow Prompts would be the place to ask about prompting
I'm seeing so many Indian creators use Veo 3. I can't figure out how it's not launched here yet. Any ideas or tips that I can try to access it?
I could only guess they're using a VPN or something if its not available there. Or, if they are using the app on their phones. Sometimes changing the region unlocks apps
Ahh, thank you. That helps. Hope they launch it soon here.
I get questions like this all the time in the discord(s) I moderate. Some countries have stricter rules making it harder to get the apps available for that region.
Yeah, true. They do
Check Flow Tv there you have a lot of sample videos and you can see exactly the prompts they used to create them. Is a nice learning tool. You can also use chat GPT and tell it to create a VEO 2 or VEO 3 prompt to make "your idea explained", after that read the results and test the prompt in Veo, if you don´t get what you wanted you can always make changes to the prompt and generate again.
is there a list of different accents veo3 can do successfully? anyone made a video of it or something?
hello how do i get in dialogue on prompt, thank you
Hello, I keep getting sent to this page everytime I try and subscribe to flow. I have owned this account for over a year. How do I resolve this?
Is there no way of using a google workspace account with flow?
Hi! Anyone using the scenebuilder? I'm having troubles exporting the scenes from scenebuilder in the right order
Hi, I already subscribe to google one, why I cant use flow?
hey! please post images and videos in #flow-videos・📹 ! this allows conversation to keep flowing and your art to be shown in the correct place! thank you :D
Audio problem:
I have a problem with my veo3 quality, where my video doesn't produce any sound. An error appears saying that they were unable to generate the audio. I tried to fix it in several ways and it didn't work. Can anyone help me?
audio is heavy robotic for me this morning
I asked gemini 😛
Are you having problems generating audio for some veo3 quality prompts?
Yes, it appears that many users are experiencing problems generating audio for Veo3 quality prompts. This is a common and frequently reported issue.
Quick question for Flow/Veo 3 users: Is it true that if I use the 'Ingredients' (image reference) feature for character consistency, it forces me to Veo 2 (no audio)? But if I want audio, I have to use 'Veo 3 Quality' which doesn't support Ingredients? This makes character continuity with sound impossible right now, unless I'm missing something. Thanks!
I had one of my gallery posts deleted. It was a pastoral scene by Rousseau of sheep by a tree with farmers in the distance. I made a wolf jump out from behind the tree and drag one of the sheep away to be devoured. The sheep left a trail of blood on the ground. At first I thought the moderators deleted it for defiling a classic painting, but then I realized it was probably because of the blood. Sorry!
I tried to have a bagpipe player flattened by a steamroller but the steamroller kept bouncing off him! More violence guardrails.
See "What are Flow's Capabilities" on https://labs.google/fx/tools/flow/faq for what each model supports. We're working on improvements there to support more use-cases.
Google Lab's AI video generation tool, powered by Google DeepMind’s latest model: Veo 2
@unreal viper
How is the quality for Veo 3 Fast?
Am I the only one who cannot create scenes with veo3?
IMO. Much better than Veo 2 Quality setting, but slightly less than Veo 3 Quality setting.
what do your prompts look like?
I have had 6 "Couldn't generate video. Try again later." happen and it has taken the 100 credits every time, it's not the prompt I tried using the prompt of a previous clip that was successful. Do we get refunded since it doesn't even generate anything?
Flow using VEO3 stopped generating with audio today all settings are good HELP Lost 13 generations
If you're using a VPN try turning it off.
No VPN
No VPN
is it possible to make vertical videos for reels?
There's something in your prompt it doesnt like.. either violence, drugs, or a referece to something percieved to be a child - run it through chatgpt and ask it to analyze why veo3 is killing audio.
How can I create video with Veo 3 now?( I had Gemini Pro)
I have that a try and trying to regenerate it
What country are you in? If you are in Europe, it is not yet available. You have to wait.
Can i use Canada VPN instead?
The image literally tells you. 
ok Yup turns out having a family in the back yard playing makes it mute the audio
hi there! Flow is not currently available in all countries where Google AI Pro is offered.
Please note:
VPNs will not enable access in unsupported regions.
Before purchasing a Google AI Pro subscription, check the list of Flow supported regions: https://support-content-draft.corp.google.com/labs/answer/16353544
If you’re in a supported region country, please reach out to Google One support: https://support.google.com/googleone/answer/9003266?hl=en&sjid=8888414863080858817-SA#about_ongoing_benefits:~:text=Get help with Google One
With Google One, you can claim your new benefit or trial from the Google One app or the Google One website. You can also get notifications about the latest updates and offers from Google. Tip: Googl
Google Drive: Sign-in
I prompted: Show me a geared elevator machine standing upright . This is what I got:
https://youtu.be/BrAu1s8YoIw?si=Xs6MvT57xBRrJgOu
My device volume is set high.
How do I hear what the guy is saying?
How are people making consistent videos in scenebuilder? Have tried both simple and more detailed prompts and can't keep the same aesthetic or character for each new scene
takes a few tries. It's never perfect
my audio is also not working in veo 3 using flow. last 7 generations produced no audio in the veo3 text to video setting.
Just a description of people and what they should do
you at least run it through gpt?
No….
Should I?
have gpt make extremely detailed descriptions
use the same description in every prompt
tell it what it's for too.. veo 3.. consistent character
it knows 😛
it will be easier when/if they ever get the frame stuff working
I don't think it's possible at the moment. Jump / Extend send me to Veo 2...
I would agree. You can get it to the point where people won't notice so much if at all, but it takes quite a few tries and gets costly
write things that incorporate the 8 second and random nature. I've seen a few people doing it pretty well.
Exactly. You better come up with different 8 second videos, where character differences are okay.
I'm working on two videos that are sending the messaging for me, that is needed, but with different characters.
Here's a video where a character and it's differences is not much of a pain 👉 https://x.com/avax/status/1936061495928115491
is that yours?
I have a couple history standup jokes I might try something like that with. When my credits reup
New here
No, this is mine. 3 more ready, that are a 1+ minute long - https://x.com/Quantinium_fdn/status/1937139031621415272
The world’s most famous farmer is now farming… Quantinium 🛜
It’s not much, but it’s honest work... 🧑🌾
Start farming by joining our early waitlist: the Q‑ueue 👉 https://t.co/SM1vhdkhZI
P.S. Thank you, @Google and @GoogleLabs
Anyone having audio issues my shit’s not working
I want to be able to add a phot and have a video generated with audio and sound of that person but VEO3 does not do that
I can use image as a reference, but AFAIK it goes to Veo2, where there is no sound.
Yeah that is the problem I'm having I want it to do sound which it claims it does but it does not
It will get there
the amount of bugs is astounding. might as well refund everyone the money
they still have the cheek to charge sky high prices for a failed product.
within a short span for 2 days i have encountered almost every single bug in the bug channel. when will google do a refund?
other AI Gen videos at the very minimal least follow my prompts. Google Flow AI doesnt follow prompts and does whatever it likes to do. this is insane..
I am finding it is VERY sensitive to the words used in the prompt. If it even sniffs a hint of violence or sexuality... NO AUDIO for you!
This prompt wouldn't make audio... I had to have GPT "soften the rhetoric" to get audio.
======
Create an 8-second cinematic video set at night in a quiet, dimly lit alley behind a restaurant or stadium. The only light comes from a few scattered streetlamps and the faint blue glow of moonlight bouncing off the tops of metal trashcans. The pavement is cracked and slightly wet, reflecting small puddles of light.
Suddenly, a group of 5 to 7 ultra-realistic raccoons, all wearing slightly dirtied Trash Pandas baseball uniforms, comes charging toward the trashcans from behind the camera. The raccoons are in motion—some sprinting on all fours, others upright and charging like miniature warriors. A few carry baseball bats over their shoulders like weapons. Their jerseys are still proudly printed with “Trash Pandas” across the front, though some are untucked or stained from the game.
The camera is positioned behind them, handheld and low to the ground, following as they run forward in a pack. The motion should feel dynamic and chaotic, like a war movie charge scene or sports tunnel sprint.
As they rush forward, one raccoon toward the front yells loudly and proudly over the chaos:
“Let’s get after it, boys!”
As they reach the trashcans, the raccoons leap, climb, and dive into them, immediately tearing through bags, tossing trash into the air, and rolling in discarded food wrappers. The bats clang against the cans, banana peels fly, and the frenzy of celebration begins. The mood is triumphant, wild, and primal—this is their victory parade.
I don't know bro, I would wait just in case.
.
What other AIs generate sound for you? That match on delivery, I mean. Not ones where you have to extra generate audio, and then put it on top to match.
I did, no audio tho...
google flow doesnt have consistent audio as well
It has some audio, especially if you want to build different scenes.
no kissing as well. but i managed to get it to kiss.
it glitches through doors and chairs. try it
it gltiches through other human heads as well
i use gemini pro 2.5 to build a very detailed prompt. i tweak the details to my specific needs, and yet it only captures like 20% of my instructions
Hey everyone, I’m having trouble using Flow. I’m already on the Google AI Pro plan, but when I open Flow, I get a banner telling me to subscribe. When I click it, it just takes me to Google One, which says I’m already subscribed. 🤷♂️
- Anyone else run into this? How did you fix it?
- Are there any Flow support folks in this server?
Thanks in advance!
i tried shortening the prompt as well. doesnt make a difference. it prefers to do whatever its training model learnt. i watch alot of videos online and i saw and recognised many reused assets from online videos.
you will find many users with the same issues in the bugs section
hey there! welcome to the community 🙂 Flow is not currently available in all countries where Google AI Pro is offered.
Please note:
VPNs will not enable access in unsupported regions.
Before purchasing a Google AI Pro subscription, check the list of Flow supported regions: https://support-content-draft.corp.google.com/labs/answer/16353544
If you’re in a supported region country and still having this issue, please reach out to Google One support: https://support.google.com/googleone/answer/9003266?hl=en&sjid=8888414863080858817-SA#about_ongoing_benefits:~:text=Get help with Google One
With Google One, you can claim your new benefit or trial from the Google One app or the Google One website. You can also get notifications about the latest updates and offers from Google. Tip: Googl
I use chatgpt for prompts
Bro, my audio is not working on my ultra now
Hello everyone! Just joined the server! excited to connect with other people also interested in VEO 3 and Google's other AI products!
i need help malking those funny vlog vids in flow
20$ for 15 secs lol
Used chat gpt for prompts, used all 1000 credits, didnt even come close to what I wanted. and credit are only avaliable for ultra users? thats some insane stuff right there. What a waste of money honestly. I'd spend 8 hours trying to get a 8 second clip how i want it even with ChatGpt.
