flow-general・🌊

[upbeat rivet]

It might be because, I haven't added instructions for the sound in the prompt.
For videos like these, what should I put in the prompt?

https://youtu.be/Faz3sNwL4rw?si=z5tmhWK5Yxoex3uO

https://youtu.be/G3uo32MNkZE?si=hjbaKHUd8Wq7m-yH

https://youtu.be/mcAJE7xLbxA?si=fy7RFUgDf_Ywaolp

[viscid fjord]

veo2 doesn't generate sound, people add it with something else, I a few tools have been mentioned here

[upper badge]

让她动起来

[strong karma]

https://m.youtube.com/watch?v=WFVLWo5B81w&pp=ygUTZm9sZXkgc291bmQgZWZmZWN0cw%3D%3D

I recommend recording your own sound FX. It's super easy and fun too!

[rocky matrix]

I've seen this as well.. anyone noticed it?

[craggy pawn]

Does anyone have a way to access veo 2 ?

[hexed marsh]

yeah

[craggy pawn]

for free?

[hexed marsh]

yep. through early access

did u fill out the form?

[craggy pawn]

yes about a week ago

[hexed marsh]

give it some time

also sign up for the whisk animate

[craggy pawn]

is it veo 2?

[hexed marsh]

yeah it utilizes veo 2 under the hood i think

[craggy pawn]

Do you have an idea of the expected time

[craggy pawn]

oh nice

[hexed marsh]

sign up for that

and make sure u look out for any other announcements in the server

[craggy pawn]

i will trye

[craggy pawn]

ok

[craggy pawn]

i sign up here

[hexed marsh]

good luck, hope u get access soon

[craggy pawn]

thank you

[craggy pawn]

Do you have access to veo 2?

[hexed marsh]

yes

[craggy pawn]

How much do you rate out of 10?

[hexed marsh]

10

[craggy pawn]

Is it really that great?

[hexed marsh]

it's best model for AI video generation hands down

there is no other model that produces the consistency and overall quality of VEO 2

[craggy pawn]

Yes, I have watched a lot of videos about it on YouTube

[craggy pawn]

You're right. 100%

Does it support 4k

[hexed marsh]

check out #flow-videos・📹 to see some videos

and if you have prompts, send them in #flow-tips-n-tricks・✏️

i think its 720p for now but the quality is so good you won't even notice

[craggy pawn]

Yes, I'm see it

[hexed marsh]

everything else makes up for it

[craggy pawn]

Yes, even the videos it generates appear exaggeratedly realistic!

[hexed marsh]

it impresses me every time i use it to say the least

whisk animate will be a lot of fun too

[craggy pawn]

Yes, it's a strong model, even stronger than sora

[crisp basalt]

is it just me or

im noticing the artifact in the generated video now

like it become more obvious now

[near gate]

i just noticed it now.. it ruins everything :/

[hexed marsh]

hm i haven't noticed that

[crisp basalt]

Check what zen posted in gallery

[waxen lodge]

i need videofx access please

[nova ferry]

happening since a few days at least 50% of my renders

[rocky matrix]

Yes same here. Check the background. @main saddle I just wanted to also make you aware of this and so many generations that is not useable due to artifacts

[crisp basalt]

Feel free to post your vid in thread I made in bug channel https://discord.com/channels/1172568727942860810/1352132020368572437

[dim glen]

Guys is it possible to upload a picture and turn it to a video on Veo 2?

[rustic topaz]

please don't ban me

[crisp basalt]

@edgy holly Poe got image to video mode now 👀

[rustic topaz]

have you tried it yet? how was it?

[crisp basalt]

Oh I’m not subscribed to Poe , I’m on Freepik plan atm

[rustic topaz]

damn I want to know how good it is compared to Kling

Kling is god-darn expensive to me atm

[crisp basalt]

Oh

When it comes to visuals, Veo excels at generating realistic visual effects and accurate physics-based collisions.

Kling 1.6 Pro is impressive, but what makes Veo stand out is how naturally the video looks,

I tried Veo image-to-video on Freepik a few days ago, and I’m really impressed with the results. It still struggles a bit with following prompts, but overall, the experience has been great.

[rustic topaz]

the biggest problem I get with veo is that it keeps changing scenes, it gives cuts and edits I don't want

[crisp basalt]

oh so far i havent got any ,,, jump cut or changing scene especially with image to video mode

[mental mulch]

Hey how do i gen a video thats not in slow motion

[rustic topaz]

i mean, you can ask the model to do fast motion, but it surely will have some ai shenanigans to it
you should try repeating the word fast motions, fast movements, etc

[rustic topaz]

to be honest, does anyone even use TV mode?

[sharp field]

Does anyone know how to get into early access to VEO2?

[rustic topaz]

the register form closed a while ago

[crisp basalt]

Yeah, the non early access users use it all the time

[hexed marsh]

made me chuckle ngl

yo did anyone ever get access to whisk animate via the recent form yet?

[crisp basalt]

that was corny

[solid tide]

Yes, same here! Looks like Moiree, or as it was filmed from a screen. More than 75% of my shots have it, would say since 3-4 days

Look at here cloth here

Thats what usually might happen with older cameras, when fine patterns are filmed, you get a moiree effect.

[near gate]

@main saddle I have been waiting for months pls invite me my email is: [e-mail] pls invite me!

It feels like eternity

😧

Btw please do not kick me out of the server for begging but I have been waiting tooo longgggg

[hexed marsh]

lool

[dim glen]

Is it true that Google is planning to add Veo 2 to It's AI studio?

[crisp basalt]

They’ve been working on it for a while , it was called codename robin

[crisp basalt]

Ohhh TIL it’s called moiree

[dim glen]

That's great.

I think Google is winning the Generative AI race.

[crisp basalt]

Looks like the moiré effect isn’t showing up anymore, at least based on what I’m seeing in my latest generated video.

