flow-general・🌊

[brisk frigate]

how to fix texts

[crisp basalt]

Probably gonna take … Veo 5 pro max model to finally able to generate text on video without issue

[idle compass]

I think you are not getting an answer, as your question is hard to understand. Flow is used to create videos, not images, and Flow has never been able to create portrait videos, if that's what you mean?

[stuck adder]

Veo needs to be atleast 15 seconds

[tawny sequoia]

I had selected scenes in scenebuilder, 10 scenes, I tapped on the project icon next to "Scenebuilder" to go back to the page to select a scene I want from the 3 outputs. Anways, after adding that scene to scenebuilder, I tapped on scenebuilder to resume but all my scenes were just gone. Like really?

[umbral smelt]

Hello is anyone having issues with the frame to video cropping? The cropping size is not allowing the full image, this just started happening a couple days ago. Normally I can use a full image

[tawny sequoia]

Anyone has had similar problems? To where your scenes in scenebuilder gets wiped? This would be 2nd time I have to go back to select the scenes again in the project.

[umbral smelt]

Hello is anyone having issues with the frame to video cropping? The cropping size is not allowing the full image, this just started happening a couple days ago. Normally I can use a full image

[warped arrow]

same issue

[warped arrow]

i remember before there's like fit to size option

[umbral smelt]

Exactly! @warped arrow i thought i was going crazy. Is there any way to fix it? You basically cant use frames to video like this

[warped arrow]

there's a new message that will appear if you use "frame to video", i guess when they added that message, somehow this cropping issue got added as well, lol

probably need to report it as bug

[umbral smelt]

@warped arrow thanks, i hope they fix it soon. It really makes it difficult to use.

[warped arrow]

i reported it to bugs-feedback here, hope the devs will notice

[umbral smelt]

@warped arrow where do you report the bug?

[warped arrow]

on the left side here, there's different threads, like this one is "flow-general" , there's another thread called "flow-bugs-feedback"

[midnight temple]

can someone query something for me?

i spend my 100 credits 🙁

[fierce nimbus]

We hear ya. Stay tuned!

[fierce nimbus]

Fixing soon. Sorry about this

[worldly yacht]

hello Nick hope youre well 😄 i have a question about prompting if thats alright

[midnight temple]

its imposible to scroll horizontally

[stuck adder]

First frame to end frame generally have better quality

[warped arrow]

but yeah, before you don't need to do that, and you have the option to just upload the photo without cropping, and video will have black margins on the side, and sometimes, the AI will complete/extend the side also.

anyway, the devs are fixing it already, thank you devs

[tawny sequoia]

I forgot to add a scene in scenebuilder so i
I added it. 🤦‍♂️ Problem is, the scene is at the end of scenebuilder order. So I am trying to slide the scene in the order I want it to be, but it is impossible in Mobile .

[fierce nimbus]

Whats up!

[worldly yacht]

is there any trick in the book to make the flash move in the speed he does in the series ? ive tried EVERYTHING and nothing has worked 😄

[warped arrow]

is this a question, or just a statement?

[snow mantle]

تتضمن اللقطات في هذا الفيديو، الذي هو عبارة عن محاكاة بالكمبيوتر، شخصية أنثوية تشبه القطة وتحمل قطة صغيرة بين ذراعيها. القطة الأنثوية ترتدي فستانًا ورديًا فاتحًا. يتجولون في ممر مركز تجاري. هناك العديد من السيارات الملونة المعروضة على أحد جانبي الممر، والقطة الصغيرة تظهر اهتمامًا كبيرًا بالسيارات.
عندما توقفوا أمام السيارات، مدّت القطة الصغيرة يدها إلى إحدى السيارات الحمراء الصغيرة. بعد ذلك، بدأت القطة الصغيرة تصدر صوتًا لطيفًا يشبه صوت القطة الحقيقية.
إليك وصف للبرومبت الذي يمكن استخدامه لتوليد هذا الفيديو:
صورة: فيديو بجودة 4K لشخصية أنثوية تشبه القطة وتحمل قطة صغيرة، تتجول في مركز تجاري فاخر. ترتدي القطة الأنثوية فستانًا ورديًا فاتحًا. القطة الصغيرة تمد يدها وتلعب بسيارة حمراء صغيرة. الإضاءة ناعمة ودافئة. التركيز على القطة الأنثوية والقطة الصغيرة. الأجواء سحرية وجميلة.
صوت: مزيج من الأصوات: صوت أنثى تهمهم، مواء لطيف من القطة الصغيرة، وأصوات أخرى في الخلفية للمركز التجاري.

[fierce nimbus]

Super fast is challenging. I would try phrases such as “in an instant” or something like“superhuman speed”. Think of how else would you describe the action. No guaranteed method i know of

[worldly yacht]

ive tried all that :/ thank you anw !

[stuck adder]

It’s really easy

Is it moving too slow?

[worldly yacht]

Not moving at all , he runs and thats it , he doesnt get into superspeed

[stuck adder]

Show me how it looks

That won’t work unless he extends it 4 times and just speed up the footage in post with motion blur or motion trail

[nocturne otter]

have you tried making everything around him seem slower while hes running

[worldly yacht]

yeap doesnt work

theres wording to get around this , say Flash instead of barry allen , instead Tony stark say a man dressed as iron man

all my videos are superheros 😄

[stuck adder]

So what do you want him to do?

Are you using image to video? Or text

[worldly yacht]

tried both bro

[worldly yacht]

to run at hyperspeed through a city , the speedforce behind him , like watching a movie , ive managed to get it done on another app , but i want to get it to work on flow too

what sketch to command ? where is that

You see from the videos above how i want flow to show it , i cant get any results like that

because maybe someone else has a better idea and wording to try make it work , otherwise i wouldnt ask obviously

3 ideas , theres loads of people in here that might have a lot more ideas than 3 people

if you read properly , i said those ideas have been tried before , but anyway , why is this bothering you so much ? you ok bro ?

no ,im actually asking , are you ok ?

^dont give it half an attempt ^ do i really have to show much prompts / credits / videos ive tried ? maybe in here , theres SO many people ...MAYBE someone tried something good and it actually worked

so if youre going to act like a know it all , please dont , its only you and your 12 ideas , hundred of people might have something better to say and help

[kind bison]

That guy is a troll - ignore him.

[opaque sand]

For now, we have to deal with this nonsense. Quality is lost, time is lost. But deadlines are looming.

[vague jetty]

Guys, where i can edit already filmed videos with AI for advertising?

[frigid pier]

When do you think Flow will allow us to extend videos with voice consistency?

[stuck adder]

How is the quality lost it was the same thing

We just need to add our own black bars

Have you guy ever thought of rotating the image? Then rotating back after generation?

[opaque sand]

Every time you re-save a picture, the quality is lost.

[stuck adder]

Not all the time, but you would need to upscale it anyway cause nothing is USABLE

[opaque sand]

I already had a whole pool of images ready. Now I have to edit them again

[stuck adder]

You used those generations without an upscaler?

It’s not at all point yet where it doesn’t need an upscaler soon we hope

[opaque sand]

What upscaler are you asking about? I always upscale images to 2k. I never upscale videos in flow - I am convinced of its complete uselessness. I do upscaling in davinchi20

[humble trail]

why flow has limits?

[stuck adder]

I’m saying after the generation, you would need to upscale the videos anyway cause none is really usable without it

And not the built in scaler either

[fiery stump]

Looking for people who are experienced with VEO 3 that can prompt well and are able to help us creating videos for our performance insole brand. Paid opportunity. Just dm

[humble trail]

okay , how can i run it locally if i want is there any model open source that i can use to genereate videos like veo3

i tried that one it's too slow for me i had 5060 and it sucks on that.

[rustic temple]

Is it just me or does it feel like veo3 fast is slowly being nerfed or degraded as far as quality and reliability?

[stuck adder]

They def need an option to run locally on your computer

Especially the nano banna with all those censors lol

[vagrant laurel]

Google Ai ultra still have unlimited free generatations for veo3 fast?

[stuck adder]

I do wonder why they haven’t made veo 2 the free one

[warped arrow]

it is just annoying that something working well before, just somehow disappeared, and making things more inconvenient

[tawny sequoia]

Anyone has tips for importing audio files from elevenlabs into a Flow project? There are some scenes with no audio that I want to mask over with ElevenLabs Monologue that I downloaded.

[peak hatch]

Looking for an answer to this too, i havent done anything yet but so far it seems its best to use some lipsync ai tool

[foggy heron]

I’m not buying anymore credits too much money

[tawny sequoia]

I have my video created with some VEO 2 generations w/no audio. I figured I can use ElevenLabs to provide some narration to fill those clips audio absence. Basically I want to overlay the ElevenLabs audio that I downloaded onto the Veo 2 non-audio Clips

[peak hatch]

Yeah i want to do the same thing in some parts of my project but i havent gotten to editing the clips yet, so i havent looked in depth into the issue... could be a simple solution

Try googling lipsync ai might be there, i dont have the needed content to experiment yet

[kind bison]

You can't do that in VEO yet.

[stuck adder]

Interesting so the more people feed it the more the core model is learning

[boreal dew]

hello all, may i know is that ultra free to use veo3 fastmode now?

[stuck adder]

Nice I notice they allowed people now finally to be uploaded

And I notice image to video start frame to end frame is usualy a better quality I guess cause there isn’t any amount of guessing with colors/pixels

[stuck adder]

Yeah I rarely ever use text to video lol

They just need veo 3 start to end now

Well the open end still doesn’t give that quality that end frames does

That means it only guesses the pixels and colors you feed to it it’s not guessing

I find myself still upscaling with topaz

And I add grain later

[stuck adder]

Yeah I hope they bring out an open for 5sec aswell

Sometime it takes too long to reach the end frame

[tawny sequoia]

Created another video..I learned that Tiktok can help me input those downloaded ElevenLabs saved sounds into my Downloaded Flow Projects.

[midnight widget]

can I upload to adobe stock the videos I created on Flow VEO 2? I purchased the pro subscription

[rustic temple]

I hope they drop veo3 quality down to 20 credits per generation. Feels like they've nerfed the fast ever since switching to unlimited.

[halcyon atlas]

Hey, some of the clips I download in 1080 come bugged after the download in Flow, is anyone else having this issue?

[peak hatch]

@tawny sequoia I found my way to lip sync, I listen to my narator audio, see what he says in 8 seconds, write that in veo 3 so the character says the same thing.

Next i take my generated audio so i can have consistent voice and i cut it up for easier sync, i edit my videos in Da Vinci so i found a tutorial there.

Next i do this tutorial https://www.youtube.com/shorts/s3JGz1BtHtc it automatically syncs it by wavelenghts, if thats too tehnically u can manually match wavelenghts by looking at them

[heavy drift]

I used to be able to import 9x16 images into frames to video but now it wants to crop them all to 16x9. How do you fix that?

[winged canopy]

Ao

[stuck adder]

You have to manually do it

[heavy drift]

How so? When I upload and image it gives me a bounding box that will only do a 16X9 format. I can't seem to turn it to 9X16. I have been using flow for a while and I used to have no problem doing this until this week.

[unborn drift]

I can't select the entire image, I can only cut it into a rectangle, can anyone help me?

<@&1173646484311384114>

[kind cave]

Same issue as Cardoso here i worked around it by enlarging the PNG in MS paint until i could crop the part seems to be working

[versed peak]

What's going on with the audio generation folks??? It's like singing

[stuck adder]

Easy just create it manually in a program before sending it over

Wow I could image

Crazy

[wheat anchor]

Brand new to using Flow and struggling to get it to create a static shot - it always wants to 'dolly in' through the scene.

Below is my prompt. Any ideas what I might be able to do to make the camera/frame remain still?