It would be good if all tiers could buy credits. It seems like not doing that is leaving money on the table to be honest.
why are all the videos black?
I wonder if they will eventually let us buy more credits.
Hey, guys. What do you think? I'm receiving a lot of heat on this one: https://www.youtube.com/watch?v=S4ZSoTPuWf8
Tools used:
- Google's VEO 3 [video generation]
- Google's Gemini [lyrics + prompt generation/refinement]
- UDIO [audio backing track generation]
- Ableton Live [audio backing track embellishment + mastering]
- Adobe Premiere [editing, golor grading]
You can freely access all generated assets [videos, audio tracks], plus the exact prompts used...
how do u make those talking animals like the big foot and yeti
Heey guys, someone know why when i put a prompt in Flow the video that gives me is in black? I try several times and Flow take my credits but do not work
Hello, is it that in the pro plan
We only use once the veo3
Coz I have tried to generate the first video
It came with audio about 8 seconds
WHile I was trying to add , it takes me to veo2
Which does not produce with sound
??
investigating this! thanks for reporting!
if others are seeing this, can you please report in #1319424187730755636?
Whenever I try to switch to veo3
IT says "switching you to a compatible model for this feature. Submit again to confirm (or check settings for details"
Check in your Flow library to see if the videos are available there.
audio is not working for me
we're investigating - please report this in #1319424187730755636!
Same here
Hey, good news from Spain. It will be released in the next few days. I just got this new one.
?
Yes. Veo 3 is not available with flow's extend feature, or image to video. You should have received this notification telling you.
Why won’t it let me use itV
Just sent you a DM
you'll have to scroll down in order to have the 'continue' button populate!
Can anyone help me - why is it that sometimes I can use "Extend" in Scene Builder and use Veo3, but other times it forces me to Veo2?
I have a pro plan, when I am logged in. It says you don’t have any credits
hey! for any bugs please report them in #1319424187730755636 so the team can see them easier!
Annyone from the netherlands? Who is using flow
"Black screens" bug -
if you're seeing a 'black screen' or anything similar to this, please reply to my forum post, here: #1386797321836298330 message
When I try to access flow this is what it says every time
just about to do that georgia! youre on it!
It won’t let me use it can you please help me @main saddle ?
what link are you using?
Honestly it's kind of wild to see them using Veo2 in the tutorial video because i've not been able to get Veo2 to make anything even close to the quality in this video
Built with and for creatives, Flow is the only AI filmmaking tool custom designed for Google DeepMind’s most advanced models - Veo, Imagen and Gemini. In this tutorial you will learn how to seamlessly create cinematic clips, scenes and stories with character and object consistency. You can prompt how you speak, craft cohesive narratives, and e...
hey hey, does ingredients work with veo3?
The difference in both video quality and the AI's capability to interpret text is so wildly different between Veo2 and Veo3 that it's kind of crazy that you can't do this yet
I have been trying for hours now to use it and I can’t
no, only text to video, and you can't extend or jump to using veo3 either
Where are you trying to access from? Which country? It depends on where you are. It's not available yet, but the Ultra plan is already available. It might be available in the next few days.
Ahh, if you're from the USA, it's strange that they don't let you.
try using another test....
I finally figured it out
But can I make videos longer than 8 seconds using a free trial?
Yes, it depends on the credits you have.
hello guys what do you think about flow 2.0, it can happen?
A lot of people are still getting fantastic results with Veo 2, especially using frame to video. If you don't need audio for the shot it's still great and it costs less credits. The team is obviously planning to make those features compatible with Veo 3 eventually.
Yes I'm really looking forward to when Veo3 gets the upgrade to frame to video and ingredients. Those are amazing features and if they can get it to work with Veo3's quality it'll be very powerful indeed
I'm finding quality hard to match between them though
Any chance someone knows how to download a whole scene strung together instead of just the individual parts? Can't find the buttom
Hover over the scene and the download button is in the bottom right.
TY!
the generated videos still have black screen
But thanks for answering!
enjoying Flow VEO 3. But all my videos went missing. Anyone else?
Can anyone explain where my videos went?!
hey, the generated videos are Transparent! why is that ??!
please, let me know how to generate transparent videos...
Same bro what’s going on?
Can anyone help with my problem, i subscribed to a google ai plan, but when i try to make a vid with veo3 it tells me that i nees to upgrade my plan, and when i press on "upgrade" it says that im already subscribed, please help
bro why is screen black?
Same bro
kind of unfair to lose credits when veo fails to generate
Dam that was kinda nice bro
Looks like Aikido is too much for AI. Lol!
Same thing here
used up so many credits with no video creation
this sucks
very frustrating
the pricing model is a bit ridiculous tbh
i work in engineering and i know waht it costs, that profit margin tho
also why is there no black people?
hey, im using chrome, in flow non of my videos want to load/download UNLESS I go directly to my library to view them.
is it racist?
anyone get similar issue?
Did you break it? Those faces look terrible!
its what all my generations look like ai sucks still
You're asking it to do too much. It can only handle so many ideas right now. Think of it like an adult with a child's body. Needs more time ot learn.
Is google flow working for you guys?
Yes, I wouldn't use it commercially. Has a long way to go. At least its free for students.
My google flow isn’t working at all
What do you mean
Whats the point in making it paid if its shi tt
hi everyone - does anyone have a prompt for getting audio in veo3 when adding another scene to scenebuilder. thanks!
Pimping ain't easy. Just have to make do. The key is learning what prompts get these to come out better. There is a leaked list of VIP prompts on Github. Research those and see if it helps turn these lemons into gold.
Wan2.1 is rumored to be a better video generator and is free, but it takes a hours to get that thing to spit out videos.
I knew we could take him. No one wanted to listen! 🙂
I mean prompting is not hard when ai can just create the prompt for you
If you want to do simple/regular videos yes
i need like big foot and yeti prompts but cant get good ones from ai
Anyone have issues with no audio on VEO3?
no just they dont have that kind of money
nothing is working....anyone else having issues?
I have a problem and it is that the videos I make with veo 3 do not have audio.
Does anyone know how I make have?
Same here
quick question, is there a way to get more credits?
They call it money.
It only shows an upgrade but it doesn't show where to buy more credits straight up
You have to be in the ultra tier to have the option to “top up” - it’s really annoying
it’s already overpriced but you have to commit to overpricedness to continue to overpay - fairly buckwild
lol, I see
yeah if you wanna hammer the expensive ai chips you gotta pay 😛 fact of life
Hello, how do you create consistent characters in different Flow account?
I have couple flow account since the pro plan has credit limits
Has anyone tried the video generator Wan2.1? I set that thing up around 10 this morning. It's 12 hours later and I am still waiting for my first video to generate. 😭
Yeap! Big time
is it me or the quality has been off since the black screen issue?
Black person?
Learn how to make consistent characters in Veo 3 using Whisk and Gemini! Plus, I show how to use AI tools like Runway or CapCut to remove unwanted Veo 3 subtitles.
Watch my movie "More Than One Reality" where I used this character:
https://youtu.be/25fPB5Vr8ws
I forgot that I removed music from a few clips. Here's how you can remove the music ...
no
Black people dont wear flipflops with v shape white top and blue jeans and glasses
never seen that in my life
Yes they do
it’s just clothes AI doesn’t really know how to dress people right anyway.
hello! is there a way to prompt weight? I'm trying to get certain aspects to my videos that are just not happening
Who can tell me how to set the ratio to 9:16? Thank you!
Here's a fatman eating a hamburger and a skinny man eating a hamburger.. Use descriptions like Lanky, or rounded..
doesn't help me, but a FANTASTIC video. good for him
or if you know a character to reference you can say similar too
well that's how you get different builds
more for audio. The scene is a character panicking in battle so I want a rapid heartbeat SFX
so like "with an emphasis on"
Is there any way i can add myself in picture for short story? I tried but scene not giving me accurate character of me
Hello guys
If anyone could help to use the Veo 3 I’m form the Middle East and it’s not working out
I’m struggling
Guys can you generate this if you can
How many time it tack like the duration
..?
Hello
can use iphone flow de veo 3?
Hi Guys,
Everytime I land on the Flow page, I get a message on the top, you need to subscribe to Google one, when I click that -> you are already subscribed. And the vicious cycle continues. Anybody faced the same issue ? I am just not able to use Veo3 on Flow and Gemini restricts me creating only 3 videos in a day.
Good morning 🌅 all just having a look around and noticed a mention of a wait list is this still active or did I miss it 😬
Good morning!! I am fairly new here and and running into a problem when trying to create video with frames to video. When I try to create, it automatically switches to veo 2. How do I stop it from doing this?
Ahh so no wait list, ok so 30 day trial then 18 a month
Its not supported via vpn
Hi team, I subscribed to Google One AI Pro with the account hongvanhn69@gmail.com.
I’m currently located in the US (San Jose, California), and I’ve confirmed my subscription is active.
However, I still see "0 AI credits" in Flow and can’t generate any videos. Google Support was unable to fix it, saying it's related to rollout limitations.
Could someone here help check or escalate this? I can provide screenshots if needed. Thanks a lot!
If they can't do anything, there's nothing we can do. The credit system is terrible. Don't invest so much into this thing. There's only so much you can do with 8 seconds. Before my 15 month free trial is up if this thing doesn't improve I'm gone.
Hi mods, sorry to follow up — just wondering if anyone might be able to check my case regarding Flow access. Happy to resend details or screenshots if needed.
@heady prairie VPN is Not working. Flow is only avaiable in 70 Countrys
As i Said before 🤓
Are the black screens fixed in Flow?
In wich country Flow is available ?
So i figured it out how to get unlimited credits for free without violating google policies, legit way
Am I the only one having issue navigating around Google One, like where in the site is my next Subscription payment, or when am I getting my AI credit refresh.
Anyone with quality issue since the black screen?
Will flow ever be included for Google Workspace paying users?
Hello
I did this one in google gemini pro using VEO 3. Why the audio was not included?
@woeful fossil
what module did you use?
Welp that'll do it. Veo 3 does not support audio with children.
This, using video
Ohhhh I see so how could I modified? Changing human beings into fruits or vegetables? Maybe animals?
I made a video on flow this morning, text to video using Veo 3 Fast and it had audio, I used a prompt off Gemini too
Big and small racoon?
Yeah those modifications will likely work.
Okay cool. Thanks
Is there any suscription focused in video generation? Google AI ultra is too expensive and google ai pro is too basic (only 3 videos x day)
Pro gives you 1k credits/mo. You can use Veo 3 Fast and get pretty decent results. Otherwise ultra is the only way with flow.
I agree that happened because of the kids, but probably off topic, but sometimes missing audio is caused by a veo bug, and 2 out of 3 of those missing audio bug videos, they are like your best generated videos with just the missing audio lol.