[balmy mica]

Does anyone know when the next open registration is?

[crisp basalt]

I don’t even bother opening the videofx library page on my phone anymore, it always refreshes itself before any video loads, then suddenly crashes with a error message.

[crisp basalt]

Nobody really knows for sure, but if you’re looking to try out the Veo 2 model, there’s another signup form available,it’s for Whisk Animate, which uses Veo 2 to bring your generated images to life.

Never mind it’s closed now lol

It was opened for like almost one week

[balmy mica]

🥲

I’m eager to use it for my business

I’ll check it out, tysm

The wait list with VideoFX?

What’s the difference by the way? (I’m new to all this) If I can use it in Poe, is it still a good idea for me to join the wait list on VideoFX for example?

[crisp basalt]

Yeah it’s a good idea to join the waitlist anyway,

[balmy mica]

Oh wow

Is it possible for me to feed it a png image of clothing that I designed and have it generate a video with that piece of clothing in it?

[crisp basalt]

Maybe Veo 3 😝

[balmy mica]

Oh wow

What a time to be alive

I haven’t

[balmy mica]

I watched one of his videos and got hooked, I’m still binging 😂

[mental mulch]

Can I get some help

how do i have videos generate on videofx without having the videos come out with slowmotion

is there a certain prompt im missing

[surreal flicker]

Try using keywords such as fast, motion and etc

[mental mulch]

like fast motion

[surreal flicker]

Yes

[mental mulch]

how do I make a keywords

[surreal flicker]

Ah just use those types of words in the prompt

[mental mulch]

oh I have

it doesnt work lol

[balmy mica]

I see 2 versions of Veo 2 on Poe, there’s Veo-2 and Veo-2-Video but both are marked official and only one is by Google, what’s the difference?

[mental mulch]

Good question

[balmy mica]

I’ve been trying to feed it my fantasy prompt but it ends up coming out cartoonish/3D rendered with awkward movements

[crisp basalt]

Yeah that’s another thing Veo model struggles atm, well I shouldn’t say struggle but it’s more of how user prompted it.

[balmy mica]

Hmm, alright I’ll try it out

Should I use a prompt generator to help me?

[young moon]

does videofx only generate 8 second videos?

can it generate youtube video lengths like 10 mins or sth practically for those who have access to it?

[surreal flicker]

Yes

From what i’ve seen

[young moon]

oo, is there a way to practically generate youtube video clips like 10 mins practically by appending 8 second videos or is it just not practical and just manually video edit yourself for best effectiveness?

[surreal flicker]

For now you can only combine the 8 second clips separately to form some sort of video

[young moon]

oo i see, thanks for the information man 🙏

[surreal flicker]

there was this instance where a website provider had veo 2 along with an image to video option which google shortly told them to remove

Implying i2v may be coming out with the opportunity to extend the previously generated videos

[lusty nymph]

Where is this from?

[crisp basalt]

And luma still can’t do it right with their generated video

[wet kiln]

This new story im generating with ChatGPT and Veo 2 is gonna be sick I can feel it

[wet ice]

guys ... we got a new update ?

[crisp basalt]

WOAH

HOLY MOLY @edgy holly ITS HERE

image to VIDEO

ACTUALLY NO ! LMAO

Ahhhhhh finally

[wet ice]

🙋‍♂️ 🫡 😎

[crisp basalt]

this is a very good update

so now your generated video from the past is shown on main page too this is actually nice!

[wet kiln]

OOOOOOOOOOOOOOOOOOOOOOOOOOOH?!!!

Great! Time to create 38 scenes for this 5 minute story :d

[crisp basalt]

video file are named like this now , very nice

[wet kiln]

It feels like its faster

[crisp basalt]

it is!

[wet kiln]

:gasp:

Its already at 30%

and i just started generating my next scene

Oh dang I thought the generations got more cursed and then I realized I accidentaly clicked Veo 2 - Fast mode

[hushed rain]

If anyone wants to check it out, I just released a short dark fantasy film: ALL Veo-2!
https://youtu.be/YDlME4qvER8

[hardy halo]

Wow you guys are fast! Just launched a new UI update. Let us know what you guys think 🙂 #flow-announcements・📣 message

[crisp basalt]

@hardy halo generate five videos !!! That's very generous of you guys!!

[hushed rain]

and by "film" I mean...short. It's like 2 minutes. haha

[crisp basalt]

iT'S Even ui mobile friendly now!!

@paper inlet come take a look at the new update

[fierce nimbus]

We listen we listen hhaahah

[hushed rain]

haha, seriously? I've been cranking on an AI Film for the last 4 days and we get this update 30 minutes after I release it?! hahah (kidding: thank you very much!)
I guess I gotta go make another one now...

[wet kiln]

This is AMAZING. I'm actually in the works of creating a film too

I hope it comes out great

[crisp basalt]

oop - i think those link contains y our .. stuff thingy

[hushed rain]

apprecate it!! LOTS learned here! If you go into VideoFX Prompts, check out the FWD that @crisp basalt did (not sure who the original author was) on Scene Split in your prompt. That is a KILLER hidden feature.

[thick umbra]

yay! new ui! now we need image to video!

[thick umbra]

hey tim!

kinda image to video? but i can't import my own images :<

[crisp basalt]

its actually text to image to video haha, im sure its not ready yet for early access users.

[thick umbra]

its on the API though. so maybe they are still making it better on labs idk

[wet kiln]

https://www.youtube.com/watch?v=TG1BNR-wC9I Trailer of the story I'm creating

[crisp basalt]

Overall, the new VideoFX UI update looks much cleaner and more polished.

The layout is more organized, and the larger video previews make it easier to browse generations. Overall, it feels more modern and user-friendly compared to the previous version. I'm really happy with the update ❤️

[wet kiln]

Yeah the previous version felt like watching videos on a card or something

[crisp basalt]

Really appreciate all the improvements you guys made to the VideoFX tool, especially the UI!