A static, ultra-high-definition shot of the Amazon rainforest, as seen through a large invisible window. It's very important that the camera and frame MUST NOT MOVE - no zoom, no pan, do not dolly in, do not push in. It should look like someone just set up a camera in a rainforest and just left it recording with no interaction. The lack of camera movement is important to create the impression that we're looking through a window in a static building, out into the rainforest. Dense green canopy fills the view. Dappled sunlight filters through the trees. Gentle, natural movement in the foliage — leaves rustling softly in a light breeze. The scene should feel immersive, calm, and luxurious — like a serene window view in a high-end retreat. Rich cinematic colour grading. No humans, no text overlays. Intended for use as a video wall background in a luxury interior design installation

[vernal kernel]

Jesus christ that is f.ing annoying that when I try to get to another line by hitting enter, it starts generating when the prompt wasn't finished

[rich echo]

Adding clips to Scenebuilder - Is there any way in Flow to put 'perfect generation 1' and 'perfect generation 2' in the same scene without regenerating based on frames when trying to put 2 clips in the same scene? Not seeing this specifically yet in the docs so I'm guessing no?

[vernal kernel]

Yeah, doesn't come naturally on mobile.

Yeah, I agree with that experience.

[midnight widget]

Can anyone tell me if the videos created on Flow are free to commercial use? I purchased the pro subscription and I wanted to make videos for my adobe stock portfolio
If I upload the videos, will adobe ban my account?

I mean does flow veo 2/3 gives me license to use commercially?

[bright agate]

Does anyone know how to generate 9:16 clips in Flow. Any trick or something?

[wild jetty]

scene builder is not working properly

[flat hinge]

upscaled veo3-fast downloads are not usable. Please fix this.

[flat hinge]

hope it gets fix soon... I need to continue my project 😩 ... breaking the flow ... maybe Google should buy Topaz Labs and offer it as a value bundle with the Ultra plan and I'll keep subscribing 😋

[serene flume]

I guess we are all having similar issues. unable to download normal Upscaled videos on my end too

[dusk wasp]

Has this happened to anyone else? I created this scene and then wanted to make some adjustments, but now it won't let me. I get this error message. Obviously, I'm using a free account.

[flat hinge]

I just used Topaz upscaler via Fal. So far it works the best for me while waiting for Veo's upscaler to be up.

[molten seal]

I understand your right to express yourself but you should think about it twice. That's creepy.

[abstract pebble]

editing, I'm going to make a longer version of it and keep on improving the prompts, it should be a bit creepy as I'm going on a video that has a style of a horror movie trailer

[midnight widget]

Isn't this one from whisk? I search through Flow faq panel but I didn't find anything. Do you use your videos commercially? @crisp basalt

[crisp basalt]

https://labs.google/fx/faq

[midnight widget]

Ayeee! So I can use those videos and can upload to adobe stock as they are my original content, can I?

[midnight widget]

@crisp basalt brother

[crisp basalt]

yes

[midnight widget]

Thanks mate

[crisp basalt]

make sure to label it as generated ai content or somthing

[midnight widget]

Yes definitely I will

[midnight hedge]

Is there currently a server side issue with upscaling? Apologies if this was already answered somewhere.

[peak hatch]

I have issues with upscaling too

[twin mortar]

Same issue here 😬

[midnight hedge]

The race baiting creator to rethink his actions before trying to post a viral clip fooling boomers for clicks.

[abstract pebble]

It’s not about race, it was definitely inspired by a REAL event!
I’m not rethinking anything it’s about creation edits and suggestions, if you have nothing good to say better not to replay at all.

[raw oak]

Hi upscaler is broken

[feral zodiac]

Anyone in UK know when we will be able to use realistic pictures of people?

[silk nexus]

Jesus christ that is f.ing annoying that

[abstract pebble]

Yeah we know it’s in the bugs channel

[abstract pebble]

You are entitled for your own opinions and I respect that ,you don’t see any value in it but I do.
I think awareness is a very important value especially if it’s something that got very little attention imho.
Ill say again that we here to discuss about flow related topics and my message was clear about it from the beginning.
Let’s try to shift our conversation to related topics

[midnight hedge]

Granting you charitability you likely do not deserve, it's still not a good idea. Snuff films or even something leading up to it in AI are in incredibly poor taste and do nothing to "raise awareness". It's just fetishization of violence. Dropping the charitability, if you're going to operate with those motives just say it with your chest and don't be a coward making justifications. Just be honest ffs.

[abstract pebble]

I hear you, and I get that some people feel strongly against this kind of content. My intention was never to create “snuff” material or glorify violence. What I shared was inspired by a real event and aimed at awareness, not fetishization.
We may simply disagree on whether that has value, and that’s okay. I respect your view and I’ll take it under consideration in my future posts here, but I also stand by mine that shining light on disturbing realities can have meaning.
That said, I don’t want to drag this channel into an endless back and forth. The purpose here is to share creations, edits, and feedback. If you don’t see merit in mine, fair enough — let’s move on so the discussion stays focused on flow-related topics.
❤️

I think we’ve reached the point where we’re not going to agree. I want to be clear that my work was not meant to glorify violence or create a “snuff film,” but to raise awareness around something real.
That said, I also respect that others here feel strongly the opposite way. I don’t want to derail this channel any further. I’ll keep your feedback in mind and move forward focusing on creation, edits, and flow-related topics.

[flat hinge]

temp solution... FYI i paid $0.20 for one 8s veo3-fast video so it works for me now and when I really need. anyway, upscale is working now.

[fair gull]

I think Veo 3 is stuck in Broadway Musical mode. No matter how I word the prompts or ask for no music - I get my characters singing to a showtune. I think I've generated at least 50 clips and all have music and singing. Even when prompting no music and removing most of the prompt.

[fair gull]

[CAMERA MOVEMENT] – Static shot, no movement.

[SHOT TYPE] – Medium shot

[CHARACTER ACTION]

1. Judge Silverback, a distinguished silverback gorilla wearing wire-rimmed glasses and black judicial robes with white collar, sits upright in his leather chair behind the mahogany bench
2. He leans forward slightly
3. he peers over his glasses with a measured, attentive expression
4. He opens his mouth to speak, jaw moving deliberately as he addresses the court

[ENVIRONMENT] – Traditional federal courtroom with rich mahogany wood paneling, ornate carved details, brass nameplate reading "Hon. Judge SILVERBACK," American flag with gold fringe on flagpole, official court seal mounted on wall, warm incandescent lighting from overhead fixtures

[SOUND FX / MUSIC / AUDIO] – Courtroom ambiance with subtle air conditioning hum, distant shuffling of papers, muffled footsteps on carpet, leather chair creaking under weight, fingers tapping lightly on wood surface, no music

[DIALOGUE] – JUDGE SILVERBACK (deep, authoritative, measured tone): "Tell me about the day of separation."

[tardy ridge]

remove that "He leans forward slightly
he peers over his glasses with a measured, attentive expression" and reword it, his posture, etc, etc.

rty that but I have the same issues especialyl when it's something emotional

might have to change the dialogue

[slim eagle]

https://tenor.com/view/oh-i'm-gonna-cry-lala-ri-rupaul’s-drag-race-all-stars-s8e12-i-could-cry-right-now-gif-6558079541713868269

[fair gull]

Thanks, I'll try that. lol.

[fair gull]

Thanks

[crisp basalt]

below

in text file cuz discord nitro is whack

[wind grove]

@crude cloud is there a way to lock in a visual look for an app. everytime I try to add or adjust a feature it creates a new looking app.

@Dion Almaer is there a way to lock in a visual look for an app. everytime I try to add or adjust a

[peak hatch]

U can generate images first and go from there

[wind grove]

No I like images that i had. I just don't want it to change each time. for example. I am creating a calculator and each time I try to adjust what it does. It then changes the total look of the calculator. Where things like input fields are and colors etc...

[peak hatch]

Sorry i dont understand what you're trying to say, if you want to generate more images in a similar style u can try on whisk, you can upload up to 3 images in the side panel under style and try emulating the same style over again.

Also if you just want to change details i think they incorporated the nano bannana thing in whisk so once you generate the image and want to change something slightly you can click on it and say "Remove text" or "Change that to this"

[crisp basalt]

Make sure your prompt mention the calculator and/or any objects that’s in the image you’re using

[dark wigeon]

Lately, I’ve heard many people talking about starting dropshipping, but scaling remains the biggest challenge.
I went through the same phase—trial, error, and figuring out what actually works.
Everything changed when I focused on data, automation, and building a real brand.
With the right tools and ad strategies, growth became consistent and sustainable.
If you're looking to scale efficiently, I offer guidance and strategies built from real experience.

[flat hinge]

Nowadays Veo 3 hallucinates a lot via Frame-to-Video (Image2Video) ... but I'm surprised sometimes I get really good output even though it is entirely off script.

[midnight hedge]

It can certainly be a crapshoot, but I keep finding more ways to prepare my images better to safeguard against it. Prompt matters a ton of course, but certain visual cues can deinitely send it off into dreamland.

[tawny sequoia]

Flow scenebuilder should definitely be added on mobile

[midnight hedge]

You can always do it manually with a screen grab, but that is much more cumbersome on mobile I'd imagine.

[tawny sequoia]

I usually have to "open in chrome broswer" but sometimes after such a long time creating a prompt for text to video, when I am ready to create that scene, the whole page refreshes back to mobile and I just lose all progress on the prompt. Not even saved because it did not attempt to generate.

Funny thing is that most of the time, If i use Gemini app to create a video, it's easier to prompt and the video is very close to your vision of what you want to create.

[midnight hedge]

I personally don't use text prompting very often, but I can see that being convenient.

[molten locust]

hello guys

[crisp basalt]

Cuz Gemini auto-expands your prompt. In Gemini app Veo, they’re designed more for casual users and it doesn’t really require much knowledge or skill to craft prompts for video generation in Gemini app

[molten locust]

guys i dont know how to generate a video our company needs video generating tool can any one suggest me a while

[crisp basalt]

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

Anything else just ask ChatGPT or Gemini lol ..

[molten locust]

company ad i can i create through this video generation

[versed peak]

These audio errors are so annoying 😭 it glitches up every night

[tawny sequoia]

Because of the fact Gemini is a conversational AI model for "thinking" I assume. So like you mentioned, it doesn't require much knowledge or skill from the user. Makes sense

[midnight hedge]

Man, image to video animated fight scenes are tough. Finally got something usable, but sometimes you really gotta brute force it. Not sure what to do when it stops being free.

[kind bison]

When is 1080p generation coming to Flow? Already able to generate it VEO on 3rd party sites.

[stuck adder]

Says who? Does it do it natively? Or it’s being upscaled

[kind bison]

Natively. Google announced it today. I have already generated some clips.

But on Freepik and Krea

[stuck adder]

Where did they announced it??

[kind bison]

I saw the announcement on X

[stuck adder]

On flownygoogle??

[kind bison]

https://x.com/googleaidevs/status/1965160822260318702

Google AI Developers (@googleaidevs)

Veo 3 and Veo 3 Fast are now generally available in the Gemini API.

We're also announcing ~50% price reduction for both models and support for 9:16 vertical and 1080 HD outputs.

Read the blog ↓ https://t.co/55JTxWnQ0l

X · Sep 8, 2025, 9:10 PM

Not on Flow yet - only via APO apparently

Which is why I am asking about Flow's ETA

[stuck adder]

No freaking wayyyy

How does it look compared?

That means it should be on flow soon then?

If the api has it

[kind bison]

Looks a little better than 720 > 1080 upscale, but I haven't done enough tests with different input images to get a good feel for it

[stuck adder]

Wait is it only text to video 1080p? Or image to video aswell?

[kind bison]

I only do image to video

[stuck adder]

Awesome so it looked better you think

That’s a great start

[kind bison]

I would post a sample but the clips have all been bloopers - still working on getting a good one.

[stuck adder]

Show me a sample of 720 vs 1080p

And looks like 9:16 1080p isn’t supported by the API as yet only 16:9

9:16 goes up to 720 apparently

[midnight hedge]

Really threw it for a loop tonight. Knew it was a long shot, and wasn't incorrect unfortunately. Describing Portal portals in a way it understands is a near impossible task. Cool liminal space though. Hi doggie.