My veo 3 quality video has audio, but when I try to download it, there is no audio.
[Does anyone know what it is?
Sometimes you have to wait a minute or so before trying to download it, especially if you're upscaling to 1080. I've had videos that initially didnt have audio update a short time later.
So just wait?
hey, thanks for sharing this! i'm going to forward your message to #flow-videos・📹 where we like to keep videos like these 🙂
Do you guys think in the future veo 3 can detect which part of the script is TOS and highlights it. It’s a little bit annoying playing the guessing game trying to limit test the prompt I’m sure everyone who actively uses it knows what I mean.
Hey, does anyone know when usually credits gets renewed?
anyone else get charged their monthly bill and credits still not show up?!?!! tf
been over 24hrs since got charged
sheesh same here im talking to support rn
dumb af insane AI tech cant give us credits on time. okay google
I got charged 2 days ago actually
Please let me know how it goes
yeah they said it still processing in their ssytem even tho my bank has been charged adn money gone
so dumb tf
bro I already spent half of it
Dude, I've already spent half of my credits trying to make a video but they all come out without sound. I've tried everything from prompts, positions, gpt chat prompts, I've watched several videos but they all come out without sound. Can someone help me?
Please make sure your model selection
they said to me to wait another 24 hours
Lol
google...... like cmon bro. its google just give me credits i paid for on time?! tffff
I already checked, I change from video 2 to video 3
But it doesn't result in anything
I've already used 500 credits trying and trying but they all come out without audio
What is your prompt?
Two 14-year-old Brazilian girls, Aline and Maria Luísa, are standing in a realistic school courtyard during recess, surrounded by other students watching the drama unfold. They are arguing loudly, using authentic Brazilian Portuguese slang and expressions, over who gets to date a boy named Andrey. The scene is extremely realistic and cinematic, with detailed facial expressions, intense emotions, dramatic hand gestures, and a natural environment with realistic lighting, background noise (birds, students chatting, distant bell), and authentic school architecture.
Aline, wearing a pink crop top and tight jeans, rolls her eyes and says in Brazilian Portuguese: "Tu nem gosta dele de verdade, Maria Luísa! Só quer porque ele tá famosinho no Insta!"
Maria Luísa, in a school uniform with messy hair and a backpack falling off one shoulder, claps back: "Pelo menos ele me responde no Whats, coisa que tu nem sonha!"
They begin a playful but intense argument, filled with dramatic movements, sarcastic facial expressions, hair flips, and dramatic music building in the background. Friends around them react with gasps, laughter, and classic comments like "Eitaaa!" and "Vai dar ruim!"
The boy, Andrey, is standing far in the background, nervously hiding behind a wall, peeking around the corner with wide eyes and holding a soda can. The camera occasionally cuts to his shocked face with zoom-ins and dramatic music stings.
Include realistic Brazilian teenage slang, authentic body language, and school ambiance. The audio must be in natural Brazilian Portuguese. The scene should be hilarious, chaotic, and extremely lifelike — so much so that people watching might think it's real footage from a hidden camera.
Are you sure it allowed 14 years minor girls prompts? Its red flag right there
Do I try to change their age?
anyone struggle with instructing veo3 to have the camera stay in one spot and it just ignores you?
I changed the age and nothing changed
maybe it's also that they're students and that's another red flag for the age gate
A hyper-realistic and cinematic short film showing a Brazilian man being hilariously fired and replaced by a robot at work. The man, in his 30s, wears a worn-out uniform with a name tag that says "Marcão". He’s in a factory, sweaty, tired, eating a pão com mortadela. Suddenly, his arrogant boss walks in with a smug look and introduces the new "employee" — a shiny, emotionless humanoid robot. The man reacts with disbelief and exaggerated expressions. He yells in Brazilian Portuguese slang: "Tá de brincadeira, né? Eu dou meu sangue aqui e cês me metem um 'robôzão' desses?!"
The robot responds in robotic Portuguese: "Não reclamo, não canso, e não como pão com mortadela." Marcão throws his sandwich on the floor and says: "Isso é desumano, véi!"
Make the robot hilariously overqualified and doing ridiculous things like welding, calculating, and even dancing samba while working. Show dramatic slow-motion scenes of Marcão leaving, looking back emotionally while holding his lunchbox. Add hyper-realistic lighting, factory sounds, camera shake, and subtle background laughter of coworkers. The entire dialogue must be in Brazilian Portuguese with real-life slang. The tone is comedic but emotionally touching, almost making the viewer unsure if it’s real or satire.
End the scene with Marcão sitting at a bar, venting to the bartender: "Fui demitido por um robô... que ainda dança melhor que eu."
Is this prompt acceptable?
I have no idea but seems like it might be more approachable yeah since it comes nowhere near children
i can't even get the damn camera to stay still
hi hello fam
I was creating a video on Flow and used the VO3 Quality setting, which costs 100 credits. However, when generating the video, it didn’t apply the resolution settings I specified in the prompt. Also, the video that was supposed to be 30 seconds long ended up generating only 8 seconds.
I’m on the Pro version, and VO3 Quality is available in my plan. Shouldn’t it at least be clearly stated that, even when using VO3 Quality and spending 100 credits, there’s essentially no difference between using the normal and Quality modes in the Pro plan compared to Ultra? This isn’t mentioned anywhere.
PROMPT:
"[Subject Description]
A molten gemstone is poured into a cube mold, solidifies instantly, and then is sliced diagonally with a steel blade.
[Camera Perspective / Style]
Smooth transition from overhead pour to close-up side slice. Warm cinematic tones, slow motion.
[Setting]
Polished white marble slab, glowing studio lighting from both sides.
[Sound]
Molten pouring hiss followed by a soft solid crack sound during slicing. Ultra-clean cutting audio."
8 seconds is standard every clip. no matter quality/model
you can't prompt for how long it will be lmao
research before you waste all your credits
Even Ultra is limited like that?
How do I generate videos with audio here?
Can you help me?
yes, you can't prompt for aspect ratio or length of video. it's all 8 seconds long clips no matter what
You cant generate within the discord, you need to use Flow or Gemini with Veo 3 Fast or Quality to generate audio.
So there's no way to change the resolution from 16:9 to 9:16 either?
thanks
you can prompt for it and very rarely it will turn out. I've gotten a video to come out sideways then i can flip to 9:16 in editing but yeah you will waste credits trying to prompt the aspect ratio. maybe better to prompt for wide angle then just crop it and have your subject in the middle. depends on what you're doing
Is there really any particular trick to finding an idea that can go viral... it is more about the production value and quality is it not?
hopefully that is an update soon. you should be able to choose the ratio 9:16 before you generate. shouldn't have to be prompting for it. because prompts are hit or miss. Sora you can choose before you generate so its a guarantee.
consistency
Thnks bro, help me a lot
it's about being actually funny, or realistic or different. maybe following a trend to an extent but also standing out and putting effort into your prompt. not just basic "bigfoot is in the woods talking about yeti" or whatever. gotta stand out.
I've got a really strong format and prompt i feel i can do anythign with... it is really just picking an idea that i am struggling with seeing in action when clipping it all together
Gandalf (or a wizard) doing everyday modern tasks or dealing with modern day problems...
GrimReaper doing the same
Choosing a background character that is universally recognizable (like the stormtrooper archetype that went viral) it really becomes just getting a good understanding of what is good and a script that can be...funny and natural
@woeful fossil your stuff is dope youv've done your research for sure i like it. You and i are the same it seems
Thanks, I appreciate that. I've definitely done a lot to pay attention to what works with prompting to get solid outputs from the Veo 3.
But, at the end of the day I'm just trying to provide the rubric for others to follow. What goes into a prompt is just as important as what comes out.
Hello, can anybody help me please: I'm adding a video to the scene builder in flow but when I go to the scene builder, the video isn't shown to me, instead there is just a grey screen. Also I cannot click on the + to extend the added video
Hello, am using veo 3 ultra, am trying to generate text in my scene like a neon text saying motivational words on a background and everytime VEO generate something stupid like a succession of letter anybody as a tip? the scene look perfect impossible to obtain the right text thks
a fish speaking to fisherman, is possible?
I'm sure it is, you can choose the style such as animated, old fashioned black and white movie, or realistic or surreal etc and where the fish is, such as in a bowl with the fisherman stood next to it...
Then just decide what your fish is going to say and id imagine it should work!
I tried, but of all the ones I tried, the fish didn't speak, only the fisherman =/
Make it very clear, for example the fish says "hello fisherman" the fisherman then laughs and says "hello fishie"
Hello, can anybody help me please: I'm adding a video to the scene builder in flow but when I go to the scene builder, the video isn't shown to me, instead there is just a grey screen. Also I cannot click on the + to extend the added video (I'm using the ultra membership) please help
it didn't work out either, but thank you for the help
Maybe you could ask Gemini to structure the prompt for you to make sure that the fish speaks, and ask it please ensure the fish is speaking and saying this exact thing etc
No problem
What was your prompt?
A fisherman stands at the edge of a foggy lake, holding a medium-sized silver fish with both hands. The fish looks directly at the fisherman and says, "Damn it, put me back in the water, it's cold out here!" The fisherman replies to the fish, amused, "I get you, fish... it's probably warmer down there." After saying that, the fisherman jumps into the lake together with the fish. These are the only two lines spoken in the scene. First, the fish speaks to the fisherman. Then, the fisherman responds to the fish. No other characters or dialogue are present.
Try that and let me know how it goes. Of course review and make any adjustments you see fit.
thank you for your help, I tried, but only the fisherman speaks, I will try something else, thank you again
Hi there, is frames to video with voice not working or buggy?
So i run it through on fast.. make sure it's a good prompt before doing quality.. this is the fast result..
basically what i wanted, but low quality
Here's the quality version.. same prompt 😛
it's trying too hard
trying to work on over the shoulder shots and different camera movements
Really wish we could get 15 seconds. It would be so much easier.. maybe just for the ultra tier? I'd pay extra 🙂
the audio is not wokring right?
like i tried 5 times lost a lot of credits its not workng
I'm not having audio issues check your setting make sure it's not switching you.. make sure you're not extending or using frame grab..
Have you tried "extend"? How's that working for you?
hello guys
it would be awesome to have a download button and audio where i can upload and match it in that page, but the audio track is too muchos, the download button would be awesome.
this is my card to santa, wish me luck
extend only supports sfx no dialog
hey guys i use gimini but dont have sound on my videos i also tried on iphone but still no chance i life in the netherlands is that te reason why its not working? please someone help me
Can someone please try to tell me how to get sound from the videos I create in the app for Google Flow? The videos I've been getting have a random dude talking about it, but, I can't figure out how to hear what the dude is saying?
hello i just subscribed to gemini pro but for some reason in flow it says as if i am not subscribed to the google ai pro
Is there a way to get a 9:16 framed video? I've tried prompting it but it still ends up being a full frame
I'm still working on a consistent Bigfoot character and I think I'm about to nail it this time. Stay hard!