[verbal hatch]

did the flickering effect disappear after the update?

[wet kiln]

NnnnnOPE

[wet kiln]

Oh dang it's kinda worst than before

[verbal hatch]

has the quality gotten worse?

[wet kiln]

No

[crisp basalt]

invm i think its fine

[verbal hatch]

the flickering never went away...

[wet ice]

what flickering guys?

[wet kiln]

It's called a moiré effect. A visual effect that two similar patterns, when superimposed, create a new, often unwanted, pattern of wavy or striped lines

It was barley visible but now it's slightly more visible

[crisp basalt]

i had to put it to output only one video for text to image to video so it wont introduce jumpcut

at least from what ive tried so far

[hot bough]

gosh I need access

[crisp basalt]

from what ive tried so far yeah

i dont know dont quote on me that tbh, but, i tried with 2 output only

only 1 out of 2 didnt do jump cut

[fierce nimbus]

losign which prompt?

[crisp basalt]

Oh yeah i was about to say that

[shrewd bone]

ah ye, kinda odd but just gotta click "Use Prompt"

seems like the library loads a lil faster

[wet ice]

great visual ex... now i understand

[crisp basalt]

One thing I really like about the new update is that in the text-to-image-to-video process, it now generates four image variations, and we get to choose the one we like best.

[verbal hatch]

I wonder if they'll ever remove this flickering effect? ​​The update is certainly cool, but I'd like to see the flickering effect gone.

[crisp basalt]

just a suggestion but is it possible to add the delete video button on the main page too?

That way, it’s easier to delete a generated video we don’t like without needing to go all the way to the library page

[ashen yoke]

Say more about what you mean by overshot it?

[mental mulch]

The UI looks okay

Just dont like how laggy it becomes now

[shrewd bone]

ballet

(sped up x2)

[winged plinth]

My fav style.

[ashen yoke]

I'm not running into lag -- I kinda miss the gray'd in little prompt space because I like that feature in Whisk .. but it's not a complaint

[wet kiln]

https://www.youtube.com/watch?v=IvsFkn4j8D0&ab_channel=Veo2Movies My newest story

[fierce nimbus]

which prompt space?

[unique abyss]

Like the consistency you are able to achive.

[wet kiln]

Thx!

[ashen yoke]

The space with grayed out sample prompt

[hardy halo]

Everything is public and free to share!

[visual sapphire]

Just need someway to stich videos together + continue from last frame etc

[tacit aspen]

Amazing Update!

[sage hare]

new UI is goated

[torpid cedar]

Interesting UI. The text to image feature seem bugged tho

Just release the img2video. 😁

[sturdy atlas]

is there something in particular that is bugged with text to image?

[torpid cedar]

@sturdy atlas it doesn't complete the full task. I generate the image and then the scene gets bugged error related

Whisk seems to have the same problem

[sturdy atlas]

hmmm will investigate, thank you!

[hexed marsh]

whisk animate 😮

[ripe jetty]

I want pizza now name dropping Domino's

[small lake]

Does anyone have good prompting resources for using VideoFX to generate realistic videos of people?

[upbeat rivet]

@small lake I believe that the resource you are looking for is the Google Gemini app. I used it to create videos of Bruce Lee fighting characters. I believe that it is okay with doing Bruce Lee because, no one owns his image. But, I don't think you would be able to create videos of someone else who is still alive.
My newest creations :

https://youtu.be/Faz3sNwL4rw?si=z5tmhWK5Yxoex3uO

https://youtu.be/G3uo32MNkZE?si=hjbaKHUd8Wq7m-yH

https://youtu.be/mcAJE7xLbxA?si=fy7RFUgDf_Ywaolp

[ornate kelp]

anyone else having issues with text-to image-to video. Generating the image fine, but it’s not generating the video. Just gets to like 10% and disappears

[crisp basalt]

What was the image and prompt if you don’t mind

[ornate kelp]

It seems to be for everything I try, so it’s def not a filter issue I.e “hotograph capturing an elegant grey Siberian cat lounging gracefully on a plush, modern couch. The cat’s luxurious, dense fur is a silky gradient of shimmering silver and smoky charcoal tones, accentuating its majestic appearance. Its vivid emerald-green eyes, almond-shaped and expressive, gaze thoughtfully ahead, reflecting gentle indoor lighting that softly illuminates its whiskers and highlights intricate facial details. The ears, tipped slightly forward in a relaxed yet alert position, add to the feline’s sophisticated charm. The cat sits poised with a serene dignity, its thick tail neatly curled around its paws, emphasizing its regal and composed demeanor. The couch beneath is upholstered in soft, textured fabric with subtle hues of muted grey and taupe, providing a complementary backdrop that enhances the cat’s natural beauty. The setting feels cozy, peaceful, and inviting, expertly framed to showcase the refined textures, subtle shadows, and meticulously detailed clarity characteristic of professional animal photography”

I get the image, then it when I prompt for the motion it just disappears. For even something simple as “the cat leaps off the couch””

[crisp basalt]

Oh

You might wanna try this @ornate kelp

[ornate kelp]

Thanks will do!

I feel like I had a similar issue in the past when I was logged in on 2 devices, but I can’t tell with the new UI

[ashen yoke]

Thanks for these latest posts -- I cleared my cache because I was running into similar problems running on Firefox; the image generating but then crashing out when going to video

[obsidian summit]

Did that fix the issue?

[ashen yoke]

Spoke too soon; added it to the bug list

[obsidian summit]

I'm experiencing it too.

[ashen yoke]

nope 😦

[ashen yoke]

well that's no good

[obsidian summit]

I guess the positive thing is I'm not the only one experiencing the issue.

[crisp basalt]

Have you tried other browser?