[kind bison]

After running some more tests, it looks pretty similar to a 720 upscale to 1080 - not a huge difference.

[stuck adder]

What’s the resolution of the input image?

Think it’s because it’s not on flow?

Is it in vertex ai too?

[kind bison]

Using a Midjourney image - 1456x816

[midnight hedge]

Whoa, think I just got held-up by Spider-Pig, or he really badly wants to give me that hotdog.

[stuck adder]

I’m using vertex studio now and have 1080p with 9:16

However it’s saying image to video not available for veo 3 only veo 3 preview lol

Idk what that means

Honestly I see what you mean with the 1080 it looks the same as the upscaler lol

The 720 version still looks better

[kind bison]

I need to do more tests using an upscaled image

[stuck adder]

It looks like they just upscaled it before spitting it out so it’s the same as the 720–1080p upscaler they have

[karmic kindle]

Not sure if this has been spoken about before, but are there any plans of letting videos be longer than only 8 seconds? I know you can expand them with the Scene Builder, but let's be honest. The Scene Builder is quite unstable. Often it will expand a scene, but will generate different characters or redo the first scene. All in all, it's not something I enjoy using that much. I'd rather make one video that's 20 or 30 seconds long.

[stuck adder]

Way too sharp

[stuck adder]

Just keep using start and end frames bro

That’s the only way for now

[kind bison]

Problem with both start and end frame method and the scenebuilder is that there is a gamma shift of the image at the transition, which is a major hassle to remove in post.

[stuck adder]

Ahhh yeah i noticed it too

Easily fixable, use smooth cut in resolve and it’s fixed

[kind bison]

Nah, doesn't always work in Resolve

[stuck adder]

It has for me everytime

[kind bison]

Has failed for me lots of times. Really depends on the look of the clips

[distant socket]

how use veo 3 in russia

[kind bison]

OK, just tried an upscaled version of my input image and I have to say that the 1080p feature is a fail. Too sharp, too much contrast, too blown out.

[stuck adder]

Yeah I agree

If you use 720 and topaz it gives a better image

Maybe it’s because it’s not directly from Google?

I doubt they’d release garbage lol

[muted matrix]

Well now that I've been using Veo 3 for months and I'm just loving it. However there are a few bugs here and there but its been good so far. Really waiting to try out the 'Last Frame Feature' and also the Ingredients feature for Veo 3. That I am really waiting for!

[rich echo]

Agree, still not understanding why you can't just connect gaps between pre-existing clips in your project in Scenebuilder. Its like shot continue builder not Scenebuilder if you can't directly work with your 'shots' and only generate forward from your current shot. Dying for Sora storyboard edit basics here.

[crisp basalt]

Is there any other ai video service that can do more than 10second

... so ... are we expecting something on september 16

😄

well not really, i was speculating if flow team is bringing us something from the store on september 16 cuz

ya know...9/16...

[crisp basalt]

but aint no way they are gonna, keep us edging that long

edging is such a weird word to use im so sorry

hold your cows or something

hold your horses

[stuck adder]

Did you try the native 1080p?

[crisp basalt]

its just okay, i havent notice any difference

[stuck adder]

Yeah at 720 it’s just a bit soft but it looks way better

You’d need to use like topaz and you good

[stuck adder]

What do you think is happening?

[crisp basalt]

Idk how to answer that question lol

[stuck adder]

I thought they were natively generating it? Or are they running the 720 through a upscale and calling it that

Yeah it’s super resolution isn’t bad

[iron agate]

Fix the image-to-video portrait crop bug! We're well into another week, and it has still not been addressed!

[obtuse tree]

New vertical video feature in VEO 3 is not availalbe for use on flow yet correct?

[crisp basalt]

not yet

[obtuse tree]

I think you have to use dev tools, right? Can someone with zero experience in dev/api figure it out, or just best to wait?

[crisp basalt]

dev tool for what

u can just wait until they announce something

[obtuse tree]

Freepik generates vertical video?

[crisp basalt]

oh u mean vertex ai

yeah its on there

[midnight hedge]

Anyone messed with Runway ML? I don't need the tools, but curious if there's enough time saving stuff to make it worth checking out.

Can't wait for seed control so we can someday generate AOV passes.

[crisp basalt]

yeah

its fine

i dont like it

but yeah

[midnight hedge]

If you can already do it in After Effects is it just a money pit then?

[crisp basalt]

i dont know i dont do any of that

[vivid fossil]

Why can’t I find the 9:16 option in Flow? Is it only available via the API for now?

[midnight hedge]

you're supposed to tell Veo to do that

[mighty mortar]

pls teach me

I joined a few minutes ago

[stuck adder]

It’s in vertex ai right now

[stuck adder]

I’ve used runway aleph it’s actually pretty fire if you know how to composite

[midnight hedge]

I do. That's what I would use it for since After Effects roto brush has definitely improved, but it's still tedious. Same worry across the board with all of this. What happens when the free and cheap generations run dry and we're all chained to the expensive generations and subscriptions. Ugh.

[blazing egret]

A bug. They mentioned it should be fixed by next week

[obtuse tree]

Yes. Ive been using whisk to create scenes in 9:16 and canva to put black backgrounds to make them 16:9 then upload to flow.

[rustic temple]

aww

[stuck adder]

https://www.facebook.com/share/v/1C7x5gacjs/?mibextid=wwXIfr

I was able to run aleph with this changing the suits

[blazing egret]

Thanks! I'll try the canva line

[stuck adder]

However it sucks at keeping the face so I had to rotoscope back on

Oh and I used topaz to upscale the 720 outputs too

[midnight hedge]

I'm probably alright giving it a pass until it's a bit more powerful then. Sounds like it wouldn't be much better at roto than Premiere's new automated one. Still beats keyframing masks for hours, until that's the only thing that works for some shots.

[stuck adder]

All they need to do is just turn it sideways though lol

And it works

[midnight hedge]

I've been depressed the past few years as Veo invalidates my decade plus of experience in motion graphics, but getting more into these circles and embracing the tech I'm realizing it truly is still a tool at this stage. If you don't have critical thinking skills or creativity you're still not going to make anything of value. Heck, even the wherewithal to ask Gemini for simple workarounds. Don't even have to think much.

[stuck adder]

Exactly for me AI is a tool

You still need a creative plan

That nano banana and veo3 combo is crazy

[midnight hedge]

Disgusting. Still have to use it in combination with Photoshop, but it's pretty bonkers. especially the image extension capability. Much more complex than Firefly in a lot of cases.

It does, with their aleph platform.

oh, yeah, no. I just do passes with lime-green backgrounds.

You can force Veo to generate with a green screen via text prompt or just have banana isolate the subject and then fill the background with green screen for image to video.

That's the kind of thinking you develop learning the old school way first.

[stuck adder]

Yeah I rotated mines sideways then generate then back to upright

But this was first frame and last frame

This was generated rotating sideways then after it generated it rotated it back

[obtuse tree]

Are you refercing the ability to create 9:16 videos, or the ability to upload 9:16 images?

[wild flicker]

Hey is 9:16 supposed to be 20 credits for ultra accounts in fast mode?

[crisp basalt]

its out ?

[wild flicker]

EU yes

[crisp basalt]

ooh

[wild flicker]

hopefully US soon

[crisp basalt]

wait can u screenshot how the aspect ratio option selection looks like?

[wild flicker]

I dont have it yet, but someone in my AI discord has it, says that 9:16 is 20 credits and 16:9 is unlimited but yeah he posted screenshots

[crisp basalt]

if u used chatgpt for this, i wish you would use chatgpt and ask them how to generate video in Flow website

[crisp basalt]

sickkkkk

[wild flicker]

But yeah hes saying 916 is 20 credits right now

even with start frames

[crisp basalt]

oh

aistudio got veo 3 now

nvm

[wild flicker]

haha yeah thats what he was saying

[stuck adder]

Yayyyy about time

What about the 1080p option

[crisp basalt]

..? just use upscale button lol

[stuck adder]

No I mean if the native that they announced looks better out of flow

[crisp basalt]

idk if its better or not it looks same to me

[stuck adder]

What this mean

Vertex?

[crisp basalt]

comparison
native 1080p
https://v3.fal.media/files/lion/NNCXOXuZWfEr0HzKSip-q_output.mp4

flow upscaled - https://cdn.discordapp.com/attachments/1347690644981940374/1414931353862148177/A_noir_detective_202508162125-UP.mp4?ex=68c15d02&is=68c00b82&hm=29c8b626d914433c12662b41973b584cd46d87ab38745d02e93bac9349d3ffb5&

@edgy holly for the image 🔥

[stuck adder]

Okay can you try it with real images?

[crisp basalt]

no

ye\

i rather just stick with Flow since its free

[stuck adder]

Ah I thought he used vertex okay

My bad

[crisp basalt]

no worries

[stuck adder]

I don’t see much of a difference there

[crisp basalt]

i forgot only 16:9 can do 1080p , not 9:16

[stuck adder]

Yeah that’s what I saw on the api

Yeah 720p does look better

It has a slight blur on it that make it seems real nice a slight

[crisp basalt]

(not my pic btw)

aistudio

[stuck adder]

I saw this on vertex ai last night

But I wasn’t able to select veo 3 for the image to video only veo preview

[crisp basalt]

yeah its not available yet

only veo 3 preview

the faq page updated

[crisp basalt]

Nice!!

[stuck adder]

You go it?????

[crisp basalt]

yeah its out

[minor storm]

????

[stuck adder]

Yes I have it too portrait mode

[minor storm]

how to do it?

[iron matrix]

ahh gross... We will fix this ASAP

[kind bison]

Portrait just popped up for me on Flow as well.

[stuck adder]

Dang, start to end frame isn’t available for portrait as yet

[kind bison]

Now, if Google will fix 1080 generations to reduce the extreme sharpness and high contrast, we will be cooking.

[stuck adder]

Yeah it doesn’t offer the 1080p with vertical

[minor storm]

Caméra GoPro fixée au bras massif et poilu d’un yéti. POV tremblé, style amateur, comme filmé à la main.
Forêt sombre et réaliste, nuit profonde, sol recouvert de feuilles humides qui se froissent et craquent sous les pas. Branches qui fouettent l’objectif, condensation légère sur la lentille. Image granuleuse, instable, typique d’une captation brute.

Le yéti, voix grave et haletante, parle en français familier directement à la caméra :
« Salut les gars… je viens de me faire un pote. Il s’appelle Tiago, il a 19 piges, il sort de l’usine. Il vient de finir de brancher les câbles. »

Coupe rapide vers un plan serré de Tiago : salopette d’usine élimée, épaules fatiguées, mains tachées de cambouis, câbles roulés autour de l’avant-bras. Son regard épuisé et inquiet fixe brièvement la caméra.

Ambiance sonore réaliste et immersive : souffle lourd du yéti près du micro, vent qui traverse les arbres, branches qui craquent, corbeaux lointains. Mix sonore brut, sans musique, donnant un effet documentaire viscéral.

🎯 Durée cible : 7 secondes maximum
📱 Format : vertical 9:16
🎞️ Fréquence : 30 i/s
🎥 Style : réaliste, immersif, caméra tremblée type found footage What's wrong with my prompt? I get an error every time

[midnight hedge]

A series of cursed misfortune told in 3 attachments.

[kind bison]

It's not on Flow yet anyway (1080p). I don't want to pay extra on 3rd party sites until the sharpness/contrast/gamma is fixed.

[stuck adder]

Gotcha

The portrait quality is pretty good

I wonder why they’re charging for the vertical smh

[junior horizon]

@main saddle thanks for the big news, we love you guys!

[main saddle]

🙂

[tame granite]

What’s the best way to generate good prompts for TikTok asmr? Is there a content creator or website that publishes theirs? Mine are very inconsistent and I find myself having to generate multiple clips to get it right but in the end it doesn’t meet my expectations, please help me!!🙏🙏 if it works good I’m willing to PayPal

[junior horizon]

have you checked #flow-tips-n-tricks・✏️ and #flow-videos・📹 ? I recall some people sharing their prompts with their results for a lot of ASMR stuff

[oblique gust]

I think 9:16 is also free for ultra user as well. I just tested 12 videos in 9:16. No credits being used even though it says 20 credits.