I use clip champ to convert the image
free
I'm out of credits, so come back on the 28th so I can use the flow
You've got 8 secs to tell a story. Make that 8 secs like a brief glimpse of a world that seems to exist independently - a world that your intended audience wishes they could see even more of.
and yes you have to imagine yourself with the details of the scene and how it wants to be seen in 8s
also dont be naive no one is going to tell you how to make money because they are all trying to make money themselves
A shot of three men dressed in black military uniforms, carrying weapons, with masks on their faces, cautiously walking alongside a wall on a foggy night with dim lighting.
For some reason whenever I use jump to or extend, even while veo3 quality is selected, the next generations quality is always really poor
if I just generate videos in the main window, and use frames to video, or ingredients to video it stays solid, but the editor just seems to break things
Just read the article about the movie ancestra and there are several festures mentioned that aren’t available. Are those coming in the future?
top
OK, I’m generating videos in chrome on my iPhone and there’s no audio. How to remedy
I think fast might be better than quality overall
this is all fast https://www.tiktok.com/t/ZTjc3pFTQ/
Any suggestions for how to extend a clip and have the audio continue?
It's crazy, not even Gemini can help me with this one. B But, I hope one of you can help me with this. I believed that I am a PRO user. But, it told me that I don't have enough credits ?!?!
When will I get more credits?
In fact, If I can please ask, what is the credits system about?
u get them when your sub renews every month
if you go ultra you can buy as many as you want
As a PRO, how many credits do I get every month? And, what is the credit price difference between Veo 2 and Veo 3? Fast and not fast?
1000
it's a buck for veo 3 quality and i think twenty cents for everything else.. not really sure.. never did anything with 2
I'm imagining that each costs a different amount of Credits and I get a certain amount of Credits every month. 🤔
Do you happen to know where I can check my Credits balance?
click where it says pro
For Flow Veo 3 Ultra can Veo 3 Support be added to Ingrediants to Video because I need an AI video of myself for a presentation and sadly Veo 2 has no audio and it's overall worse. Also being able to add samples of a voice to replicate it would be nice (of course with all of the checks to ensure it follows T.O.S.)
nope
best you can do is upload a pic of yourself to whisk and have it create a detailed description
is there an issue with the website for flow? theres no buttons or anything to interact with, just says "Flow ,a new state of creation" and i cannot proceed
Its working fine for me.
im working on something funny (atleast funny to me lol)
@charred sail @charred sail Do you happen to know where I can go to learn about the Credits?
Guy's I need your help , I have a problem on flow I can't reach the veo 3 although I'm subscribed at Google ai pro plan any idea how to help please
where do you live
how to get video without watermark?
Why is the limit so low for pro members, it's only 1000 credits a month at 2 vids on the app every couple hours
is the sounds working alreay
@primal thicket anything referencing a minor will generally not have audio. I suggest 8-year-old Son would be the culprit.
Same
hi! can anyone suggest what video tutorial to watch for google flow? want to try this trend now. and i am starting to build skills with content creation and all related to it such as video editting
In veo3 my credits are finished I create a Gemini pro account for 15 month student but my flow credits remain 10 and not updated give me the solution
Can I assign a challenge, someone needs to remake this video with AI, PLEASE
https://youtu.be/WPuImU4l4qo
INDEBTED Among Us TwitchTV "streamer" Jerma985 shows off his DUMPTRUCK ASS to the tune of MOZART!!!!!!
how to use veo3
I'm torn between what I want to make and what's easy to make...
Can flow be made available In india
So far I am loosing soo many credits because my prompts keep failing for VEO 3
I'm trying to make a fun music video for my daughter of a ballerina and it keeps breaking thinking I'm doing something bad or something
Nice. Here's the link to the policies: https://labs.google/404
turn the kid into an animal of some sort
that's why we do sasquatch and yeti videos 😛
I appreciate the suggestion but that is dumb if they're going to block this. Some clips work, others don't. The problem is I loose credits along the way for failed generations
I don't mind if it fails, because I can try again, but I have lost thousands of credits trying to make this dumb video
Also this is the link to their policies that pops up: https://labs.google/tools/flow/faq
Which is a 404 error
are you doing fast or quality?
Quality
Fast don't turn out right to do realistic looking footage
sometimes when you put a puncline in the clip it will add laughter or a rimshot on its own.. sometimes it's funny.. sometimes it's just wasted money..
Sure! that'll be 20 bucks
anyone know when flow will get added in sweden?
What tool do you guys use to remove random watermark images that VEO3 puts in videos? I'm sure Flow had one, but it isn't apparent anywhere.
Hi, I need some help with my Google One AI Premium subscription.
I just subscribed to the AI Premium plan, but when I try to use "Try in Flow" on the Veo page, it says I need to subscribe — even though I already have.
Also, I don’t see my 1,000 monthly credits anywhere in Gemini or my Google One dashboard.
I’ve made sure I’m logged in with the correct account, and the plan shows as active in my Google One settings.
Could someone please help me figure out what’s going wrong?
Thanks in advance!
Hello, can anybody help me please: I'm adding a video to the scene builder in flow but when I go to the scene builder, the video isn't shown to me, instead there is just a grey screen. Also I cannot click on the + to extend the added video (I'm using the ultra membership) please help
Same issue here, can some one help me figure this out?
hey there! thanks for your report, and welcome to the community!
Flow is not currently available in all countries where Google AI Pro is offered.
Please note: VPNs will not enable access in unsupported regions.
Before purchasing a Google AI Pro subscription, check the list of Flow supported regions.
If you’re in a supported region country and still experiencing this problem, reach out to Google One support.
With Google One, you can claim your new benefit or trial from the Google One app or the Google One website. You can also get notifications about the latest updates and offers from Google. Tip: Googl
i know right... so greedy.
Dear @main saddle, could you please provide some clarification regarding the absence of audio on Veo3? Furthermore, I am curious about the continued removal of credits when audio is not generated, especially considering that Veo3 is expected to include audio functionality.
I'm from USA sir...
Anyone else notice a significant dip in quality of video from the veo3 fast option?
Yes its So bad I cant use it. and added a sharpening function that puts white lines around everything.
the artifacting in scenes where the subject moves quickly is absolutely terrible.
and the MP4 / H264 Compression artifacting is similar to very early quicktime from the 1990's
Is there a prompt setting we coulld use to Increase the Video Quality upon Download? like video download A422HQ or something?
#flow-general・🌊 A muscular, intelligent-looking monkey wearing a black winter jacket, woolen cap, and aviator sunglasses is vlogging with a smartphone and a DSLR camera in his hand. The scene is set in the snowy mountains of Kashmir. In the first scene, the monkey walks through snow-covered pine forests, speaking into the camera excitedly.
In the next scene, he stops at a beautiful wooden roadside tea stall surrounded by snow. Steam rises from a hot cup of pink Kashmiri noon chai as he takes a sip, smiling warmly. The monkey looks around at children playing in snow and sheep grazing nearby. The background shows majestic Himalayas under soft sunlight and floating mist.
In an emotional shot, the monkey sits alone on a wooden bench beside the tea stall, looking at the mountains in silence, his camera hanging from his neck. He remembers his jungle life and dreams big.
The final scene shows the monkey standing on a hilltop, taking a selfie with the vast snow-covered valley behind him. He waves at the camera, ending the vlog with a wink and a smile.
Soft cinematic lighting, snowflakes falling gently, warm emotional mood. Camera includes close-up, slow-motion shots, drone-style landscape pans, and POV vlogging angles.
how can I add my image into the veo3 description to keep the person the same for each video? Please help
Hey guys i have question in Google Veo3 is there a way to get the consistent character in every scenes? In Text To video option?
Is there a way to mark videos as viable candidates for future use? Currently, I'm having to wade through dozens of iterations and re-watch each video to find what I'm looking for.
Seems a bit klunky but looks like you can mark a video by hitting the Add to scene button.
Use image to Video to help consistency. Mid journey has a method to keep the room items and character faces consistent - you have to use the (--p xmidvjc) and Seed #
upload and reference Images for characters and looks.
then use thise images in VEO3 as image to video
Ok guys.. I'm proud of this one.. https://www.tiktok.com/@sid_jenkins_atx/video/7519951126968536333?is_from_webapp=1&sender_device=pc&web_id=7515788093951526430
this took me 6 hours
yall fw it?
NOT SOUND, HOW TO FIX IT?
An 8-second Pixar-style 3D animation, rendered in 24 fps.
🎬 Shot 1 (0.0s – 2.5s)
Interior – Cozy Tree Hollow Living Room – Morning light streams through a circular window.
A middle-aged raccoon father sits comfortably in a gently creaking mossy wooden rocking chair. He wears round glasses, a tiny forest-made tie, and a neatly buttoned leaf-patterned shirt, giving him a dapper “ready-for-the-day” look.
He calmly reads a bark-and-leaf newspaper titled “The Daily Nut.”
From off-screen, his young raccoon son (around 8 years old, energetic and curious) shouts playfully:
🗣️ “Daaad, the boiled egg is too hard to beat! Help me!”
The father raccoon slowly peeks over his newspaper, raising one fuzzy eyebrow, confused.
🎬 Shot 2 (2.5s – 8.0s)
Smooth camera pan as the father raccoon sets down the newspaper, stands, and walks toward the kitchen (with a little tail flick for emphasis).
The angle shifts to reveal the kitchen interior: a snug space with tiny utensils, hollowed log counters, and berry jars stacked high.
Standing on the kitchen table:
A cartoonishly muscular boiled egg with twiggy legs, oversized red boxing gloves, and pumped-up egg-white biceps bounces side to side like a pro fighter, tiny shadow-boxing grunts included.
Its determined yolk face scowls slightly, eyes focused, ready for action.
Next to it, the raccoon son, wearing a leaf-apron, has his little paws raised in a “see?!” gesture, glancing back at dad with an innocent “I wasn’t exaggerating” look.
The scene holds for a beat on the father and son staring at the egg—wide-eyed, mouths slightly ajar, tails stiff in disbelief.
End on a comedic beat as the egg lets out a tiny “Hup!” jab toward the air.
remove the son/daughter references
don't refer to them as young
little, demure
How do you guys make video
also check to make sure you're using something that should give you sound
like new scene veo 3
Can someone make a video of a gorilla eating takis saying yeti
Come run that fade
An 8-second Pixar-style 3D animation, rendered in 24 fps.