[obsidian summit]

Tried both chrome and firefox

[crisp basalt]

hmm

if possible, try to screen record the steps leading up to the crash

Or at least take a screenshot of the crash itself, it might help the labs team investigate it more effectively

[obsidian summit]

Just posted a screen record of the issue.

[plain lava]

hi! I've been having the same imageToVideo failure at ~8-10% all afternoon. It was working for me this morning.

[silk hill]

you have image to video?

[plain lava]

I meant creating an image first in VideoFX, and then using it as "first frame" for my video prompt.

[silk hill]

bah

[crystal ermine]

Guys any of you got access to both VideoFX & Whisk Animate? Could you guys tell me are there any differences in quality, functionality etc? And maybe show us the same action done by both VideoFX & Whisk Animate? Thanks.

[crisp basalt]

It’s basically Imagen Fast Model

And Veo Fast model being used to animate

Don’t quote me on that

[paper inlet]

In Whisk, I added three subjects, an otter and a snail generated by Whisk, and an image of a metal puzzle ring from the internet. I put in a simple prompt to create the two composite images. I animated the one on the left, reusing the same simple prompt. Then I took the prompt generated by Whisk and put it into VideoFX (Veo 2 - Fast Mode).

This is the prompt generated by Whisk: "A smooth-coated, dark brown otter sits on a light beige sandy beach, attempting to solve a complex silver metal wire puzzle, while a light brown shelled snail with a grayish-brown body watches from nearby; calm turquoise sea and blue sky background.

A smooth-coated otter with dark brown, wet fur sits on a light beige sandy beach, its body slightly hunched as it manipulates a silver metal wire puzzle with its paws. The puzzle is composed of several interconnected loops and wires, forming a complex three-dimensional shape. The otter's dark eyes are focused on the puzzle. Nearby, a snail with a light brown, spiraled shell and a textured grayish-brown body sits on the sand, seemingly observing the otter. Small waves lap gently at the shore, leaving a thin line of wet sand. The sand is relatively smooth with small footprints and debris. In the distance, a low-lying, tree-covered headland extends into the calm turquoise sea under a clear blue sky. Several small, brightly colored buoys are visible in the water. The lighting is bright and sunny, suggesting midday."

Veo2 (regular mode)

[crystal ermine]

I think I actually prefer Veo2 fast mode result. Thanks guys @paper inlet @edgy holly @crisp basalt I guess I still have to get Veo2. Hope Georgia re-open the reg.

[crisp basalt]

Whisk Animate uses veo2 fast mode btw

[crystal ermine]

Yeh I read your previous post, but there's a difference in what's produced by Whisk Animate & Veo fast on Video FX as shown by heatherhelen 🙂

[crisp basalt]

I forgot whisk animate specifically uses VEO 2 Fast Mode Image2Video model

[crisp basalt]

This one was text to video right?

[paper inlet]

I think the Whisk: text-to-image > Animate workflow is similar to the Veo 2: text-to-image-to-video workflow. In other words, the workflow is your prompt > image generated > model generates a new prompt for the video > video is generated.

I generated both of the Veo2 videos, fast and regular, with straight text-to-video, using the same prompt auto-generated by Whisk.

Here's the text-to-image-to-video workflow in VideoFX with Veo 2 - Fast Mode, where you generate the image first and then create the video. Unfortunately, I can't get the video to generate. It's not working. But, you can see that the text-to-image part creates an image that looks more like the Whisk generated image.

So for the workflows, it's either 1) text-to-image-to-video workflow available in Whisk Animate or VideoFX; OR 2) text-to-video available in VideoFX.

[forest raptor]

woah!! now videofx saves my generations

nice, there's a gallery section now

[forest raptor]

oooooooops

Btw, "female astronaut 16 bit videogame" is blocked

but "astronaut 16 bit videogame" isnt

no female astronauts apparently

back in the kitchen

ah, "astronaut woman" is fine

ah no. blocked.

now the videos also have a different name, no need to rename them every time!

[iron matrix]

Were having some issues today with Video generation! Please hold tight and hopefully we will have things fixed soon!

[crisp basalt]

Thank you for the update.

[silk hill]

anyone getting this when they try to make anything in whisk

[solid tide]

For me regular T2V Veo 2 is working, but Text to Image to Video is only working with Veo Fast

As reported above, makes like first 5% then gets canceled.
Just reporting, as its stated already going to be fixed. 🙏

[wet kiln]

Really? It's working for me.

[crisp basalt]

converted the videofx web into electron based app,

[near gate]

oh image to video, please come to us🙏

https://tenor.com/view/cute-kitten-begging-attention-gif-8380139

[forest raptor]

yah

[true locust]

Slow vertical tracking shot: The camera moves upward, revealing "Tekzo", a 10-year-old boy with spiky electric-blue hair and glowing cyan eyes, standing in his secret high-tech garage. He gazes in awe at his cyber-tech glove, which emits neon-blue energy trails. Floating holograms display futuristic schematics around him. The dimly lit garage flickers with monitors, scattered gadgets, and robotic parts, while sparks of electricity and digital particles swirl in the cyberpunk atmosphere. Animated movie quality.

[iron matrix]

just to be sure, when you say "still not working" you mean you're getting "Something went wrong" toasts. Or is there something else going wrong?

[crisp basalt]

Mine was “something went wrong”

[crisp basalt]

NO WAY

IMAGE TO VIDEO IS ON

@near gate

[surreal flicker]

I’ve heard their image generation model is really good

from a video i saw

[crisp basalt]

wait its generating very fast, im using normal model right now

[crisp basalt]

i mightve spoken too soon

[iron matrix]

Sorry for now you can only use Veo 2 (not Veo 2 - Fast Mode) with I2V and T2I2V! This will hopefully change in the near future

[crisp basalt]

Ok so when u picked Fast Mode, it will show that toast

Ahh! I understand now

Thank you so much yall ❤️

[iron matrix]

this should be fixed now, please give it another try!