[fierce nimbus]

Yup 9:16 is 0 credits for Ultra users. We are going to update ASAP

[junior horizon]

is 9:16 trained on any other way? I feel it is giving me really good results

[pliant raft]

Vertical aspect ratio Lets goooo thanks ill try it asap.

[molten seal]

Salut. Il n’y a pas de générateur sur discord. C’est sur le site Flow.google que tu pourra faire des vidéos, j’espère un peu plus originales qu’un selfie de gorille.

[opaque sand]

Confirmed. Also unlimited and also free of charge.

[oblique gust]

It's game changing lol.

[junior horizon]

I was so tired of editing, I love it

[opaque sand]

Don't rush. I just tried 9x16 from the reference image 9x16.

[oblique gust]

did you change it to 9:16?

[junior horizon]

haha you uploaded it on 9:16 but generated on 16:9

you need to upload and settle on same format

or that happens

[muted tinsel]

is 9:16 working well for yall?

[oblique gust]

yes it has been. I did 16 videos now.

all 16 videos are all 9:16

[junior horizon]

same here, audio issues on some but as ultra user that's no big deal

[oblique gust]

yeah i just generated 4 videos at a time and pick out the best one.

[junior horizon]

I mean.. we all get the big diclaimers on audio xD

[muted tinsel]

I wonder what ppl are going to come up with now with 9:16

ive been waiting for this

[junior horizon]

more yetis xD

[severe sedge]

Upscale option is not available in 9:16 ratio

[oblique gust]

I'll let you guys watch a sneak peak of the intro of my Boy 😄

[pliant raft]

You can get topaz, thats basically doing the same thing with more options and models, and with the right settings thats far better

[oblique gust]

This is for full 16:9 on youtube. I'm doing a 10 minutues series.

but for tiktok I have to do another version for it but in 9:16

[junior horizon]

nice gen, the lipsync is way better than most of the yetis I've seen

[oblique gust]

dude you have no idea how long it takes for me to get the JSON prompt right.

i made two versions one is mainly use cinematic static shot and one is for cinematic handheld walk and talk style.

[opaque sand]

Aaaaa, why do I have to choose the format twice, Karl, twice? What kind of i***t would turn a vertical frame into a horizontal video???!!!)))

[junior horizon]

you would be surprised xD

[muted tinsel]

ive seen these, how much do they cost?

[opaque sand]

In this case, vertical video is not free even for ultra subscribers.((((((

[elfin fulcrum]

20 credit

[oblique gust]

it said 20 credits but i checked. It didn't cost anything.

[elfin fulcrum]

9:16 Not available on Gemini

[oblique gust]

for ultra users at least.

[elfin fulcrum]

Ah , you have an ultra subscription,mine is pro

[junior horizon]

to be fair.. video generation should always be done on flow unless you wanna do sth quick - lowQ

[oblique gust]

if you do generate a lot and making ton of videos. I do suggest switching. I know it is expensive. But I cut off eating Chinese food so i could afford. lol jk.

[muted tinsel]

did u use image to video on this?

[elfin fulcrum]

Yes

[muted tinsel]

ahh i see

[elfin fulcrum]

Need more prompt ideas , is there any bundle of prompt to learn

[stuck adder]

Yeah it feels like they’re using nano banana idk

[oblique gust]

it's best to make it your own. Learn the strutures first and then you can pick and choose what work best for you line by line.

[dense mirage]

Can i get this prompt?

[elfin fulcrum]

Two hand gently pick up the figure on the frame and rotate it slightly for inpection.the model in the packaging box remain stationary,while all other background elements stay unchanged.

[dense mirage]

Thx a lot

[quick rapids]

9:16 added but upscale to 1080p removed. always a catch lol

[muted tinsel]

what was the photo u sued?

[stuck adder]

The upscale is not good anyway why use it

[quick rapids]

oh I agree honestly it was pretty useless

[stuck adder]

They just need to bring out video to video using the nano banana technology and we golden

[muted tinsel]

wait you used ingredients to video? it doesnt let me use that for portait 9:16

[elfin fulcrum]

I used frame to video

[muted tinsel]

ohh bruh

[junior horizon]

you can cut it but also can use another tool to expand the original picture so you have a 9:16 starting frame

[opaque sand]

Who needs this 1080p upscaling? The creators of the new Resident Evil ?

[stuck adder]

Vertical is def on another model

[kind bison]

Censorship on Flow is absolutely unreal!! I'm working on a crime movie series. Not even remotely R-rated - dialog and content could be on any evening network television show in the USA - yet VEO flat out refuses to generate videos because it doesn't like my dialog (a character says "he's going to kill us if he finds out"). Totally unacceptable!!

[oblique gust]

ok nvm VEO 3 isn't too important when it come to Wagyu steak you know. I gotta have to get a second side job if i want to afford both 😄

[muted tinsel]

anyone have any sick ideas for the new uipdate?

im gonna be brain storming all night

[junior horizon]

Loved your idea 😄 made my own figurine, I was not really feeling the trend but with video things change

[muted tinsel]

where is everyone making these figures?

[junior horizon]

nano banana (gemini or google ai Studio)

[muted tinsel]

thank you

[stuck adder]

Just ADR it for now lol

Or change the dialogue

[stuck adder]

What’s wagyu

[junior horizon]

"Your sub to life is about to expire"

from CSI vibes to scary movie

[stuck adder]

Lmaooooooo

Or I am going to defeat you “dragon ball z”

[kind bison]

I have to - I changed "kill" to "keel", so at least the lip movements will match. Flow still wouldn't generate the clip, but after the spelling change I got it to work on Fal (it blocked 'kill" as well). Google should be ashamed of themselves for censoring us so severely! Making me feel like a criminal or pornographer for trying to generate perfectly SFW content.

[junior horizon]

to be fair you were just trying to bypass a filter XD

[elfin fulcrum]

Using the model, create a 1/7 scale commercialized figurine of the characters in the picture, in a realistic style, in a real environment. The figurine is placed on a computer desk. The figurine has a round transparent acrylic base, with no text on the base. The content on the computer screen is the Zbrush modeling process of this figurine. Next to the computer screen is a BANDAI-style toy packaging box printed with the original  artwork. The packaging features two-dimensional flat illustrations.

Or

Create a 1/7 scale commercialized figurine of the characters in the picture, in a realistic style, in a real environment. The figurine is placed on a computer desk. The figurine has a round transparent acrylic base, with no text on the base. The content on the computer screen is a 3D modeling process of this figurine. Next to the computer screen is a toy packaging box, designed in a style reminiscent of high-quality collectible figures, printed with original artwork. The packaging features two-dimensional flat illustrations.

photorealistic, 4K resolution, sharp details rendered in Unreal Engine 5, cinematic lighting smooth edges, clean lines, no aliasing

Use this prompt for image

[junior horizon]

but yes.. can we chill a bit with filters? :C

like.. we have explosions but one guy can't say "kill" xD

[muted tinsel]

damm even with google ultra pro i only get a few tokens for nano banana lol

[muted tinsel]

I love how these came out

[coarse tendon]

anyone else not able to post vertical vids on tiktok?

[light plover]

I love today's update of google veo3 ❤️
I couldn't upload #flow-general・🌊 because my contents are korean.
but I will share my insight on #flow-creative-challenges・🎨 !!

[covert pasture]

anyone know how to make the video 16:9?

for tiktok

[light plover]

16 : 9 is easy, 9 : 16 is today's update

[covert pasture]

like when I put 16:9 is always a error trynna make this into a ai tiktok

[stuck adder]

Is it giving you a error?

Did you mark it as ai?

[stuck adder]

Nice same with runway and nano banana, how are we suppose to create movies with such filters!!!!!

[coarse tendon]

yeah I marked it. I uploaded a 1 minute and an 8 second they said it's under review 👎

[stuck adder]

You can’t even do like shooting or gun or anything like that like who wants to make . Kid movies lol

[stuck adder]

Let me see it?

[coarse tendon]

I can upload photos?

oh nevermind they just got approved

[stuck adder]

Ok

[coarse tendon]

that was a weird 36min wait

[stuck adder]

Have you guys tried seedance 4.0 ?

[calm hollow]

Will flow get mobile app

[crisp basalt]

Exactly

The closest thing we could get tbh

[kind bison]

Can't make kids movies, either - they don't allow kids to talk in videos!

[crisp basalt]

And it still crashes few times

[calm hollow]

I mean it would be cool if we would get google store app

Btw is google Pixel good phone

[kind bison]

Yeah - nothing special. I'm sticking with Midjourney.

[crisp basalt]

Compared to videofx phase I think the one we have now is much okay now

[calm hollow]

There is one video generator with no bypasses but its hella expensive for 50$ it only lets you generate 2 hq videos

[crisp basalt]

We have Google Flow app at home

The app in question :

[junior horizon]

https://tenor.com/view/discord-ping-discord-ping-pings-mass-ping-gif-2196332929166408226

hahahaha

[dry frost]

Same. I keep all of my servers fully read and zero muting

[junior horizon]

I would be dead if not muting, my discord crashed one day I had a sale just with dms

[stuck adder]

Which is that??

[calm hollow]

Invideo ai

[stuck adder]

How would you compare it to nano banana?

[junior horizon]

comparing bananas is pretty primitive

[calm hollow]

How you got free credits

[stuck adder]

Ok so not worth using

[calm hollow]

So its for rich people only damm

I swear ai tools should be less expensive

Why are they so expensive

How

Isn't midjourney on discord

[stuck adder]

Because the power takes to run million requests is expensive lol

Atleast I wish Google gave us a local option to run veo 3

[calm hollow]

You would need supercomputer for it anyways

[stuck adder]

I would’ve thought veo 3 would be

[stuck adder]

I’ve got a 3080ti 👀

[calm hollow]

Tho I can make simple videos with no music on my pc

[calm hollow]

I got 4080 super

Yeah you need supercomputer

[stuck adder]

Question: do you think other than adding full native 1080p support what else would you see veo 3 improving on? I can’t think of anything

[calm hollow]

But they don't have sound

[midnight hedge]

So many things. AOVs would be a game changer, but seed locking and finer tuned controls would be next level.

[stuck adder]

Yeah I agree completely

It’s like every other week someone’s bringing out a new AI model lmao

It use to be years now it’s months now it’s weeks

Soon new AI models will only last 3 days

[ruby olive]

Oh, come one, there is a million things to improve

1. reference generation 2) generate from uploaded sound 3) extend 4) GET RID OF THE RIDICULOUS CENSORSHIP

[crisp basalt]

https://tenor.com/view/beyonce-beyonce-knowles-smirk-laugh-trying-not-to-laugh-gif-11309455152051537806

[stuck adder]

😂😂😂😂😂

That reference thing is rock solid

[ruby olive]

what are you talking about? There is no reference generation in Veo-3

[junior horizon]

are we getting "extend" before GTA VI? jkjk I know its hard

[fierce nimbus]

there is reference generation in Flow but not for Veo 3. This is correct. Ingredients to Video is currently only available for Veo 2

[fierce nimbus]

That is Frames to Video not Ingredients to Video. Veo 3 supports Frames to Video in which you can define the first frame. But currently does not support Ingredients to Video where you can set references.

[crisp basalt]

just pretend that youre in ingredients to video section after you select the first frame 💆‍♀️ to use with veo 3

[stuck adder]

I didn’t say there was buddy I was meaning what he suggested

[ruby olive]

I can use many external things, that is not what reference generation is

reference 2 video is what we need

the only one who nailed it so far is Vidu

so we could have our characters with exact faces and features that stay consistent through out the frame

right now, if try to go from a wide shot to a close-up, you character face will morph into whatever

if the character is standing with his back into the frame and turns around, how are we going to get his face?

[raw oak]

hi. we used to be able to download 1080p. is it now 1080 by default?