Shot 1 (0.0s – 2.5s) Interior – Cozy Tree Hollow Living Room – Morning light streams through a circular window. A middle-aged raccoon sits comfortably in a gently creaking mossy wooden rocking chair. He wears round glasses, a tiny forest-made tie, and a neatly buttoned leaf-patterned shirt, giving him a dapper “ready-for-the-day” look. He calmly reads a bark-and-leaf newspaper titled “The Daily Nut.” From off-screen, a little racoon (around 8 years old, energetic and curious) shouts playfully: “come, the boiled egg is too hard to beat! Help me!” The middle-age raccoon slowly peeks over his newspaper, raising one fuzzy eyebrow, confused.
Shot 2 (2.5s – 8.0s) Smooth camera pan as the middle/age raccoon sets down the newspaper, stands, and walks toward the kitchen (with a little tail flick for emphasis). The angle shifts to reveal the kitchen interior: a snug space with tiny utensils, hollowed log counters, and berry jars stacked high. Standing on the kitchen table: A cartoonishly muscular boiled egg with twiggy legs, oversized red boxing gloves, and pumped-up egg-white biceps bounces side to side like a pro fighter, tiny shadow-boxing grunts included. Its determined yolk face scowls slightly, eyes focused, ready for action. Next to it, the little racoon, wearing a leaf-apron, has his little paws raised in a “see?!” gesture, glancing back at dad with an innocent “I wasn’t exaggerating” look. The scene holds for a beat with both staring at the egg—wide-eyed, mouths slightly ajar, tails stiff in disbelief. End on a comedic beat as the egg lets out a tiny “Hup!” jab toward the air. Sound is important
is it better?
take out the 8 y ears old
I mean you got the daily nut,, 8 years old.. and kids. c'mon 😛
Step into a world beyond worlds. Scrollforge awakens—a realm where magic and technology blend seamlessly, where scrolls hold the power to summon bots and craft your destiny. Forge your path, command legendary bots, and leave your mark in a saga powered by Abstract Creative Studio. Are you ready to shape the scrolls of fate? © 2025 Abstract Creative Studio. All Rights Reserved.
Scrollforge releases mid-2026. Prepare to forge your legend
Hyper-realistic sounds of biting into raw honeycomb — crispy, sticky, and layered with deep textures. No talking, just full macro food ASMR.
🎧 Headphones recommended.
#honeycombasmr #asmreating #crunchyasmr #eatingasmr #closeupasmr #softloops #notalkingasmr
NFTS will be limited addition minted for consumer commercial rights. every scroll minted has real world value and will be limited in stock
each NFT bought comes with a AI bot with the knowledge injected in it already for quick use for your business or content. stay tuned
I haven't posted this yet but I spent all night trying to make this video for a friend of mine (Tatum Lynn) who is a singer. This is her song. The chorus repeats, I showed it to my friend and she wants me to try a different approach making something more realistic and less "toy story", but I thought this was kinda cute. Took me A LOT of trial and error to get the clips to be similar and fit together using FLOW/VEO3 limitations. WOW!
I might still release this official once it's finished, but this is the "Work In Progress"
It already looks amazing 🙂
Thank you! This took me about 12 hours trying to learn how to get consistency between prompts and how to use flow. It was quite an education. I thought it would be much easier using "first frame" and jump to type features, but I found those features nearly impossible to use with VEO3 because each generation got progressively worse. So I had to start using ChatGPT to build an entire storyboard and description database to create precise prompts with descriptions of every single element. WOW
Ultra to expensive don’t we think ?
I think ultra is expensive for someone playing with it for fun. I think it's not that bad for people trying to make content for business purposes (potential revenue) or promotions like people wanting to make music videos or B roll footage for something.
I think I've spent about $1000 over the past few days trying to figure out what I can use it for. That's nothing compared to the $800k i've spent making music videos over my lifetime.
As Ryan said, it's a matter of perspective. Is $250/mo too expensive for someone that's going to sit there and make the 10,000th yeti video that's been made that day? Yeah probably.
But is it too expensive for a company looking at the cost of hiring Actors, VFX, Props/Wardrobe, Set Design/Location, Production etc. Not even close.
I just wish someone from GOOGLE would get back to me about credits they're robbing me of because I keep encoutering policy violations trying to make this ballet video
@silver crow You a DP in Arizona as well? The results you got are SICK! Would have taken quite the budget to do this in stop motion for real.
Just wait till you see what I'm getting for this live action version of the video I'm trying.
Siiick! Can't wait. I just had a client pretty much bragging to me that Veo3 was gonna save him from having to do a shoot in Florida we had been planning for a bit. Decided I had lost the battle already and made a little something with Flow today for the first time.
Here are two clips I have for the live action version of the video so far. It's REALLY HARD to make this stuff without triggering a policy violation. I get it, but it's painful.
That first shot by the stairs would have cost me probably $3500 especially getting the location, lighting, time of day right, warddrobe, actor, etc. The 2nd clip who knows how much.... ya right!
Are you adding the foley or is that Flow just adding it for ya?
@deft steppe Dude, I have had to go to EXTREME lengths, you have no idea, to get consistency in characters in the different scenes ... Over the past 4 days or so it's been a crash course masterclass of what works and what doesn't all by trial and error. I've made Google Gemini gems to help, I've used multiple different versions of chatGPT models to help me work through the prompts. So yes, the clips are EXTREMELY PRECISE
I have created an entire Profile for each prodomenant figure in the music video, and I have created strong language forcing ChatGPT to give me precise information about each character and scene and setting. It has been insane, but I think I've figured out a flow that is working now
well done! Nice to see someone else from AZ doing cool stuff
The biggest problem I am hitting up against now is the content policy violations ... it is extremely sensitve making anything with kids... holy cow. That's been my biggest problem now
@deft steppe what do you think of those clips of the live action version?
Those look pretty good! I think with quick cuts amongst performance shots of your friend singing would completely sell it as real footage.
It's password protected
Nice work The first one looks to have been mostly Veo 2 right? If you're interested in trying another method I built a prompting tool that I have linked in bio. Generating anything with kids is really a feat so nice work getting what you have so far, its definitely been a huge issue with many of the users in here.
Exactly. I think that's the key, there is about 2 or 3 seconds of usable footage per 8 second VEO 3 "hilucination" ....
I just got some pretty good results for an ad 'cold open' and they're about 8 seconds each in that vlog style everyone has been posting
But instead of bigfoot vlogging it's a pool guy
Is anyone getting Veo 2 or 3 to spit out 9:16 video yet?
The trick to getting prompts to work for children is in chatgpt copy and paste the Google AI policies into chatgpt and tell it this language:
Make the prompt adhere these policies and also know that the VEO 3 LLM is extremely stupid and take things extremely litteral so do not use language that can be interpreted with double meanings because the google VEO 3 LLM will interpret it to the most extreme and worst meaning possible.
Here's the link to the policies that i copied and pasted into chatgpt:
No, some people have had luck generating the video sideways but the prompting is too inconsistent to waste credits with, better to make something centrally focused that you can crop in.
Yeah, I built a gem in gemini pro, that was helpful. Theres a gpt that has been decent. It's really challenging for narrative work.
I'm doing everything with VEO 3. I haven't found good results with VEO 2 yet
Yeah I had initially tried training my prompting tool with evasive celebrity and child vocabulary but it turned out not to work that well. I just warn users not try to generate anything that breaks their policies if they can help it.
Veo 2 is decent if you reallly need a blending shot or to start from an image, and obviously dont care about the audio.
Once you copy and paste the policies into the Gem/ChatGPT session, it works really well.
The LLM for VEO 3 is VERY STUPID. I had words like "the moon dominates the scene" and it would flag it as a policy violation because you can't have something dominating a child in the scene. So you have to remind the Gem/ChatGPT that the LLM is really stupid and litteral and to avoid any word that could imply a violation.
Yeah I did do that too, I also had tried to have it warn the user if the prompt it was generating received was potentially going to cause an issue and it was a bit overzealous.
Theres definitely a lot of work needed with Veo 3's context and reasoning. You can create such a clear prompt and it just goes off the rails sometimes.
Another example is "her foot twists to the left" trying to explain a ballet move. Yah, so of course it doesn't tell you what flagged the violation but I figured out after multiple attempts I had to say it "pointed to the left" instead and that worked.
Obviously I wasn't trying to cause harm, but the LLM didn't like that language.. And of course it doesn't flag it until after it's done rendering for some reason, so you have to wait for like 2/3+ minutes before you know if it worked or not.
I can't see your video, it's password protected
I wish it told you what language specifically violated the policy instead of it being a guessing game
Yeah I think there's a lot more that can be done to address many user complaints or misunderstandings in real time... So many people coming in here with the same questions about Veo 3 and Flow.
I really thought this would be much easier based on watching the video. I think they need to be more clear up front that this is extremely beta the problem is the price per clip makes me think this is a commercial product and I should be able to create whatever I need. And if I had to upload a drivers license or something to have a filter taken off, I would do it. Their content has a watermark in there anyway so if you posted things that is really bad it would point back to you.
I was having this conversation with my son.
In the world there are pencils, paper, and paint brushes. They don't have content filters built into them. People could draw or paint whatever the heck they want and put it out there. Sure it takes a skilled person to make something realistic, so there is a level of restriction built in just because it requires skill.
But the public form stops people from doing this. Because it would point back to the creator. Sure there are sick, dark, disgusting places out there on the internet, and they're going to find ways to create whatever they're doing whether it's paper and pen or paint brushes or photoshop or whatever.
But if there is a digital watermark on the content that is created, and we all know it, why have such filters in place? The public form will judge the creators, and it points back to the creator. It's a little bit insane
Yeah
I just canceled it I can’t afford it right now. I wanna make contact and make money, but it’s a lot of work.
How can I get the attention of anyone from google to help me?
lol
There are cheaper things out there that are amazing. I found this place that is cheap monthly and seems to do pretty good stuff for the hobbiest. https://www.freepik.com/pikaso/ai-video-generator
Yeah but I did as I made an account a big foot vlogs I gained over like 500 followers on my it’s just I wish I could get them where they blew up
You know
@tough gust @elder crow @iron matrix can you help me? I'm stuck.
Do the google people ever interact in these channels?
Woops. Now try
It says "sorry can't find the page", maybe it's a new link?
Let me try again. You’re probably right
I used VEO3 in one afternoon. This is rough right out of the program. Lots of iteration. It's as rough where you'd expect it to be to build a narrative.…
we read but normally focus on the bugs-feedback section. Def open up a ticket there if youre having issues or have suggestions!