[hardy halo]

Image to Video is live!! Know y'all have been waiting for it... Had to really fight to get this one out so please read the disclaimer and generate responsibly!!! #flow-announcements・📣 message

[near gate]

LOVE YALL🫶🫶 THANK YOU SO MUCH!!!

[crisp basalt]

Just from this message, I can already tell how hard you guys had to fight to make it happen. Major respect 🫡

Yeah, I think this one still needs some prompt tweaking. In my case, it added a jump cut to something totally unrelated to the image I uploaded.

let me cook again with the prompt

Also just want to say I appreciate how the thumbnails for image-to-video or text-to-image-to-video don’t just show the first frame, but still clearly represent that it's an image-to-video generation. It’s a subtle touch, but really helpful. 👌

[maiden depot]

wow it's so fast too! compared to t2v. this is fantastic 😎

[hardy halo]

Please share your outputs and feedback! We want to hear about what works and what doesn't

[maiden depot]

Just wish that the prompts were always visible at the bottom of every batch when i leave the page and return. That's a super useful addition

[hardy halo]

Do you mean you wish they stayed in your prompt box?

[crisp basalt]

referring to this actually!

would love to see it stayed after refreshing the page

[hardy halo]

Got it! We'll add that to the list

[wet kiln]

Yaayyy now I can create visual stories more faster now! :DDDDD.... As long as theres no weird cursed stuff in the videos that is.

[maiden depot]

agreed. especially since we have the ability to go fullscreen with the new ui anyways. but it would be nice to be able to have a ui toggle between like small and big thumbnail and then like detail view as we have a whole history now. or just having the videofx toggle in the library page would be nice.

[narrow dove]

THANK YOU, FINALLY. I knew this moment would come!

[swift mauve]

Thank you, I waited for this, I couldn't see what I really wanted to create, which I couldn't with text to image to video or text to video, but now it is the best, image to video. 🤗

[hardy halo]

Sharing this image I've loved doing Image2Video with. I've been prompting "tv static turns into x" and you can make x whatever tv show you want!

[torpid cedar]

Img2video isn't working

[stray copper]

Can you say which error do you get? In console

[stray copper]

My first try also not worked, as I understand, need upload exactly 16:9 image, you can try

[torpid cedar]

It loads up and disappears. I check library blank with 8 secs

[crisp basalt]

actually i tried upload image thats not 16:9 exactly, it still went through and generated, downside is its looks stretched

[stray copper]

Try exactly 16:9, maybe it will help you

[torpid cedar]

Hmm ok

I will have to redo some images now

[iron matrix]

we might be able to help with this in the near future 👀

[hardy halo]

Can you share what image and text prompt you used?

[torpid cedar]

Seems 16:9 only works.

This was just a test

[crisp basalt]

looks sick regardless

[verbal hatch]

This is a very cool update that adds generation Video of his image

[paper inlet]

I tried I2V with a photo of a painted car, attempting to bring the painting on the car to life. I asked Gemini to help me optimize the prompt. The first image is the photo of the car. The second image is a screenshot of the VideoFX library showing the two videos. For some reason, when I download the one on the left, the thumbnail appears to be for the video on the right, although it still plays as what it is. Also, the video on the right is upside-down. 🙃 Here's the prompt: "A sunlit street scene captures a vintage Volvo 240, its boxy silhouette adorned with a whimsical, abstract painting reminiscent of Monet's water lilies. The car's surface is a canvas of swirling blues and greens, suggesting a tranquil pond. Soft pink and white daubs hint at budding flowers. The painted water is gently rippling, the floral shapes slowly unfurling into vibrant lotus blossoms, their petals opening in a dreamlike dance. The surrounding urban environment remains still, contrasting with the ethereal motion on the car."

[stray copper]

nice

[crisp basalt]

the model was struggle cuz the image wwasnt 16:9 i assumed haha hence the upside down

[torpid cedar]

Yeah but I wanted to use my other images that aren't 16:9

[torpid cedar]

My initial image and it's not 16:9

Lol you can see where i was going with it

[paper inlet]

The upside-down one.

[crisp basalt]

Omg

[stray copper]

)))

[crisp basalt]

at least u can manually rotate it later 😅 but still thats so bizarre

[hardy halo]

I think its becuase the file size is too large

Sounds like we really need cropping 😅 Coming soon, stay tuned!

[torpid cedar]

I mean if it's 1:1 I'm fine with that as well

[upbeat rivet]

When I try to use the Videofx, it takes me right to my library. How do I get to the prompt?

[torpid cedar]

It seems to fail on a lot of stuff . This image is 16:9 and just said cinematic. It loads up and fails like the rest@hardy halo

[charred dock]

Guys im happy, because image to video finnaly here, honestly i've been waiting for it)

[lusty nymph]

@main saddle Seeing your status "Need coffee" I got you 😅

[main saddle]

THANK YOUUUUU

[rigid venture]

IMG2VID - Taking an image of a desk I had as a concept for a project and got some cool perspective shifts

[forest raptor]

ooooooooooooooooooooooooooooooooooooooooh

maybe I can try to make the upload image work as a reference to video! let's try!!

first test: ok, must consider the size of the original image

test 2: seeing if it can be used as a reference-to-video tool. almost

test 3: fixing the image size issue (also what was that at the end?)

[wet kiln]

An AI hallucination

[maiden depot]

so something that would be such an incredible time saver here I've noticed, would be if when we "use prompt" is if the reference image that was used with the prompt also was copied over instead of having to re-upload everytime i want to change something

[sharp harbor]

biggest issue with IMG2VID so far is just sticking to the original image. It loves to cut to a completely new shot very quickly most of the time. I'm sure future iterations will fix this!

[sage hare]

What a great day to be a veo tester

[edgy pasture]

There is no way I2V is here

is it coming to whisk animate too?