[ruby olive]

no, its still the same old 720p

previous 1080p was a bad quality upscale

I hope the new 1080p is something different

[prisma holly]

Thanks Google team for the vertical ratio

[raw oak]

Ok. but it's not available now. I seems i only have 720p

[raw oak]

can you download 1080p ?

[ruby olive]

yes

same option as before

[raw oak]

not me

[kind bison]

It looks pretty much the same as the upscale.

[ruby olive]

that is sad, but expected

I guess that is why the price is the same

Or it would be double

[raw oak]

Can everybody download 1080p I do not seem to have the option..

[muted tinsel]

we are so early

who else making a start up rn?

[stuck adder]

It’s not available for vertical video

[raw oak]

thanks!

[stuck adder]

And 1080p native is not on flow as yet wouldn’t use it anyway

[raw oak]

because it's not that good

[foggy folio]

Can you make avatars whit veo?? recognize my voice and gestures from a short video?

[crisp basalt]

No

[foggy folio]

Which Google tools could help me do that?? Spetially the voice??

[crisp basalt]

No Google tool can do that yet

[muted tinsel]

Hey im a little confused on how the tokens for nano banana work, do these tokens reset? or what do we do once we hit the limit?

[crisp basalt]

I think this ain’t the right channel for that question

[muted tinsel]

where should i ask it

[crisp basalt]

Idk are you using nano banana through Gemini or aistudio

[vivid fossil]

I keep getting “Failed Generation” on 9:16 videos on Flow, anybody else getting this error?

[safe nova]

no

[analog radish]

One of my favorite things in the morning, with my coffee, is seeing the notifications from the flow-announcements channel. You guys are doing great work, don’t stop, we love it!

[cold thunder]

is it possible to create veo 3 videos without audio?

[lusty kelp]

How do you avoid the AI not making spelling mistakes in Veo it seems like all my credits gots towards fixing that if i have anywords in my video

[bright agate]

They gave 9:16. Veo is a lifesaver

[clever flint]

I have zero credits now : ( Do google gift cards work for recharging my account?
If i find one

[lusty seal]

Flow Credits could rest month by month or day by day?

[clever flint]

Month by month apparently

[prisma dock]

is flow free rn?

now showing me each video cost 20 credits

[sweet pebble]

Hey! I generated my first veo3 vids on artlist.io with Hungarian language and it worked perfectly - now that I tried it with my new google labs pro subsription in Flow it doesn't work. what could be the problem? edit: the promt I use is ENG - but I want my characters to speak HUN.

[pliant raft]

thats funny how 1 image with exact same prompt so everything the same can pass the censorship one time, and then getting blocked for the next 20 attempts so i gotta stick with the one and only mid first attempt that went to the end of the loading.

[tidal maple]

Is there any information about a future update to add 1080p to the vertical format?

[crisp basalt]

probably when the api supports 1080p native for 9:16

[tidal maple]

ok thank you

[stuck adder]

I wouldn’t use it anyway it’s bad rn

If they had topaz as the backend then I would’ve

[tidal maple]

Yes, that's what I noticed today. I tested it for several hours and the results are still not satisfactory.

[oblique gust]

Good morning guys 🙂

[frigid pier]

I am experiencing the upscale issue now.

[hasty island]

same here i press the 1080 button and nothing happend

[frigid pier]

yea.. This issue happens often recently..

frustrating 😐

[foggy heron]

uh guys why am i getting auto payments?

[ionic dagger]

does the crop option is fixed in flow

[ionic dagger]

why video quality is so low when i generate a video

[signal spindle]

Hello folks, why the audio is not generating? I am using Flow & VEO 3 fast model.

[frigid pier]

Audio is still beta. So the AI model doesn't generate the audio somtimes.

[dry frost]

It also depends on what you wanted to say and from who

[tame granite]

Wheres the best place to find prompts for tiktok asmr ai videos?

[kind bison]

Anyone else go through this cycle with generations? First one will be OK, but there's one fatal error that makes it unusable. Next gen, the model does it's own thing and strays away from the prompt. Third-fourth generations: the model just goes crazy and stops following instructions. Fifth generation: it sort of comes to its senses, but now a new fatal error that makes it unusable. 6th-10th generations: just a roll of the dice at the casino if it follows the prompt, so you can maybe cobble together an edit in post.

[fluid pasture]

This is happening to me as well

Anyway to fix this today?, i just tried with Voe 3 quality and that's not working as well + tried on my second acc and its the same issue

[rocky tangle]

Hey there,
sorry if I asking you this question again. 🙏

But, if I purchase ULTRA, the Veo3 Fast is completely included?

Makes Veo3 Fast sounds to video too, like Veo3?

[rocky tangle]

It would be really great, and kind, if you could answer this question for me so I know whether I should get this ULTRA.
I can well imagine you're all very busy. But this one quick question is quick.

[rocky tangle]

☝️

[sweet pebble]

I have ultra for a week now and it has veo fast and quality with beta sound - so if you want veo 3 with sound - it comes with the ultra sub and every video is 100 kredits from you 25000

[ornate quarry]

I'm trying to figure out why this is failing - "Animate this man being surprised by a adult male with short dark brown tousled hair, a full neatly trimmed beard and mustache, medium straight eyebrows, medium-sized eyes, a rounded face with a broad jawline, wearing a black T-shirt and showing a neutral expression, who is deaf, spilling M&Ms over the man, then shouting 'Damn you Minty!'."

I've done 4 times on it and its failed all attempts

I've lost like 150 flow credits to it

[oblique gust]

Veo 3 fast is free if you have ultra plan. 16:9 or 9:16 all free

[oblique gust]

Could be the language you used in this prompt. Reword it to

[ornate quarry]

Reword it to what?

[cobalt timber]

any info when making portrait vids with ingredients will be available on veo 3? or do i need the ultra plan for that?

[oblique gust]

Inclusion of Disability ("who is deaf") Text-to-video models (including Veo) have guardrails around medical or disability-related attributes. Even though you’re describing the character, the mention of “who is deaf” often triggers refusals or silent failures.

[ornate quarry]

Nope

[oblique gust]

At least that what I think

[ornate quarry]

Exlucded it and it wont generate

Ok i used a different image and it works with that

Im so confused

[rocky tangle]

Thank you for your answer!

[oblique gust]

AI is tricky lol.

[rocky tangle]

Thank you! But in your ULTRA thing, is Veo3 Fast for free, right!?

[ornate quarry]

Do i get points back for failed generators

or are they permanetnly gone

[rare latch]

Why flow videos are loading so slow and downloading them is even slower ?

[ornate quarry]

Like all my flow videos are failing to generate...

[gaunt crypt]

All my animations which were Pixar-like and then went to Hannah-Barbera style refuse to generate audio as the model states a thing re. not generating audio for minors. Was perfectly fine on the Pro Plan. I update to the Ultra plan and all of a sudden it is flagging me on audio for every style possible.
Why would the system flag up a cartoon that resembles that of Scooby Doo and not provide audio because of the use of minors? £100 more a month and it gives me the finger. Any ideas folks? Am I missing something. Unusable c**p. That's what I'm getting all for the small default price of ~£250pm. Thanks in advance for any ideas

[keen swallow]

I'm having trouble uploading frames to generate videos on Veo3. The system keeps saying that I'm using images of famous people, but I'm not. All the faces are AI-generated and don’t reference any real person. Has anyone else experienced this or knows how to fix it?

[twilit delta]

where can i cancel my subscription?

does someone have a link? i cannot find it

im paying 19,99$ monthly

[rocky tangle]

Thank you for confirming it. I wasn't quite sure if it was still current.

[jagged jackal]

You either gotta keep trying till it accepts or move on. That's really the only two options you got as it does this to me sometimes.

[keen swallow]

I did that, I changed the ethnicity of all the characters and redid all the frames of the scenes, and now it worked. Thanks for the advice!

[modest jolt]

help fix

[rocky tangle]

Understand. Yes. It could be really like that. The situation is as follows, actually I only need the flow access, and the 25,000 credits.

Of course, if Veo3 Fast falls to unlimited, it would be even better. I don't need Google One, etc. But I have to take everything with me if I can only choose between A and B.

What would really interest me, how is it as an ULTRA customer, you still have a limit on how many videos you can create per day, i.e. at Veo3 Fast, for example? (Credits excluded)

For example, if you want to make 10x videos and have enough credits. So there is a basic generation limit per day?

[midnight hedge]

Probably have to have some hard limits set considering how easy it would be to exploit for further degredation of their other site YouTube completely hand free with bots making slop and uploading.

[rocky tangle]

Or can you use ULTRA to create almost hundreds or thousands of videos a day with Veo3 fast?!

[rocky tangle]

Oh, oh, so. Means that there is a basic limit so that this cannot happen, that any bots make nonsense.

But it would still be fierce if even then the ULTRA people have a limit. Apart from the credits.

[midnight hedge]

I don't know that there is a limit, just saying it would make sense if there was some type of hard or soft cap. I have ultra and even working on a project for 12 hours not sure I hit 100 in one day.

[rocky tangle]

Okay, 100x Videos on one day is great. This would be enough then.

Don't worry please, I took it to the point where you suspect it.

Thank you! 🙏

[rocky tangle]

Thanks.
So, purely theoretically, you could then generate 100x in a row without running into the limit. or even more.

Yes, exactly. One could have a long video generated as an example. In these 8 seconds, steps were taken. But for a few minutes, there are already a lot of clips. Hence my demand for the limits. Without the credits.

Sorry if it was expressed incorrectly by me.

🙏

Either I get something wrong, or you. Sorry.

With all due respect. I don't want to make 20x videos at the same time.

[safe spoke]

Hey guys so im trying to make a video for my twitch and the style of the character always come out the same and the style and the scene just doesn't process right is there any changes to the prompt i can do to make it closer to what i want?

[modest jolt]

helpppppppppp please fix

[rocky tangle]

Yes. But... about the same limits. Frontier Modal.

Pleaae. Look when I finish Clip 1 after 3x attempt. I'll do Clip 2... and so far.. so long, thats my some minutes Video, consisting of many 8 seconds of clips, compiled and exported as a whole video, finished.

Maybe there's something I don't know. Maybe that's why I don't understand what you mean.

[thick kayak]

I keep trying to animate my look alike pictures and it's always telling me that I'm a minor or a prominent figure

[remote crypt]

I just got an update info in Google Flow, but I mistakenly clicked it away. Is there any way to get it back to read it? Seems to have been an update. But I can't find it on the announcements channel

[sacred kettle]

IT Is a short thing. i can assure you of that. i already went thru. cannot do it again

[sacred kettle]

yeah me too, im sick of this. it has no fun. i mean i respecxt celiebrities and all but what, no paparazis? no nothing? now MY OWN FACE IS A "PROMININET FACE"

this seems tobe midjrouney, where you cannot make a spy movie short because if you go thru a hallway "it can lead to a situation of possible surveillance camera". same people that creates the images for the trending videos of strange creatures fake found footage

the irony. also im from like, descendant of polish people. i have chubby cheeks. my cousin female is chubby . she is 31. "you cannot upload, she is a minor". chubby cheeks, always a problem

wait the 3 months until the offer expire. if you were just refering to that, i apologize, i missunderstood

btw i just received that theoretically we have character refrence and ingredients, in europe, and other countries. i tried it, nothin. no character ref. no ingredients on veo 3. happened to any9one else?

[stuck adder]

It’s just reiterating the vertical style video etc

[crisp basalt]

hand me your ultra ccount 🖐️

[frigid pier]

Am I the only one who feel like Veo 3 fast voice quality got worse recently?

[midnight hedge]

God this shot was so hard to get working. Hope later versions of frame to video get a bit better at stuff like swinging.

[hardy frigate]

Anybody know where to turn on the 1080p output in google flow I've an Ultra account but I don't use it much thanks

[midnight hedge]

Not available in Flow yet. Only option for 1080 is upscaling at the moment.