I would lean away from directly pinging unless you have waited for a day or so and seen no response
Good heads up. I posted something late last night. I've been spending a ton of cash on these credits trying to make this video, I really need someone to watch my prompts and tell me what's triggering them to fail beacause I'm not doing anything evil and I'm getting super stuck. This is a really expensive product to not have any "support" to know what's going on
I wonder when flow is going to get veo3 fast to extend videos
Amazing.
Thanks . That was first weekend veo3 dropped. Beifeni had built any prompts builder in Gemini or gpt. Was trying to use frames as much as possible
dumb q, if i choose 4 outputs will it cost 400 credits to generate, right?
Yes
After wasting so many credits figuring out why my videos showed a "Video audio generation failed". I decided to switch from Chrome to another browser & now they have sound. Just an FYI for those with the same issue, change your browser.
Let me try, because i've been having audio failure issues too for last few generations.
@zealous sable turns out it doesnt safari, which browser did you use?
Unfortunately, Microsoft Edge
Hi
anyone know how to retrieve a deleted FLOW project?
Can you use Veo 3 on a tablet? (Just ordered a Samsung 9+ 11" tablet for work)
Dude... sometimes my projects just disappear. That's why I download each clip now as soon as they are done.
Edge does seem to work better. But, I also learned (after wasting 500 credits) that Veo is REALLY picky about what you say in the prompt... so freaking annoying.
mann, literally spent a week on it, going to try to find support email. Thanks for teh tip, will deff download each clip from now on
Yes x4
Good luck getting in touch with anybody. I couldn't find a way to talk to them after 3 days of trying.
I've come to the final stage of my first prompt :S
Download every scene you are happy with as soon as you get it.
New resource from the team.
Hi
can veo2 or 3 use negative prompts?
yes
followup question... how?
my guy wont stop gesturing with his hands lol
❤️ finally new content about flow
i thought jump to works with veo 3?
Plzz VEO 2 plzz
In 3 months... we're all going to think what we were working with today was stupid. lol
and who's this "we"that you were referring to 👀
I was referencing all of us. But, you don't have to be a part of it if you don't want to be.
I was talking about how fast things are advancing... that's all.
I'm hoping we get extended lenth scenes. And, i think that will happen soon.
So... you're spamming a scam?
It's in the prompt... tell GPT to tell the prompt to NOT gesture with hands at all.
You know how to make those Bigfoot vlogs
Like I be creating sometimes and it’s only him talking not like yeti talking too
ya, tried that, doesnt seem like negative prompts work in the traditional sense. cheers
Give me your prompt.... let me see it.
The man speaks directly to the camera without moving his hands............
i can ellipsis too
Gotta get WAY more sophisticated than that...
im all ears
Send me your prompt... i'll make the video... I have 24,000 credits.
I appreciate the offer, but this is confidential IP
10-4
ill keep hacking away, thanks m8
I don't know if anything in this would help... but read what it says and seee if any of it could help you.
Create a cinematic video of a man speaking directly to the camera without using any hand gestures at all. The man should remain completely still from the shoulders down—his hands must be visible and unmoving (e.g., resting on a table, folded in his lap, or by his sides). His facial expressions can be natural and expressive, but his arms and hands never move or gesture throughout the video.
The background can be neutral and uncluttered—such as a plain wall, office space, or quiet outdoor setting. Lighting should be soft and natural. Camera framing should include his upper body (at least from the waist up) to clearly show the absence of movement in his hands and arms. The tone of the video should feel controlled and intentional.
Important note: Avoid any direction or framing that might cut off the hands or allow for hidden movement. The visual focus is on intentional stillness and control while speaking.
Why ppl love to create Bigfoot vids now
It will ge old and ignored soon... just like the baby vids that went viral.
Is it their fetish or something ( I’m not judging
I mean I’m judging but hey
We listen and we judge ✨
Google the moment they announced Flow.
People : 🐒 🦍 🎥
Idk where did u get the idea that negative prompts works in “Flow”
There’s no negative input box even
and you didnt help
Why would I
why are you even in here lol
You misunderstood the whole thing I said lol, I said flow doesn’t have the negative prompt input box. Like in this screenshot
circling back to let you know I got the same result, but i do appreciate the help!
Negative prompt is for API only
It doesn't need a box. You just tell it what you "don't" want it to do followed by what you "don't" want it to do.
that = negative prompt.
think they are missing the point...
No need to be passive aggressive and look far too deep into things lol, have a great evening 👍
I think you are right.
I think i got it... "a guy with no hands is talking to the camera" haha
If I had every credit back for every prompt I tried... lol.
he cant use hands if he doesnt have any...
I'm not trying to steer you wrong... I was just trying to help.
I know you can't share the idea/prompt... so, that makes it hard to help.
thanks again for the help - was able to solve it by removing his hands , can't use them if he doesnt have any
ill definitely keep your advice in mind in the future though
Nice
and maybe even to wrigleys point, they will add a negative prompt field for us to input vs prompt crafting
You don't need a negative prompt field... just tell it "don't do this."
Example...
I typically tell it... "no music, no text overlays, no graphics, no captions..." etc.
This is a better example...
Question... are you animating first frames or animation from just text?
All from text prompts....
I think thats where the issue is, all of the ai gen platforms have much more control with natively generated media
MJ, Runway, Veo, etc...
To be honest... I am not sure what you're saying or what you are saying the issue is.
is there a good prompting guide for veo3? my videos never come out very good
It takes me 10 attemps (average) to get what I want... but, I'm getting paid for them, so I don't care.
I'm touching on what I believe to be the missing link for true character consistency.
You can get pretty good character consitancy if you work hard enough on the prompts. It's not perfect... but most people wouldn't notice the difference.
no doubt, but for my usecase consistency matters a lot - so my workflow is pretty lengthy but necessary
but we are talking about nothing... this will be solved within the next year
custom character/audio models
Yeah... that's gonna be tough... (I think Veo will fix that for us soon).
My clients want short social media clips... and "close enough" is "close enough"... if that makes any sense.
You said it might take a year.... I'm betting within 3 months.
Ok... I took something like that. Created a "project" in GPT. Added it to the instructions. Now, it follows what I fed it with every prompt.
could only hope... i'm currently moving my workflow over to comfyui so I can swap out the nodes as the tech advances
your poor eyes. put that thing in dark mode 🙈
It's over 3,000 words of instructions. lol
I canceled Hedra and MidJourney (after having both for long time). I might get Hedra again for log format stuff.
My stack is...
GPT
Veo
Riffusion
Photoshop
Premier pro
flux.1 Kontext{max} is also magic
go check it out if you havent already
think its on a few platforms right now
Flux 1 was on Hedra... right?
yes, but this is a multimodal variation
Hmmmm
it uses "kontext" lol
FLUX.1 Kontext is here 👉 https://ltx.studio/
FLUX.1 Kontext is now live in LTX Studio. This state-of-the-art multimodal model brings powerful new capabilities to your creative workflow—including precise inpainting, stylistic control, and seamless consistency—all driven by natural language.
In this tutorial, we’ll walk through how to u...
I haven't seen anything I can't do with my stack yet...
Long form is the only thing I like about Flux... but it limits you to motions and cameras.
I gotta head to bed... ne back tomorrow.
Hi Hailey, what issues are you having with your videos? I have a pretty good prompting tool that people here are finding helpful. If you'd like to give it a shot it's in my bio.
As for prompting, the tool will help nail the format, but refining the content of the prompts will always be up the the user. The better you can describe your scene and wishes initially the better it will do. Understanding Veo's strengths and weaknesses is important. I have some help info available there as well!
hey is it meant to store character descriptions? cuz mine keep disappearing
check this one out. my best yet. https://www.tiktok.com/@sid_jenkins_atx/video/7519951126968536333?is_from_webapp=1&sender_device=pc&web_id=7515788093951526430
Yes, the characters should save, you're signed in right?
yeah i sign in
refresh every session before i start
I just logged out and back in.. the last one i did is still there
Hmm, I'll dm you so we can keep it out of this chat.
sure
Hi iam use AI ultra
When i do extend on video on flow with veo3 why always no audio? And how to fix that please ?
Hi, Veo is not yet available in my country so I can't test it, but can you use it to reconstruct half of a video ? Let's say I want to use a full video cut in half and want to try what veo would guess of the other half?
when i try to extended a video i made with veo3 quality it only lets me extend further clips with veo2 fast is there a way to fix this ?
U need to change prompt sometime it doesn’t work with some prompts
gotcha, any specific key words it doesn't work with?
i switched like 6-7 prompts and keeps giving the same error
consistent character
Same question for me. Does a person need to be subscribed to anything to be able to extend with veo3?
i have the highest package and i can’t
so i don’t think so
What?
How do I generate a video from a photo on flow?
I can't see any way to do that with the mobile version
Frame to video
thanks, that's sad tho
yea agree
Extend currently only supports Veo 2
Google flow generate kese kare
Nothing to do with subscription plan. Currently only supported with Veo 2. Please check our FAQ for compatibility.
Good afternoon all 👋 Are people still having issues with generating videos with sound? I want to check before I waste any more credits and I have not heard back from support about the issue :/
is there anyway i can add videos to a scene without having to scroll and find each from main page
feel like scene building could be a little better tbh
Is there a view on when veo3 will become available to Belgium? And is there a way to add audio to veo2 clips?
In Gemini app, you can create 3 Veo3 videos a day, at least in Finland. Veo2 does not include audio, but you can in video editors add a soundtrack ofc.
Hello, could someone help me? I'm having issues creating with Flow, it's telling me i need a subscription when i already have one, and when i click on the link it suggests it says I'm already subscribed
there must be a way to extend with generated sound/voice?
Today a few hours ago, I used Veo3 in the Gemini app, and in two out of the three videos, it completely ignored my audio prompt and made up its own dialogue.🤷♂️
I suspected this might still be happing... :/ ive been using the main Flow app both scene builder and the new prompt option and ive been having issues with sound since yesterday and I'm now at the point where i dont want to waste any more credits as i will have to wait for a month for more credits with the only other option being to increase my sub and I'm not going to do that with no feedback on the issue and potentially wasting credits :/ if it continues ill cancel my sub after the free trial.
I am so confused, I have no idea where to purchase veo 3
Yeah, I don't know if it's a new issue or if it's just like that for this model and only now I came across it. To be fair, I have under 100 prompts worth of experience in Veo3, so difficult to say for sure on my part.