🔥

[edgy pasture]

I wish I was one lol

[weary tendon]

@main saddle YEEEIYYY!!! I2V works so well. Great job. One question. I noticed the "seed" input is gone. Will it come back? 🥹

[weary tendon]

https://tenor.com/view/ohh-gif-13342724363430229217

[lusty nymph]

If you click on generate image option instead of upload its there 🙂

[ashen yoke]

Super woot on the new image to video feature, Google Labs humans of humanity

[solid tide]

Not sure if its the right place for UI Improvements:

➡️ The Prompt window in the middle of the screen (desktop)** should really get a drop down / fold down button**,
so it is not all over the place, when using /reusing long prompts.
💡 Only way to make it small again atm is to delete the prompt in the window, which is a bit annoying

[edgy pasture]

Hi, good night! I was wondering if the YouTube version of Veo2 is the same as Veo2-fast. Does anyone know? Thanks

[mental mulch]

When I try to do a image to video generation from Veo 2 - Fast Mode, it'll say "Switched to a compatable model. Please submit again to confirm." and when I do so, it'll say "Oops, something went wrong.". And it persists every time. I have tried refreshing, I have tried different images. It will not work.

I also tried Clearing cookies

am I the only one getting this issue

Text to Video works fine.

It's just I2V

Guess im stuck with Text to Video until a fix comes in

No matter what image I use, same issue

So its persisting

[mental mulch]

i just want to use this damn thing

😦

[outer aspen]

Hey this is amazing!

[narrow dove]

Wow, what a quality! Now you can create consistent videos! THANKS 🙏🙏🙏

I took my little brother's toy as a reference.

Unsuccessful animations

Successful animations

[crisp basalt]

@wet ice

[wet ice]

Thanks for clarifying 🫡

[wet ice]

P.s this implies we need to be careful not to reach that monthly limit too fast? did that change considering the above?

[crisp basalt]

I'm not 100% sure if that impacts the monthly limit. Might be worth checking with the team just to be safe

[cosmic venture]

Omg image to video! Thank you so much google I cant wait to show my friends videos from my art

[mental mulch]

I cant use it

Whatsoever it will always say Something went wrong

Im still having this issue like 4-5 hours later

[sage hare]

I got an idea

What if you screenshot the end of each video, and put it in img to video

And then attach all the extensions together

Boom endless vid

[paper inlet]

Or select a video to start from and the last frame is auto-selected, so you can bypass the screenshot step.

[sage hare]

Wdym

Is that really a feature?

Pls english

[paper inlet]

It's not a feature. I was just adding to your idea.

[sage hare]

Oh wait im stupid

XD

[paper inlet]

Even more advanced would be a way to select the exact frame from the video you want to start from.

[sage hare]

That doesnt work btw

[paper inlet]

It would be a new feature.

[paper inlet]

#1349053519273001092 message

[thick umbra]

NO WAY AHAHZWHBUD NJLWKABTGKNFLMS;ZX\\

SO HYPED

[molten bear]

Hello, how do i get access to VideoFX?

[main saddle]

hi!! welcome to the Google Labs community server appreciate your enthusiasm to try VideoFX!

the best way to sign up for VideoFX access at this time is to join via the VideoFX waitlist. if you're looking to have a video generated, we have a wonderful pool of community members in #flow-tips-n-tricks・✏️ who might be able to generate one for you! you'll be able to find this channel and more within the VideoFX section of the server (where we are now) -- here's the breakdown of those channels:

#flow-announcements・📣 - for timely announcements about VideoFX.
#flow-general・🌊 - for general VideoFX chatter!
#flow-tips-n-tricks・✏️ - brainstorm prompts, and share your prompt requests with VideoFX users who may be able to generate a video for you (please ask kindly!)
#flow-videos・📹 - to showcase & view VideoFX generations
#1319424187730755636 - a place for members to submit VideoFX bugs 🪲

please let me know if I can answer any questions!

[molten bear]

thanks for the info

Oh ok, interesting, thanks 👍

[wet kiln]

@crisp basalt what are they adding?

[crisp basalt]

Credits

[wet kiln]

You're kidding.

[crisp basalt]

I mean, it’s expected, even Whisk Animate has “credits”. For early access users, they’ll just reset monthly once you’ve used up all your Whisk Animate credits.

[wet kiln]

Wait that has credits?

[crisp basalt]

Yeah

[wet kiln]

A friend of mine with Whisk Animate told me that

I thought they were lying

[crisp basalt]

It's not clearly visible in the interface at the moment, for the early access users ofc. But once the tools are released to the public, an actual credit system will be implemented 💯

[wet kiln]

Oh good god

[crisp basalt]

I don’t see it as a bad thing, it’s normal.

I'm all in and excited to see where this tool takes us ,I'll keep supporting it all the way, from early access to the day it opens up to everyone

[wet kiln]

Breathing is free 😮‍💨

Aww :c

[fierce valve]

this is cool! what was your prompt? did you give camera direction or did the model just freestyle?

[rigid venture]

'slow zoom in desk, slow zoom on desk around monitor, art, animation, 2d'

thank you also! :)

one shot with just that on these too. less clean versions but still pretty awesome to see with it^^

I'll share some more stuff once i have some bg scenes developed, the thought of animating over the top and it not be completely static really fuels me hahah

[dim glen]

Hey, has anyone tried Image-to-Video? How accurate is it with people? If it's good, I'd love for someone to help me test something. Thanks.

[surreal flicker]

You can request your prompts in #flow-tips-n-tricks・✏️

[rustic topaz]

when can we have Keyframes and negative prompt XD

[near gate]

lowkey wish we could train our own characters soon with Veo 2, but that’s probably not happening anytime soon😪

[crisp basalt]

There’s a workaround for it but I haven’t tried it

[dim glen]

The I2V?