[crisp basalt]

Unfortunately yeah you’re the only one

[fossil rampart]

sujon_ahmed_19

[frigid pier]

niceee

[thorn ether]

hello i have e problem with my account in flo, how contact the support ?

flow

[hardy frigate]

anybody know how to fix the vertical black bar on top and bottom of the generated video?

[silk nexus]

Does anyone know when the 1,000 credits get added to an account if you are on the Pro plan paying monthly?
I've just had the monthly payment taken out of my bank but the credits have not been applied yet. How long does it take? Thanks.

[serene flume]

Anyone having issues downloading Upscaled videos??

[crisp basalt]

have u check the library ? usually it actually did the job but sometimes the download button doesnt appear so you might need to check in the library page , the upscaled video should be in there

[calm herald]

When creating videos with Google Flow VEO3, I sometimes end up with unwanted black bars at the top and bottom of the video. Other times, with the exact same prompt, they don’t appear. Most of the time these black spaces remain, and I haven’t found a reliable solution so far. I’ll also share two different example images to show the issue.

[stuck adder]

Just remove them if it’s causing an issue on a prompt output you really like

[neat sierra]

i literally upgraded to the 124.99/month to be able to use veo 3 ingredients to video and it says its not supported......???

how do i get my money back, i signed up 5 mins ago soley to use that feature

[fringe yacht]

i keep getting this for every video, even tho im not breaking their TOS, and iran prompts that worked before, exact ones, but now it doenst let me:

[jagged jackal]

I've had prompts rejected for minor things. Yesterday "she sits beside statue" was constantly rejected, but changing it to "She sits next to statue" went through just fine. 🤷‍♂️

[opaque sand]

Maybe it's paranoia, but I feel like I achieve the desired result faster by NOT running 2x4 generations at the same time. Only 1x4.
I feel like the amount of “power” allocated to my account is divided between two tasks, and distortions occur more often...

Can you refute or confirm my assumption?

[supple pendant]

if you're talking about iterating 1 at a time and getting better results than 4 at a time then yes, you might get more consistent results doing 1 at a time becuaes when you iterate multiple, veo creates variety between the 4 versions because it doesn't know why you are iterating multiples. So the results you get might be a bit more wonky

[opaque sand]

I'm talking about this - I always do 4 iterations at a time.

[elfin fulcrum]

How's this

[elfin fulcrum]

CFBR

[mellow cove]

Amazing can you send the prompt over

[elfin fulcrum]

A sports film. A young basketball player is practicing alone at night on a court. The camera is a wide shot, capturing the quiet and solitude. Audio: the squeak of shoes on the court, the bounce of the basketball, quiet contemplative music

[mellow cove]

Thanks 😊

[queen lantern]

Hi all - my names Vincent have been building with veo3 since June 🙂
Grew tiktok to around 6k followers purely on veo3 edits
nice to meet you

[wild jetty]

can you export your projects?

[crisp basalt]

No

[sacred kettle]

veo 3 is in silly mode it seems. im using my own face, real photo and one with lora. also one from my cousin. so far i was unable to get only one singl e video makign variations of our faces for the ai ones. "veo doesnt make videos of prominent people". if at least i were famous...or my cousing.... somethng is not right.

you are a person of dilligent attidue. thanks for the table. saving it as a "reference frame " 😄

[tropic night]

how much KW does a veo3 use

does anyone know?

[thick thicket]

Ultra realistic cinematic jungle path, same beautiful woman, short jungle dress, slightly revealing but safe, walking, cheetah attacks, she screams, giant gorilla with ornaments arrives, kills cheetah, saves her, carries her to cave, photorealistic lighting.
‎

[sacred kettle]

is ths the official labs google discord? for all world' users right? please correct me if im wrong

[stone oar]

...

[kind oriole]

Bad news guys i don't know from when but, family member don't get AI Points right now 🥲🥲

If one member spending point, all family member spend too only 1k in family group thats really terrible

[supple pendant]

really well done! This is like such a better version of that Coca Cola holiday campaign from a bit ago lol. the characters are really cute and the pacing of the commercial and integration of the AI with the live action feels seamless and purposeful! I have one little tiny suggestion if you are open to suggestions, but if this is already shipped i'll keep mum 🤫

[meager wolf]

Really? That's bad news actually since i think flow want reaching more users

[real kayak]

Since Gemini only lets you create three videos per day? I have the 1,000-credit-per-month plan on Flow.

[onyx moon]

I’ve seen Veo3 fast when iterating 4 at a time completely change the dialogue. That’s wild. There should be a setting forcing it to stick to the script instead of creating variety without telling us.

[tender island]

is that a problem?

[tender island]

hi, great to see you. can you please share me your channel? Thanks!

[ripe yacht]

Hello, as stated in #start-here・📍 , “Currently our team can support engagement in English; the discussion in this server will largely remain in English for now." Thank you for understanding!

[sour silo]

Thanks for your amazing feedback! The crazy thing is that there is NO live action footage, the video is entirely VEO3 generated 😄 Always open to suggestions but yep this one has already been shipped

[sand current]

Dammn fcin good

[supple pendant]

That's absolutely wild!! Great job!! Completely fooled me haha

All I was going to say was if there was any way for you to improve the dialogue audio - like what I've been doing is putting the veo audio into elevenlabs to get a slightly cleaner output with less artifacting. (Google might have an audio tool that does something too but not too sure.) To me personally, a big tell when i notice something is Google Veo AI, is hearing the slightly "tin-like" quality of the audio - esp with dialogue.

But yeah, overall amazing work!! Very cool to see. Thanks for sharing!!

[violet parrot]

Hi

[tender island]

Anyone creating a serious video series using Veo ? P.S. I am a developer developing an application on Veo 3. Need feedback from serious Veo3 users. Thanks!

[wind spoke]

Hey, I posted here yesterday but its gone... was it deleted?

[sullen maple]

Hey, we are creating content based on veo3 former nasa scientists on sceince communication if that helps?

[digital escarp]

Definitely. All my content for my indie project is all veo generated. Tap in and I’ll share.

[tender island]

Thanks DMing you..!

[queen lantern]

Add me! Not sure if we can send links here

[queen lantern]

Tyty

[tender island]

sent you DM. Thanks!

[meager wolf]

Flow not working what happened?

[versed peak]

Flow's broken.. AGAIN. The audio generation crashed yesterday and since it's the weekend, there's no one in office to reset it.

[stuck adder]

Flow needs to automatically pick a landscape or vertical depending on what you just imported

Because I could see people wasting credits by forgetting to hit portrait or landscape

[proud crown]

How to build

[final charm]

good to know. This is just my second attempt to play with it, and I'm getting rather upset with it.

[young narwhal]

2 hydraulic press flow vids

[steep cedar]

Vídeo

[full flicker]

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

[gilded zealot]

Hi

A wide, cinematic shot of a vibrant and crowded Thadingyut Light Festival night in Myanmar. The scen

[high fulcrum]

@raw stump we are unable to generate images in this server 🙂

[arctic cradle]

Is the Ultra Plan discount system over now?

[stuck zealot]

[lost oracle]

hi, any news on when we can use the images to video feature in vertical screen??

[unreal viper]

REMINDER!

we cannot generate images in the server! please do not just paste your prompt into the server as it will not generate, unless you are asking for advice!

[topaz thistle]

Guys, I need help, please. Trying to make a video using this photo as a reference. I need the smoke to be slowly removed, and the person should be panned towards his left so that we can see his full body. Some light should be entering at an angle into his chest and slowly his entire body gets radiate with golden light

I'm not happy with the O/P

Any suggestions or tips for better prompt generation? please?

[topaz thistle]

Can you please guide with the shorter prompt?

Tried many but was not helping much

so finally posted here for some help

I gave my idea in Chatgpt and asked for Prompt and kept on modifying it till i got the intended one

Is there other way I can try to get the prompt, please guide

How can I improve the Prompts please?

[winged plinth]

Try around this:
A single, continuous cinematic shot. The camera starts with the framing of the original image (a medium shot silhouette of a man meditating), with dark smoke spiraling from his back. The camera then slowly and smoothly pulls back (or zooms out) to reveal the man's full body sitting in the lotus position on the rocky ledge. As the camera pulls back, the dark smoke completely dissipates and fades away. Simultaneously, a divine golden shaft of light shines onto his chest, and this light radiates outward, engulfing his entire body in a soft, glowing aura just as the camera move is complete.

[topaz thistle]

I have not asked you to fix bro....

Thats what i'm doing

Incase if there is some good case studies or place where we can get more knowledge on this, please help with that so that it will be useful

[winged plinth]

You’re doing great. We all figured it out just like you’re doing, take your time.

[copper fog]

I am sure I can find the answer to this question by using Google but I really want a human explanation.
How are these 3rd party apps able to advertise that they can let users create videos using Veo at 1/3 of the cost?

[inland goblet]

This helps me a TON, I just didn't understand the prompts like the concept is so out of my depth. I kept thinking in terms of programming and coding or something (which I know barely anything about outside of basic python stuff)

[dry frost]

Using a screenwriting type prompting style tends to work rather well too. Which makes sense since you yourself are the director.

[winged plinth]

Give SAGA from @earnest oyster a shot. It’s been the best one I’ve used since.

[crisp basalt]

idk why he start spinning but i vibe with that

[jolly rover]

oi

[slate sphinx]

Getting error that image of minors not accepted this time but i generated that image from gemini
also it's a pixar style image

how to solve this

[sacred kettle]

i was getting continous erros of any kind. i was able to work with veo 3 flow for 3 hours. now is "flow is temporarily unavailable". i live in latinamerica. anyone in this zone or any other experiencing the smae issue, didnt see any message in announcemnts or general

[sacred kettle]

i have deatl with that problem over the last 3 days . i created 2 bugs reports. believe me i tried everything, and there is no way. this is just a bug . i am using perfectly adult faces with wrinkles and the age is indisputable and yet i get them anyway. sometimes very strange few, the smae prompt with the same image that was reported as a minor goes thru without any problem, which contradicts veo 3 enocder itself

[slate sphinx]

Need solution from devs

[fading dirge]

flow keep blocking me about prominent people on my Ai generated people lol

[glacial birch]

Did you guys try the new timeline prompt? Does it work?

[kind bison]

What's the new timeline prompt?

[glacial birch]

Check flow tips n tricks

What do you guys think?

[real vapor]

goat waking

[real vapor]

#flow-general・🌊 Translucent passengers waving from train windows, eerie glowing effect

[tawny jolt]

What's the duration of the flow generated videos?

[kind bison]

Ah, OK - yeah, I've been using that since VEO 3 launched...but I use a simpler format, like this:
Shots: 1) 0-4 s | (then add the shot type and action/dialog), 2) 4-6 s | (next shot), 3) 6-8 s | (next shot).

[cold thunder]

do some of you also got extremly long upscaling time?

normally it takes 2-3 mins but currently its like easy 10+ minutes

[crisp basalt]

try check in your library page

see if the upscaled video is there

[cold thunder]

yes everything is working correctly but it takes alot more time then usually

[crisp basalt]

yeah its an on going issues i think,

[cold thunder]

Is there something like a usage graph that shows when people are most active?
Maybe the number of people creating content depends on current demand

[crisp basalt]

i dont think so

[rocky tangle]

Hey everyone! 🙌
I have another important question for you. I read that with the "Google AI ULTRA" upgrade, you can only create 5x Veo3 videos per day, and with "Google AI PRO", only 3x.

That would be really... not much for that hefty price!

[crisp basalt]

thank you for sharing

what was the question lol

[crisp basalt]

i think its broken now lol

clicking upscale does nothing for me now it seems

[rocky tangle]

Yepp.
We have the credits. And we have the limit.
Today we're talking about the limit.

The rest can be found above.
5 Veo3 videos per day is really low on an ULTRA.
Can anyone confirm this limit?

[crisp basalt]

the important question is , where did u get this info lol

[rocky tangle]

Google

[crisp basalt]

so you decided to not check Flow Faq page?

[rocky tangle]

Its about Flow, Whisk, Gemini

[rocky tangle]

But I see you don't know either.