Vertix AI have a pay as you go plan I believe, but it get's crazy expensive fast, if you are not made out of money😂 I was stupid enough to try it with Veo2 and I set up the budget in Google AI Studio and I did not realize that the budget actually does not stop you from prompting your way to massive bill, it just sends few hours late a message that is like "btw, you went over your budget, haha - now pay us $500"😭
im very new to VEO too but yesterday i hit a point where sound was just not working and i was using the same prompt that i have been trying to refine and I didnt have any sound issues before
Okay, I hope it's a fixable temporary issue🤞
Not sure, what even is the best place to be lurking for these kind of updates... Just taking a look at this servers flow announcments channel now and then, if anything pops🤷♂️
I imagine its a temporary issue 🤞 ive posted here and also submitted feedback on site. They probably need more time to fix and or reply but as a trial user its not ideal or encouraging :/ I wonder how many people are wasting credits... will they have a way to refund any users credits that have been wasted? I hope so as I would expect that from a premium sub service especially for those thinking about upgrading to a higher package. The budget issue reported above by you too... I would refuse to pay that and ask that the feature either be refined or explained better as I feel that might be misleading..
whoa we can buy credits now!
Imo, it would be fair to give credits back to users, if there is a new bug as a result of some update they've done and people are losing credits in failed attempts as a result.
yeah I just lost 20 right now
my video was at 99% and got stuck, I refreshed and lost he credits and no video.
I had this issues at least once yesterday but I think I was refunded automatically for that one
Please use English only in chat. Thank you for your understanding! 🙂
hello, iam still subscribe but when i go on flow, i cant use it, the site told me to subscribe
Veo3 is actually laughing about me, i'm pretty damn sure 😂 Doing a video with "things you can do in under 8s" , wanted to use the USB stick meme. Asked it to create a video where someone succeds with an USB stick on first try, guess what the person in the video FAILED to stick in the USB. Changed the prompt to "someone fails putting in an usb stick" and in the video the person succeeds in 3s? 😂 
I'm so mad i cant generate audio, tried about 6 generations already with different prompts and audio does not come at all. its totally mute. am i the only one with this issue??
<@&1173646484311384114>
Hi guys, I just started to use flow and it is amazing XD is there any place to learn how to use it at full potencial?
In Flow, Click on "Pro" (or what ever package you subscribe to) next to your avatar. Click on the credit count. Then scroll down. It will show your credits used and credits refunded
DO NOT include any text referencing a minor
@tacit halo i generated this video with technically a minor, given the perspective etc is that ok?
The veo 3 model will generally not produce audio for prompts that reference minors. Its pretty strict but sometimes you can get lucky like you did.
In order to not waste credits, its best to avoid it if possible.
Is pro still a good option ?
why
yes
pricy
im just wondering if pro will work
Hi Sid,
I saw that you were using "Ai Remove" in capcut for the subtitles. Saw that on youtube also but event with the pro version I don't see the feature while looking for using it.
Are you still seeing it on your side ?
Is anyone else having the issue ?
8s is not much you say? There's a lot you can do in 8s 😄 😂 😂 This only took 6!
How to get consistent character and consistent voice in veo 3
Did you manage?
An anime boy saying "my name is Michael"
Yeah..I think that's enough consistency for a main bigfoot character. Take a closer look, my friend.
Hey, is anyone else having issues with Veo 3 Quality not generating audio anymore? It was working earlier, but now none of my clips have sound, even with Beta Audio turned on. Any fix for this?
I am also having this same issue. Did you find a solution??
hi, i recently just subscribed to google ai pro, but when i tried going on flow it says that i must be a google ai subscriber to generate, although i am a subscriber 😭
when i go on this subscribe button inside of flow, it redirects me to this page from google saying that im literally already subscribed
what shuold i do?
its also the 2nd time this happens, my brother tried it as well on another account before and he was getting the exact same stuff
ask gpt to do it. Then tell it to make it better and more detailed.. then use it every time
I don't use quality anymore. 80% of the time it was lower quality that fast 😛
my standup video was 100% fast mode
can you share me your prompt?
@iron matrix
can you share me your prompt?
Turns out it was because i had the word "kids" in it. I replaced the line "kids cheering up their grandma" with "friends cheering up their grandma" then it had the sound come in,
Best to open a ticket in the flow-bugs-feedback section.
1: are you sure youre an AI pro subscriber? it should say here: https://one.google.com/u/1/plans?g1_last_touchpoint=39&g1_landing_page=1
2: are you in one of our supported regions? https://support.google.com/labs/answer/16353544?sjid=7912190084831390621-NC
Google One has cloud storage plans for everyone. Discover what each plan offers, including access to Gemini Pro with the Google AI Pro plan.
We’re working to bring Flow to as many people as possible. See below for a list of supported languages, as well as countries and territories where Flow is currently available to users ages 18+. Suppo
yes i am an ai pro subscriber
and yes im also in one of your supported regions
should we switch this over to flow-bugs-feedback section or is this fine
if you’re in a supported region country and still experiencing this problem, I'd recommend reaching out to Google One support 
With Google One, you can claim your new benefit or trial from the Google One app or the Google One website. You can also get notifications about the latest updates and offers from Google. Tip: Googl
LOL just did a line of dialog where for some reason it whispered the first word
i should try that on stage.. just whisper random words
if youre using the RON that means youre based out of romania right. I don't think that is a supported country sorry!
i will take a look into it, thanks
is there a rule against using a remote server based in the US to use it?
i am living in the us, but i chose to pay using my brothers debit card and he lives in romania
so im guessing thats why its the ron currency there
VPNs will not enable access in unsupported regions - before purchasing a Google AI Pro subscription, check the list of Flow supported regions.
Not a vpn
virtual server located in the US
they're like $5
you log in use the desktop
goggle thinks your from the US
is that ok?
asking in an official community channel if its ok to get around limitations and break the rules is funny 🤣 🤣 Even if its working or ok they likely wont tell you
I am a comedian 😛
and i just figured out over the shoulder with dialog
last thing i need to figure out is a consistent audience
but taht's a lot of people
so for now
i'm gonna switch from right to left with the front
should confuse them enough
different angles
they'll see the same part of the crowd every 24 seconds
How often do ppl generate a video that does not have accompanying audio with VEO3 and VEO2 . I've got about 20 generations in a row with no audio, feels like I'm in a rut and wasting credits
anyone know of a solidly consistan way to get dialog to start at a specific time..
they tend to start fast on me
like i want all my standup to start 1 second in
I do 6 seconds and have 1 second in and out
gives me room to cut
and yeah i've tried.. at the 1 second mark he says.
Why is audio not working for veo 3
I used a few different prompts but one i tried to use was " make a man say "Hello my name is Jeff" "
Not quite "specific time" but adding some kind of action/motion seems to help give it a gap
It has been going in and out for me aswell
that's not consistent either.. tried it
yes also i cant and someone tell me can we make vedio here private im new to discode
I tried (beat)
I already wasted like 300 credits and i can't get any audio this is frustrating :/
can you tell me how you create this vedio in here?
first off, ask chatgpt. we all use gpt. ask it for a prompt. ask it to make certain areas more detailed. add details to the character until you get the same character every time.
but it does things that aren't supported
like accents
so it's a learning process.. you learn to take that out
but then you'll learn that other things actually give them accents 😛
like for a female character add the description, with a Canadian lilt
for british just reference an actor
I almost always get audio if I specify something like:
The [character] says "some dialogue" in a [emotion] tone
Haha me too. very annoying. I tried like 6 time and was getting all pissed. It randomly started working for me again though.
I was as well and it was still not working
I haven't had single issue with audio since day 1. It's usually user error.
Yeah I understand that. but how do I get around that
I can't do even something simple like : man saying :"hello how you doing"
It still says error to generate audio
depends on what your doing
are you being downgraded in flow?
Bro I literally tried the same thing 😭😭😭😭
trying to do something unsupported?
No
It didnt work even under the most simple circumstances nor did it work under a detailed prompt
he was supposed to sigh then deliver the line...
is there a way to force NOT subtitles generation?
I also would like to know that
nope
💀💀💀 lame
although
?
try fewer words
i have a theory its a wpm thing
like it can't figure out if it will get all the words in so it does the subtitles
i do a 1 6 1 rule
sometimes i didnt generate a dialog, but stills appears some random text on generation. I use "Do not generate any subtitles or on-screen text." as part of prompt
do you have an example prompt using to test without modifications?
god damn it, sometimes Veo is puuure evil magic. Want to know the full prompt for this?
-> A character from the scifi game Wildstar suddenly realized the player is no longer connected, a stylized sci-fi world inspired by Wildstar, with bright cartoonish terrain, floating crystals, glowing tech structures, and distant stylized mountains
nice
how do you write your prompt in 1 6 1 ?
@charred sail is it fine if i shoot you a dm?
sure
i dunno.. just speak it and think about how long that took? lol you'll figure it out.. trial and error
you got 8 seconds. don't cram 15 seconds worth of dialog in it
gpt can tell ya too
ask it how long it takes
ah yes. prompting only short dialogs.
thanks sent
got it.
check this out
i cant generate audio video with single prompts.. stills generate action but no audio. Check this prompt: "A 30-year-old male person standing in the middle of a busy street in downtown Chile during daytime saying only once time: "¡YA CONCHETUMARE!" audio and dialog."
Is it just me that gets demonetised on Tik tok for using ai videos?
Hi has anyone run into issues with projects getting 'blacklisted' so the prompts fail even though they are for innocuous things
i dont, but when use female as characters audio is ok, but with male doesnt work
My first video I created with the free trial credits,
Can't post it here, since it is 104MB (Discord doesn't allow posting videos over 100 Mb)
Titled: Mongols at the Door of Egypt in 1260 AD
A 2-minute video short film
Hello, why when I use Flow, it tells me that I need a subscription, even though I have one?
Maybe you have utilized all the free credits / or the Google AI pro plan credits which is 1000 credits per month
Click on your profile whilst in the dashboard of the Flow
You will see how much credits remaining
hey, welcome to the Labs community! Flow is not currently available in all countries where Google AI Pro is offered.
***Please note: ***VPNs will not enable access in unsupported regions.
Check the list of Flow supported regions before subscribing!
If you’re in a supported region country, please reach out to Google One support, they'll be able to better assist with this!
We’re working to bring Flow to as many people as possible. See below for a list of supported languages, as well as countries and territories where Flow is currently available to users ages 18+. Suppo
With Google One, you can claim your new benefit or trial from the Google One app or the Google One website. You can also get notifications about the latest updates and offers from Google. Tip: Googl
this doesnt work for me, only generate plain videos no audio
@main saddle do you have any roadmap or info when more countries like France+Germany are being added?
thanks!!!
no immediate news to share yet, but when we expand to more countries, we'll share in #flow-announcements・📣!
Do you know at least if we talk like couple weeks or rather months?