[surreal flicker]

Ye

I2v prompts

Send the prompt and image in that channel

[dim glen]

Ok thanks lemme do it.

[crisp basalt]

I can confirm that workaround actually worked ahaha

[analog sedge]

Img to vid pretty underwhelming

[hushed rain]

the Split Scene trick? Used it quite a bit. Can confirm it works.

[crisp basalt]

yeah ive tried it too, it works really well

[torpid cedar]

Waiting patiently for the fix on img2video. None of my images working just goes into oblivion...

[winged plinth]

All i2v?

[torpid cedar]

Basically even at 16:9

[silk hill]

is there any way to get image to video to not cut to a completely different shot 😦

[edgy pasture]

Oh I recently saw your video and tried it. Works great!

[mental mulch]

It was situated tyvm tho

[mental mulch]

It was taken care of

[crisp basalt]

thats good to hear

[mental mulch]

I cant make my videos in real time/fast motion

No matter what prompt i type in, itll be slow

[paper inlet]

What is your image and prompt for I2V?

[crisp basalt]

Maybe don’t prompt it to be fast or slow, just leave it like that and do 4 output

[mental mulch]

https://x.com/TribalChiefMike/status/1906556580817998016

Michael (@TribalChiefMike) on X

BREAKING: A Fan-Made #Yugioh Movie Project is now in the works, courtesy of @GoogleDeepMind's VideoFX/Veo 2 AI Generations. Planned to be released later this year.

X · Mar 31, 2025, 3:58 AM

[visual sapphire]

is the stretched videos from img2vid fixed yet

[mental mulch]

It hasn't been stretched for me

[mental mulch]

Is there a limit on generations on VideoFX?

It keeps giving me Something Went Wrong again

[crisp basalt]

When there’s a limit it will let you know

[crisp basalt]

This one is probably the model safety filter doing its thing

[mental mulch]

Its persisting

I dont understand what im doing wrong

[crisp basalt]

does it persist when u do text to video

[mental mulch]

No

Just image to video and text to image to video

[crisp basalt]

yeah that means safety filter doing its thing

what was the image and whats the prompt if u dojntmind

[mental mulch]

Oh I see, Im just trying to make a movie and I guess the safety filter is preventing me from making dark scenes

Which is understandable

[crisp basalt]

The safety filter’s suuuper sensitive to some words in the prompt, even if the user didn’t type anything sus, it still gets flagged. The image-to-video thing got this "prompt enhancer" thing going on, where it basically looks at both the image and whatever you typed in the box together.

So even if your prompt clean, the combo might still trip the filter.

[mental mulch]

Okay so what should I do

just do a whole new image then?

[crisp basalt]

i mean ,

For the img2vid feature, think of the prompt box like giving instructions. Just write your prompt as you normally would, but make sure to include something like (Do not include any words that might trigger the safety filter) at the end.

In my case, i did it like this

[mental mulch]

It's still resisting

[crisp basalt]

can u dm me the image

[mental mulch]

Its just a Stone Carving

of a female

[crisp basalt]

safe to assume that it detect something from the image that cause it to trip the safety filter

[mental mulch]

I tried a news reporter talking and it still resisted

[crisp basalt]

same image?

It would be impossible for you to specifically hit limit just for img2video feature only

[mental mulch]

I mean whats the limit

[crisp basalt]

Nobody really knows. A few people here have actually hit the limit, but the exact video quota isn’t clear. It resets monthly, but how many you can make before that reset is still a mystery

but those who hit the limit can still generate videos using Veo Fast Mode, assuming the limit they reached was for the normal mode.

@mental mulch

[mental mulch]

I cant even generate images

On VideoFX whats going on???

[crisp basalt]

@mental mulch

[mental mulch]

Im on Opera GX

[crisp basalt]

you can still do it lol

[mental mulch]

I hard refresh and cleared

still resisting, I cant even generate images anymore

[crisp basalt]

If you’re still having issues, go ahead and create a thread in #1319424187730755636 It's best to create a thread so that the staff can keep track rather than getting your issues buried in this channel by other people's messages.

You might get better explaination from the staff if you properly create the thread.

Also, the staff do read the threads, so even if you don’t get a reply, assume they’ve seen it. 🙂

[mental mulch]

Also where do I switch from 2.0 to 2.1

Is that Fast mode or?

[crisp basalt]

yeaj

2.1 is fast mode

[mental mulch]

Fast Mode doesn't work with Img2Video tho

Itll switch back

[crisp basalt]

yea' its disabeld for img2video only

im gonna assume you have exhausted your quota for normal mode (2.0) for img2video

[mental mulch]

For the month? 😭

So im just screwed for a month??

[crisp basalt]

possibily

but dont quote me on that, wait until a staff clarify

U got whisk animate?

does your whisk animate still able to generate animate video ? (i dont think whisk animate credit is linked to videofx credit )

[mental mulch]

I can assume so

And yes I do

Of course -_- (Whisk Animate)

[crisp basalt]

Yeah, Whisk Animate credits are separate from VideoFX credits.

oh you finished your credfit

[mental mulch]

This movie creation is gonna take longer than I thought

[crisp basalt]

Whisk Animate gives you 100 credits per month, with each video costing 1 credit.

[mental mulch]

lol

[crisp basalt]

let me dm you

[mental mulch]

I DMed u already

and added u

[crisp basalt]

im gonna assume he was trying to get the model to generate real-time speed videos, which probably led to him cranking out a bunch of videos and eventually burning through his i2v model quota

[hexed marsh]

100 gens a month is good for 3 gens a day

now i'm wishing i didn't do the math lol

[magic flower]

What's the best, accessible ways to access veo2 image to video?

[crisp basalt]

through api or videofx?

[tacit aspen]

The last update refreshed my limits, so just not doing as much this go round so it lasts.