[crisp basalt]

I already know the answer , I’m just weird out how you managed to google it but didn’t bother checking out Flow official faq page

Since when Flow uses “vids per day” quota type lol

That’s only on Gemini

Flow uses ai credits

If your ai credit is finished then that’s it you can’t generate video

[rocky tangle]

Where am I supposed to read up on all sorts of things?
Google Support, Google Flow, Google Gemini, Google Whisk, Google Community... sure. I don't have anything else better to do.

Thanks for helping me anyway and telling me. 🙏
So as an ULTRA customer at Veo3 Fast I could create a lot more videos per day!?

[crisp basalt]

Yes if you’re an Ultra user you can use veo 3 fast , for free since it’s 0 credit

As for what you read about 5 vids per day, those are only for Gemini/Google Vids and whisk (I think?)

[balmy adder]

Is it possible in whisk to make video's in my native language instead of English?

[crisp basalt]

So don’t worry! 😉

[crisp basalt]

It’s fine I don’t blame you for the confusion

But do remember , the Veo 3 fast 0 credit only applies for Landscape ratio video

[rocky tangle]

Yes, there are so many platforms here these days. Getting a complete overview is a nightmare!

That's where a helping hand like yours (and from some other members here) really comes in handy. Thank you very much.

If it were a bit more affordable, I wouldn't have to ask, read up, and calculate so much... but these euros per month are already a lot for me. Since I'm ONLY using it for fun and won't be making any money from it.

[crisp basalt]

@rocky tangle you can check out this in Flow

[rocky tangle]

I´ll do it now!

🙏

Can i add you and Frontier Modal as friend here by Discord?

please

[rocky tangle]

If its allowed.!?

No. Thanks. I meaning with it... its allowed to add wrigley to my friend list.
I havent much, but this one in my list, are all the best! ❤️ i´m so thankfull.

[tawny jolt]

Hm

I see

[wet spoke]

I can't crop the video in scenebuilder. Am using it in phone with desktop mode

[oblique gust]

well ultra user don't get charge on credits at all. All Free.

[kind bison]

Only on Fast mode.

[oblique gust]

i mean isn't that enough? >"<

[tawny jolt]

How did i generate it Lol if it was nsfw

[tawny sequoia]

Man, so I have a 4minute 32 second long video. I have to use desktop site version through my phone to use Google Flow Scenebuilder. This is the tricky part, I export the video and then the notification pops up to either download or dismiss, which I tap the "Download". The whole Google Chrome app crashes and my phone is at the home screen.

[oblique gust]

how is it incorrect? Veo 3 fast doesn't charge you any credit.

[ebon cradle]

How to make long 9:16 videos if Flow doesnt support portraits in scene builder?

[inland goblet]

Yeah that's kinda what I was thinking would help but I second guess myself so often I go from "I think I'm getting the hang of this" to "I have no idea what I'm doing" in like a split second

[analog sedge]

You’re talking to a bot

[midnight hedge]

Are we getting these features in flow as well? https://techcrunch.com/2025/09/16/youtube-announces-new-generative-ai-tools-for-shorts-creators/

Particularly very interested in the video motion to image feature.

[mild abyss]

I have ultra plan and im making videos using veo 3 fast. I don't know if im being crazy or not, but i feel like videos im getting from Gemini (which uses veo 3 fast right? or does it use veo 3 quality?) are simpler better than the results im getting from Flow (using veo 3 fast in flow too). Does Gemini use Gemini LLM to refine the prompts before passing them into Veo 3 whereas Flow just feeds the prompts directly? Or am i just being crazy here?

[tawny sequoia]

No, I mentioned that already and someone explained simply that, Gemini is a good conversation AI so it can interpret basic prompts into videos where as Google Flow may require more specific prompts for your videos. Both are good but have pros and cons.
Pros of Gemini Videomaking: Simpler prompts, easy video generation.

Cons:less consistent outputs for complex requests, different level of granular as compared to Flow's interface.

Flow pros: may offer more advanced, specialized controls (camera control, scenebuilder, "ingredients to video", etc) that allows more cinematic control for the user.

Cons of Flow: May require more complex directions and a richer set of instructions for desired result.

[mild abyss]

Ah! Interesting to see that you noticed something similar.

Thanks for the explanation. So basically my hunch was correct? Gemini will do more of interpretation where as Flow is more direct and requires you to be more specific

I wonder if it's possible to get Gemini to tell me the exact prompt it feeds into the veo model

[hard badge]

According to the announcement I saw, it's now possible to make vertical videos but when I try to use it, it doesn't work.

[wild garden]

Hello, I just subscribed to Ultra today. Back then I saw an update in google flow that starting 8/26, veo 3 fast is free for Ultra subscribers. But when I try it now, I get deducted? Any particular reason why?

[crisp basalt]

❤️ ❤️ thanks labs / flow team ❤️

[crisp basalt]

You can also throw in “basic prompts” in Flow too. Flow runs on Gemini behind the scenes , it takes whatever text you give, processes it, then passes it over to Veo

That’s why with Veo 3 you dont really need to stress about crafting the perfect prompt. As long as what you type makes sense and is clear enough, Gemini will expand it behind the scenes for you.

Honestly, i underestimated that line about “Veo 3 having better prompt adherence” , turns out its spot on

ngl my fav thing about Veo 3 is when I keep the prompt short + clear, it still throws out tons of variety but always keeps the main idea intact

when u go more detailed w/ the prompt and let’s say you want to add descriptive details about the subject you want in video , veo 3 locks onto the the subjects . for example I mention like dude in the hawaiian shirt, the woman etc), but switches up the shots/angles around them. so you get variety but it still sticks to what you wrote

in conclusion, BEES! 🐝

[distant elm]

As reminder we cannot create images and videos in the server! Also as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now. Thanks all!

[dull pelican]

Cropping tool is still not proper . i have a 9:16 image and still i am not able to crop properly in free form . Anyone else facing the issue?

[weak gull]

i'm in love with flow veo 3 😍

thank you google!

[crisp basalt]

another neat update ! 🍌

[surreal rain]

your picture doesn't look 9:16 to mee, the crop looks like 9:16 ratio so it should be fine

[twilit pebble]

I cant' recognize the UI, what are you using? O:

[crisp basalt]

It’s in the “generate image” !

[twilit pebble]

gemini right? totally different to me the layout ahah

[crisp basalt]

Oh nah all is in Flow !

Click generate image

Then you can generate image or use your uploaded image

After it generated images , you can finally use nano banana

[sand pond]

Wasted my free credits on stupid ideas/vids 😂😥

[primal hearth]

I hope I am in the right place - I have just been on with GoogleOne and they told me to ask over here. I have a google one Pro subscription, I am in the UK, I am well over 18 Google have confirmed that my subscription is correct and active, but when I go to flow, I keep getting the same error message.

It looks like you don't have access to Flow. Check availability here and reach out to support for issues.

I hope someone here can help.

[tepid tundra]

What a pickle, I'm doing a zombie series but I cannot create a scene of a zombie atack on a person haha

[heavy drift]

Man, the extremely strict content filtering really stifles creative flow. So, I have an online influencer that never does NSFW but more product placement, fashion, etc. I can barely get video out of VEO 3 and Flow anymore for her, as an example. It has lately started labeling her as a prominent person...maybe becasue I am making her one but not because she is one, or even looks too close to someone else. I can take it over to another provider, run it through VEO 3 there, no problem. I was trying to create a morning coffee video, she is wearing a half shirt and gray sweat shorts in the morning sunlight from her window and holding a cup of coffee. I wanted to her to say "I start my day with a morning cup" You get the idea. It labelled it as sexual content. I reran it and it finally took it. My question, is there value in keeping a premium plan here or just use VEO 3 through some other tool? Will Google every review and make the content filter reasonable?

[tepid tundra]

I had some issues also with nudity while doing a shower scene but it was just a closeup to the face nothing else

[sour grove]

hustling with Veo 3

[crisp basalt]

seems fine on my end

oh sht the first vid generated was

😬 ...

ill just go ahead and flag. anyways!

[crisp basalt]

i didnt use any picture for reference for the subject but yeah have this with side of google translate TTS voice low quality edition

[tepid tundra]

hahah yeah, I'm just doing scenes excluding the explicit gore and the rest will be up for imagination I guess for whoever watches it

[dense frost]

Any info on making portrait vids with ingredients? I also tried adding rotate 90 degrees in the prompt, but it didn’t work

[crisp basalt]

ingred to vid?

[dense frost]

yess

[crisp basalt]

it doesnt do 9:16 tho cuz it uses veo 2 model when u use ing to vid feature

only veo 3 model can do 9:16

[dense frost]

I tried using 'frames to video' with veo 2 fast model and added an image where the head was on the left and the body was on the right (rotated 90 degrees). The result was a 16:9 video, but it was rotated by 90 degrees. I downloaded it and rotated it back to 9:16. With this method, I got a video with good quality for 9:16. I thought 'ingred to video' could do the same, but it didn't

[solar idol]

Hi! As stated in #rules-and-guidelines・📋 ,
Please keep conversations as English-only (for now)!
This helps us moderate our community as we continue to grow.

[kind bison]

I did a zombie video last night and got so angry with VEO 3 that it wouldn't follow instructions. Wasn't looking for gore - only for a zombie to grab a woman, then she screams and runs away. It kept making the zombie act like her friend comforting her and she barely reacted to the zombie (and no, it's not a prompting issue - I create multiple VEO 3 every single day, and have done so since it was launched - VEO 3 just refuses to follow the prompt for most Horror-related content, while it will follow the prompt for non-Horror stuff).

[tepid tundra]

Yep it's annoying, but it seems to work fine for someone shooting another person tho 😂

[analog lava]

Hello, as a friendly reminder images or videos cannot be generated directly in the discord server thank you!!

[supple terrace]

New here, what’s the general consensus on benefits of the ludicrously expensive AI Ultra vs the laymans AI pro currently going for $20/mo.

[kind bison]

Free VEO 3 Fast generations on Ultra.

[vernal cedar]

Is there anyone here who’s handling a faceless channel in youtube with veo3?

[tacit aspen]

Faceless meaning your not on camera and its AI ? Yes

[nocturne depot]

hi, how can i get to contact with support about avaibality of flow in poland. I think i have a bug

[acoustic sparrow]

Hello all, i need help with some prompt, i need make a video about earthquakes , any prompt for get that? i need the camera shake

[tepid tundra]

A chaotic, shaky handheld shot

there are many kind of shots for it, try to learn the basic about types of shot/angles/takes

if i'm not mistaken, the earthquake destroy building bottom to top, so the first thing to colapse should be the bottom floor

[acoustic sparrow]

im gonna try

[tepid tundra]

might as well run your prompt into Gemini and see if it improves to make it more realistic

[acoustic sparrow]

ok, i using flow and well, looks bad

[winged plinth]

What's your prompt?

[peak lark]

hi guys. how I can bypass the error with proeminent people? for a while worked very well, now my business is depending on this.. there are AI persons not real ones.. does anyone have a clue?

[acoustic sparrow]

EXT. COLONIAL LIMA STREET, 1750 — A chaotic, shaky handheld shot — continuous violent shake (low-frequency ground roll + mid/high-frequency jitter); horizon tilts and drifts; sudden seismic jolts; strong motion blur; subtle rolling-shutter skew. Adobe-and-quincha houses with carved wooden balconies; baroque bell tower swaying; roof tiles cascading; dust clouds vibrating; cobblestones heaving with a widening fissure; 18th-century civilians as silhouettes only. Gray Lima winter sky (garúa), desaturated palette. Everything in frame vibrates; absolutely no static objects. Negative: no gimbal, no dolly, no smooth pans, no fixed horizon, no modern elements.