Gpt and/or Gemini can craft exact 8 sec dialogue within prompts. Always works for me.
anybody know how to generate a prompt for a cyberpunk drone jet flying over a medieval kingdom launching missiles at the buildings? I wanna make it a top gun sort of shot where it's like the camera is following the jet but every time I do the video either has the jet floating doing nothing while fire is going around or the jet is flying backwards.
ssup guys, anyone already tested including negative prompts on flow prompt? It worths?
This is most likely due to content policy limitations and the way the model was trained. Most generations I've seen that try to include violence other than hand to hand combat fail. Missiles don't fire correctly, guns don't fire at targets, blades don't cut targets or move as you'd expect. I would assume that the model hasn't been fed (justifiably) a bunch of real combat footage to accurately depict these scenarios.
Ok so what can I do?
Edit around the limitations, get a bunch of sweet cyberpunk jet shots, including one of it arming the missiles or something, then cut to explosions in the village, rubble flying, etc. let the viewer make the connection.
So many people want the model to make a perfect 8 second cohesive shot that tells the whole story, and sometimes it's too much to ask of it. The amount of "failed" generations that actually have 3-5 seconds of a usable clip just being thrown out, or that can actually take your story a new direction, is silly to me. But people like wasting credits I guess.
any good videos explaining how to setup your prompts?
I have the template and have been playing with them in flow but I want to understand a bit more of what im doing.
first attempt
generating 2nd try rn lol
A futuristic, sleek cyberpunk drone jet streaks across a dark night sky illuminated by the pale moon, flying low over the ancient stone walls and torch-lit roofs of a sprawling medieval kingdom below. The camera follows the drone in a dynamic, tracking shot as it rapidly descends, unleashing a volley of glowing missiles that impact several key towers and structures in rapid succession, causing fiery, debris-filled explosions against the historical architecture. The visual style is a high-contrast, cinematic blend of gritty cyberpunk and fantasy, with the scene filled with the intense sounds of the drone's hum, missile launches, devastating impacts, and the crackle of flames engulfing the old buildings.
3rd try
A sleek, dark cyberpunk drone jet tears through a dramatic twilight sky, captured in a high-octane cinematic tracking shot directly following its relentless path. Below, a sprawling medieval kingdom unfolds, its majestic castles and ancient stone buildings steeped in the atmospheric, fading light of dusk. As the futuristic jet screams over the historical landscape, it unleashes a barrage of glowing missiles downwards, which streak with intense visual trails towards targets below, resulting in powerful fiery explosions that rip dramatically through the old architecture. The visual style is hyper-realistic and intensely cinematic, underscored by the roaring engine, the distinct whoosh of missile launches, distant thunderous impact
Results might vary, especially if you're generating Veo 3 videos in Gemini. Gemini auto-enhances prompts by default
Just my take based on what I saw in the official Veo 3 API docs, looks like prompt rewriting is enabled by default for the Veo 3 model. I'm not sure if Veo 3 in flow got prompt rewriting enabled too
We that are not in supported Country we need something reduced the cost
drone
this is veo 2, same prompt
plese can you tell me where do you genarate this vedio directly in flow https://labs.google/fx/tools/flow or in discod in here?
Google Lab's AI video generation tool, powered by Google DeepMind’s latest model: Veo 2
plese can you tell me where do you genarate this vedio directly in flow https://labs.google/fx/tools/flow or in discod in here?
Google Lab's AI video generation tool, powered by Google DeepMind’s latest model: Veo 2
in Flow website only. Not on discord.
oky thank you and where are your country? becouse im in italy and its not supported
Malaysia
thank you dude
if you have Google Pro subscription.
try use https://docs.google.com/videos/u/0/ (Google Vids)
Access Google Drive with a Google account (for personal use) or Google Workspace account (for business use).
???? the huge font lmao
and in Gemini. https://gemini.google.com/veo
problem is they block evrythings i dont know why
however thank you so much for replying to me
i have google pro subscription but they bloked
Had to use a VPN to see the feature 😩
Did it work for you with Flow IA?
Yup i'm in EU
and what do you use or which is recommended?
Any VPN that can change your location to US i would say (personnaly have NordVpn)
Am I correct in understanding that the scene builder is supposed to preserve the scene seed so that subsequent videos maintain visual consistency? It's not working for me
From the Flow/veo3 launch video. I understood that scene was just a way to organize different projects and scenes for better organization. Prompt is what keeps the scene coherently consistent
Do you guys ever get perfect videos but the dialogue is off?
yep
put another quarter in the slot and try again
I'm working on nailing down 5-10 front row audience members to be the same each time i cut back to the audience
driving me crazy
close
i got the characters.. its just getting them to sit in the right order.. lol always something
please keep chatter in english thank you!
must be a heavy workload.. seems to be really struggling right now
finally got all the characters in the right seat.. i think it's just luck though. The question is how many time do i have to try each time to get them 😛
Been having a lot of fun with flow and made short about my current workflow.
Gonna start doing tests using only google products next I think.
Want some more Fresh Cake? Subscribe to the channel and check out some of the link below.
2 tries
Super realistic, I feel like 99% of people would assume that's a real clip if not looking really closely
Would probably be easier to get consistent clips by just showing the comedian on stage but looks great with the audience
hi, I'm going to forward this to #flow-videos・📹 where we like to keep images and videos created in flow! if you create anymore please post them in there next time so we can keep chat focused and you videos in the right place! :D
Actually, that makes it harder. You need a different scene to cut to.. different angles.
at least it did for me
once i get it nailed down it's on!
those pertained to the general discussion.. hard to have a discussion jumping about to 3 forums
I see, I wonder if just doing a simple red curtain background with him standing centrally , I know what you mean though the cuts are obviously difficult to line up and different angles makes sense to transition the cuts , you will get it nailed I'm sure!
I do greenscreen
slide a brick wall back there
I don't know if i'll get this one done today, but it will be my best one by far
gonna make him his own account
Blimpy Thompson AI Comic
Jim Jefferies used to drink beer while he did his act.. Blimpy is gonna take a burger break..
What is the official link to buy veo 3? There's so many and it's confusing
I just subscribed to google Ai to use flow and veo 3 or 2, but it says that i need to upgrade my plan, when i click on this link provided it takes me to google one site that says, you are already subscribed. I cant use flow. Do you have any recomendations?
you can use veo 3 in a few place, flow is my favorite https://labs.google/flow/about
but you can also use gemini!
I subscribed to Google Flow Veo 3 because I read online I can make videos with sound longer than 8 seconds long (up to 2 minutes). How do I do that? It is by default producing only 8 second videos...
So i been trying to reach cs for flow but no luck anyone know a way to reach someone ?
Im trying to get a conversation between two people in two different accents, anyone know how to fix it? wont do it
trying to do a british and an american and its very frustrating, wont do the american one
Easy once you finish the video click extend scene than click jump to scene and promp it i made a couple of 40+ second videos
yes i did this and its working but flow still not work and other can genarate vedios but without sound or vedio music why?
Thanks! If I buy Gemini pro can I use veo 3 or is it separate?
google one pro works!
hmm the videos should be able to generate with audio, if youre having issues could you open a ticket in flow-bugs-feedback describing whats going wrong?
Only some smallish % of videos should generate without video entirely
the problem i cant access flow but i have ai pro subsription its redirect to this link im in italy: https://labs.google/fx/tools/flow/unsupported-country and i can create vedios with ai studio, gemini, and vids but without music or soud only vedio can genarate.
Google Lab's AI video generation tool, powered by Google DeepMind’s latest model: Veo 2
https://support.google.com/labs/answer/16353544?sjid=16912860712475405381-NC I'm sorry were not legally available in italy yet 😦
We’re working to bring Flow to as many people as possible. See below for a list of supported languages, as well as countries and territories where Flow is currently available to users ages 18+. Suppo
I freaking love google
does it work now with voices or just plain audio?
hello ive had a lot of problems with flow not following my prompts casuing me to loose a lot of credits
🔥 "I WAS STOLEN AS A BABY – The Lie That Shattered My Life | True Story"
For 27 years, I believed I was Eleanor and Daniel Hartwell’s daughter—until I found a hidden box in the attic that exposed the terrifying truth. My entire life was a lie.
📜 What was inside?
A birth certificate with names I didn’t recognize
Pleading letters...
Awee one of them is named after me
is there anyway to get creddits back from bad/inncaruate videos?
no, they want us to waste all our money so we can spend more money on buying more tokens just to do more mistakes and then repeat the process /joking, but i might not be, who knows
i mean if flag output uses ai to determine if its the case and gives it back, then sure that would be great but idk if it does or not
Does anybody know how to download all the videos in the Flow Library? I have like 500 videos and it really isn't working well
I thought they would be in my google drive somewhere
Nice!
Thx
hi guys, is it possible to make character refence in generating video? i want to use my photo / face to generate video.
@inner hornet this is one i've been working on.
guys someone can help me ? i'm trying to do some videos with VEO2 in frame to frame mode. But i don't now why , when i get the last frame of the last video , the video be like "orange". Someone know what can i do to doesn't get this "orange skin" for my character ?
Check this out: consistent character, some voice enhancement and background music from other softwares: https://youtu.be/K4lhPZLJu1A
Meryl Bleats the over-actress. Barnyard goat humor!
You will have to use "Frames to video" and use the + sign to add images. Else you can do "ingredients to video", which I think is available only on premium plan
i made a video with flow but theres no audio on it, how can fix this
The captions (never spelled right) is really starting to cost money and super annoying.
Doesn't matter how harsh you are with the prompt about "NO CAPTION, NOT TEXT, NO SCREEN OVERLAYS."
It keeps doing it.
What?
I've found the frame to frame thing to be useless. It seems to degrade the image with each generation. It's probably because the video source itsself is h.264 at low bitrate and then the sampled image is a jpeg instead of a PNG which introduces artifactes, and then the result is gradually worse and worse and worse with each generation
@novel knot the only success I've had is using VEO 3 with very string prompts and descriptions for each of my characters or items or settings
@novel knot This guy has an interesting youtube & Google Gemini Gem workflow that seems to help a lot.
Here's a link to a youtube where he explains how to us it:
https://www.youtube.com/watch?v=NKmEJWuPP24
This is his website where you can download the gemini gem prompt from:
https://franklineh.com/prompts/zQInLMJPlXfkcKKekgvW
too bad i don't have premium subscription. I tried adding the generated video in the scene, then chose "jump to" it doesn't allow veo3 to continue. is this expected?
Try prompting like its a script in addition to using the negative prompts. I get subtitles in less than 5% of my generations.
try this -> Negative prompt: no subtitles.
Good morning All, trying to access flow but it says I am not subscribed however I am subscribed to google AI pro. Anyone encountered this issue before? Thank you
Are you in supported country?
Using vpn
maybe you have multiple gmail accounts and wrong one logs in to flow? Happened to me
No unfortunately not the case. Raised a ticket