[lavish field]

anyone having issues with flow and add to scene?..mine has just started going blank, I click add to scene but the short wont show up it is just a blank spot there ..tried over a few videos Ive done that it was working with no issues just few moments agao..tried clearing cache and downloads etc

[pseudo coral]

I have the same problem too, I was working on a project with scenes already added, but a few moments later it just went blank., even though it says my video has already added into the scene

[zenith shore]

Would anyone know what AI-video generator was used to create this video? I'm guessing that its made with Veo 3, but I'm not 100% sure.
https://www.instagram.com/reel/DOq4snzCCVg/

juventus

This is where dreams kick off ✨
#JuveBVB

Likes

265459

Instagram

[blazing mulch]

weird. i thought we can already create 9:16 aspect ratio on Flow ? then why won't it accept a 9:16 image upload? it's still giving me that landscape cropping tool. Thanks in advance!

[blazing mulch]

weird i don't get that cropping option. it's stuck at "landscape" 🙁

[crisp basalt]

Are you using Flow on mobile or desktop?

[blazing mulch]

im on my mac book pro atm

[crisp basalt]

Weird, can you screenshot the part where it ask you to crop your ingredient?

[blazing mulch]

that's the exact interface, it's just that it's on landscape, it won't let me choose anything else, it won't even show "portrait"

[crisp basalt]

What browser are you using?

[crisp basalt]

even then i dont think it matters tbh. prob workaround is try flow on another browser, or on your phone, just to see if it still happens on diff devices

[blazing mulch]

thank you

im using chrome.

[woeful ice]

How do I download without water marks in flow ?

[blazing mulch]

you can't. just adjust the canvass until you are able to hide it after generation.

[proven nexus]

hello guys

i want to know that i have a demo video of myself Now I want that video to go full ai ...like the movements and all remains the same but the background and other things has to change as per prompt.....is this possible?i saw a video on yt for the same but cant find it now....plz help

[opaque sand]

Hello. I don't quite understand how you added four frames to a single video.
Please explain...
I'm very interested in this idea...

[clever axle]

wasup my dudes

[uneven coral]

hey people

I like to make werewolfs

from pictures

hahah

[flint summit]

I was surprised to find that this monring I was able to make Veo3 videos with dialog in Gemini, even though I have zero Flow credits on my Pro plan right now. In the past it would only give me Veo2. Did something change?

[quaint lake]

Anyone have an idea of when they will add seamless looping (first and last frame) to portrait mode?

[ripe sun]

Hi everyone! I was generating an invertiew with "Veo2 Beta Audio" and the last 2 days it's impossible to generate a decent clip. Flow is not generating any clip with Audio and the video is awful. Is this a Bug? Anyone is struggling with something like this?

[ebon cradle]

I am waiting for this exact feature!...without it, everything is useless!

[quaint lake]

https://discordapp.com/channels/1172568727942860810/1418316534383906926

[quaint lake]

tell them

[quick rapids]

flow / gemini have different caps. gemini is "x per day" and flow is credit based

[iron matrix]

Is anyone having issues with scenebuilder similar to (https://discord.com/channels/1172568727942860810/1417712178114330674). If so lmk, i need to collect some more info on whats happening to fix things!

[stuck adder]

Veo 2 doesnt have audio

[copper fable]

Has something changed in last couple of days? seems i cant make avatar say some words, cause it says that its sexual content (ofc it isnt.)
But while i tried on one attempt it didnt give me error but the avatar said the same words in another language( english). If i try to make avatar say the exacts same words in my language, its block for sexual content again. its very strange. what can i do?

[ripe sun]

@lavish timber Thanks for the reply... I meant VEO3 fast and VEO3 Quality (both audio models)

[solid vault]

Hey, i am very interested in the Ingredients to Video feature of flow and saw it was announced to come out in june for veo 3, is there any updates on how long this still takes?

[crisp basalt]

come out in june?

oh u saw the announcemnet in june

[unique abyss]

My sci fi short Stellar Beacon, made with @FlowbyGoogle is an official selection at the Brooklyn Sci Fi Film Festival and will have its world premiere next month.

It's amazing that, thanks to Flow, in 6 months I went from being someone who loves film, but never made one, to someone who has a film accepted into a festival!

[main saddle]

AAAAA!! you're kiding - that's outstanding! so happy for you! let us know if we can catch a glimpse of it.. 🙂

[unique abyss]

Thanks--I've very excited! (And thanks to everyone on the team who made Flow/Veo) I'll post the film on YouTube after the festival is over, and share it here, as I want to let the festival have the "world premiere"

But right now, I'll post a link to shorter short, set in the same universe, in the gallery, in case anyone wants to take a look.

[main saddle]

yes please!!

[unique abyss]

Just want to encourage people, I know that sometimes, because this is all new, it can sometimes be a bit discouraging if a particular prompts fails, or the model gives us something unexpected. But, over time, we get better prompting, and also level up our skills and taste. I couldn't have made the film that got in the festival back in December when I got hands on with Veo as a trusted tester, but by practicing and using it every week, I got better at it, and was able in a few months to make a film that a festival liked.

[sweet falcon]

How are people adding text annotations for editing videos or images? I've seen some videos. Are you just uploading the video to Canva or something similar to add the text? Or is there a feature within flow where you directly add the prompts? I couldn't find any and it seems time consuming to download something and add the annotations everytime you need to suggest some changes

[peak rain]

Where do I go to submit improvement or feature requests for flow?

[main saddle]

you can submit in #1319424187730755636 or within Flow itself by clicking the 3 dots on the top right > send app feedback

[halcyon vector]

has there been any progress being made towards making it possible to select a last frame with veo 3?

[last thicket]

Sorry to bother sir, do you know why veo 3 with i2v feels always slow motion?

[kind bison]

I use PowerPoint - I'm sure lots of people use Canva (or similar).

[lavish field]

hah what the hell is going in with Veo today??..my prompts are completly being re-written by Veo...its making its own prompts not what I put in and im not new to this...all was fine last night today its gone off the rails..anyone else having issues?

[kind bison]

Got so angry with Flow tonight - made a clip with an image of woman I created in Midjourney - got prominent people blocked at first, but did the blur and resize and crop trick and finally got her through and made some decent clips. So, I upload a reverse angle shot of her from behind with a thumbnail annotation using the blurred front version that I just used, cropped to the woman's face, because I want her turn around at the end of the clip. Bam, the prominent people bug strikes back. So frustrating! It's the exact same shot that just got approved for a clip, but now it won't take it! Luckily I got it through on Fal. Why won't they fix this????

[crisp basalt]

take a break and go rest lol

[last thicket]

Hey

[velvet apex]

How to make intimate scenes for films with AI <@&1173646484311384114>

[last thicket]

You can't. Sadly

[velvet apex]

It defeats the purpose because if you’re making a film and suddenly we stuck with intimate scenes

[last thicket]

You could probably get to work to make the viewer think they had an intimate encounter about portraying them in bed talking something and turning off the light (no nudity, you can still make the shirtless man, something like that)

[kind bison]

At some point this censorship regime needs to fall. Filmmakers need the freedom to express themselves - we can't have huge corporations dictating to us what we're allowed to create - that's infinitely more offensive than any distasteful videos that some might generate.

[last thicket]

I can only agree with you in a text-to-vid

But it's very annoying to be honest, i think some of them are heavily censored because of CC processors

[sick prism]

any one know how to remove verafication useing phone, i want to use it and log in when i dont have a phone with me

Nanobanana is cool

[restive forum]

Hello why does flow ai not work on my phone and my computer while my age and check, the country so I am there is normally available (France) and I have my GIMINI pro subscription, you have explanations I send you the capture with.

[last thicket]

It happens to my mother's account too, she cannot access even tho its available in my country and she's 56

[restive forum]

Is there a technique to solve the problem?

[last thicket]

I'm not sure, is the account a new account?

[true pumice]

How long does it take to generate one 8 second long video with Veo 3 fast ⏩?

[last thicket]

40-50seconds

[bleak walrus]

How can there be no appeal mechanism for the issue of "prominent people"? When you are not using prominent people? The project just grinds to a halt, is that it?

[prisma holly]

Why my ai credits are not refilled

[safe wadi]

anyone has a proper prompt for making a guy dance ?

[turbid lintel]

ultra plan veo 3 fast unlimited till when?

[blazing mulch]

i believe it's not a promo and that's gonna be as is. One of the USP for Ultra 🙂

[blazing mulch]

LMAO. that has a lot of repercussions if allowed. People can actually use it to blackmail innocent people after generating fake intimate videos of their target and extorting them for money. Google doesn't want any backlash either for creating such a powerful tool. Great power comes with great responsibility, I'm sure you learned that from Spiderman already,. haha

[velvet apex]

I definitely understand all these things I am just saying Google needs to think someway out there to give justice to films. For eg I was recently making an ad where the scene was. A female actor was sitting on a chair in makeup Vanity van and a person was doing her make up. Now the scene was to make it feel that earthquake came and entire vanity is shaking, props falling down. As soon as I wrote this scene I don’t know how veo 3 says "we can’t create sexual content" also i definitely tried so many attempts with changing prompts nothing happened. As a paid user its concerning to make films on this platform

[blazing mulch]

I get what you mean, if i may suggest, you can try using Hailuo 2.0 / MiniMax for your "intimate scenes" although it doesn't have audio yet, they're not as strict as Veo 3, their filter is also slow, so if you get to generate it, you can download it right away before they censor it. ( just as long as they're not XXX or all the way , kissing / lingerie etc are acceptable. , if you want to explore more you can try some free source materials like Wan 2.0 or so 🙂

[blazing mulch]

It's hard to evade filters for veo 3 because they have 2 filters, one for the prompt and then one that filters the content as it is being generated/after generation. So there is no middle ground where you can actually swoop a successful NSFW generation . They're also pretty sensitive with babies/ harmful actions and etc.

[velvet apex]

@blazing mulch check this

Its in Hindi I don’t know you’ll get it or not

[crisp basalt]

For real 💯 there’s always plenty of alternative ai video tool people can use

[kind bison]

So sick of VEO 3 right now!!! Have a looming deadline for a client on a video and the prominent people bug has brought the project to a complete halt. All of the characters in the images are AI generated and don't resemble prominent people. Why the hell am I paying $250 a month for a service that doesn't work most of the time?! Even crazier, it accepted one of the images, then totally rewrote the dialog - not even close to the dialog in my prompt - then when I re-ran it, I got the prominent people block! What is even going on??!!

[turbid lintel]

for 8 second video stick to 18 word only

adjust the word on that bracket count

[kind bison]

That's ridiculous. I make many VEO 3 videos every single day, ever since launch. My prompts work fine when VEO 3 is running properly. It's just broken right now.

[turbid lintel]

well i use the 18 word rule and my video diolouge stay consistent till now

[kind bison]

It has nothing to do with the word count.

I've tried 8 different images now, 4 different people, 2 different angles for each, have applied grain and gaussian blur, cropped, converted back and forth between JPG/PNG - nothing is working, all blocked, on all VEO 3 platforms. The filter has gone haywire tonight.

[cold thunder]

hey everyone is there any possibility to create a filename before downloading veo 3 videos?

[rare raptor]

HAPPY 21/9

[crisp basalt]

https://tenor.com/view/sound-like-skill-issues-skill-issues-skill-issue-wii-gif-27389090

[crisp basalt]

I dont think so

[narrow lake]

I need the standard prompt for this, anything i enter generates crap

High quality crap

[errant grotto]

bro what happened

[pulsar field]

hello guys, i'm just trying flow vith veo3 in the free credit version and i like the output videos it makes but i would like to understand if i can do video long more then 8 seconds

[cold thunder]

no only 8 seconds per video

[cold thunder]

try to use ai to generate a prompt for you & optimize it everytime

[bleak walrus]

I've been trying to generate an Aussie rave scene during the late 80s. I've been at it for 20 mins now. While I appreciate the people of SA, and everywhere for that matter, the scene calls for Aussies. This is not the first time that I detected some form of "ethnic bias." Do an Ultra subscriber a favor and let him represent his own community why dontcha.

[sleek vale]

Can someone explain to me if the jump to and extend options found in flow for the video products are available in any API format? I can't see to locate the option anywhere. Thanks

[pulsar field]

i will go for pro plan