[crisp basalt]

Oh by the way, VideoFx has a new early access registration form

[magic flower]

Right, but I don't have early access..

[crisp basalt]

Well, you can register , they do rollout on

gibe it a shot tho, I've seen a few people here get access recently. They review all submissions on a rolling basis.

[copper hawk]

Only people in America can do this right?

via USA (you probably already know what i mean)

[copper hawk]

Oh yeah im slow, thats probably why I never got accepted when it first rolled out last year in May

[crisp basalt]

This is the new form tho

Also, they did discord early access rollout like months ago

which is how most of us got in

[copper hawk]

Oh damn

[copper hawk]

How did I never notice this

[crisp basalt]

well did you notice the whisk animate announcement?

[copper hawk]

oh yeah this is different

[copper hawk]

Ye I got accepted into that, im just using that rn

[crisp basalt]

ahh ok

Yeah, you’ll need to keep an eye on the announcement channel from now on 👀

[copper hawk]

Yeah

[atomic bough]

the videofx created videos that people make are... interesting

[ashen yoke]

Just a general note of appreciation for VideoFX. It's been helping me visualize some of my learner's digital storytelling projects, helping me visualize some of the ideas I'm trying to put together, and generally making me happy when there's a whole lot to be sad about these days. So thanks, Devs. Your work is appreciated.

[crisp basalt]

Google is cookin somethin big for us behind the scene im sure 👀

[wet ice]

what makes you think this? 😎

[snow remnant]

How Do I get access to videofx??? I've been waiting forever 😖

[flint dagger]

Hola cómo puedo hacer para que una foto cualquiera la transforme en cualquier estilo artístico que está de moda?

[snow remnant]

Gotcha hopefully beta will be over with soon

[snow remnant]

😂 well I'd rather get in for free but there's still freepix for now

[stark warren]

Does anyone know what the video switched style half way through? Any thoughts and tips are appreciated!

[crisp basalt]

Can you clarify your question? is it the transition to the "daytime" of the video "?

[mental mulch]

This is seriously a forest fire, i try to gen 4 image2video videos and it stops at 5-7% with the error

[crisp basalt]

I want to assume you were asking “why the video switched style” uhhh if the user prompt contain something like “style change” or something then it will give that result I guess

[mental mulch]

My PC has nothing to do with it

[mental mulch]

Just think obviously VideoFX is still in Early Access

[winged plinth]

Did Veo Fast disappear? I’m not seeing it anywhere now.

[fair gust]

mine won't work at all

anyone else having an issue?

[viscid fjord]

fast shows up normally here

[fair gust]

mine is gone too

[obsidian summit]

Yes - saying "oops something went wrong" both t2v and i2v

[fair gust]

yup mine too, nothing works right now

[winged plinth]

Might be some updates happening.

[main saddle]

hang tight y’all! I’ve flagged this to our teams 🙏

[obsidian summit]

Thank you.

[winged plinth]

Fixed, thanks/

[crisp basalt]

Wait, we can use “fast mode” for img2video now?

[small lake]

is there a portrait mode on VideoFX?

[viscid fjord]

nope

[viscid fjord]

tried it, it said switch to a compatible model

[crisp basalt]

Oh

[shell drift]

Hello, if I need to take this to another channel, please let me know. But I signed up for the waitlist in December and I haven't heard anything. Is this common to be this long before hearing anything?

Okay I figured but also wanted to make sure I wasn't a one off case

[rain hatch]

Hi mods,
I’ve been on the waitlist for videofx since December.
Sometimes I imagine a version of me who got access. he’s having fun.
I’d like to meet him. Could you help?
thanks

[maiden depot]

something I would love to see is if the video I generated was with Fast Mode or regular Veo 2 mode on the flip card...

[rain cipher]

Does anyone have an idea when the early access message from Google will arrive? I really want to try it. Google, please message me.👀 😩

[wet kiln]

You can give me a prompt and I can generate it for you if you like

10,000 people? Dang Google deserves a break

ONE HUNDRED THOUSAND?!!

[acoustic hill]

Me too (┬＿┬)

[upbeat rivet]

@rain cipher Only if you want to... You can send me the prompt you want to try, and I will immediately show you what it makes. 👍

[crisp basalt]

Ah new update!

Cropping yess!!

@edgy holly also img2video can now be use with Veo Fast Mode 👀

@paper inlet i think youre gonna like the new update 👁️

Img2Video Fast Mode

img2video fast mode is actually giving less jumpcut to random scene now! This is very nice!

Maybe the image you uploaded is already cropped to fit/16:9 ?

[crisp basalt]

Actually one of my image i input was exactly 16:9 and it still shows the crop tool,

[crisp basalt]

ah finally a proper error message that shows if the prompt/image couldnt be generated due to filter

[hardy halo]

You guys are so on top of it! Just pushed an update so there's cropping, better error messages, and I2V works with fast mode now! #flow-announcements・📣 message

[stark warren]

I uploaded the image and prompted the motion with no expections of scene change. All other generations had this issue as well. I was wondering if anyone had any idea on why this was happening? Or if anyone can provide tips on how to mitagate this behavior.

[hardy halo]

Can you try the same image with fast mode now? it may have different quality

[stark warren]

Will do

[stark warren]

Here's what I got:

"Skier swaying from side to side, heavily breathing causing the shoulders to raise"

[hardy halo]

hmm i think the issue here is that the image is very different from your prompt. I would add more detail to your text prompt around the camera movement (zoom out to see the skier on top of a hill) or try a different image that makes it more clear he's a skier

[stark warren]

Thank you, I just tried a couple prompt varations and I'm recieving results close to what I want.

[mental mulch]

Still wont work for me

I2V just straight out hates me

@hardy halo Can I get some serious help, please?

I've been waiting 4 days on this issue of instant errors when I generate with a image, I tried every image that I can think of and it refuses to generate.

[iron matrix]

Were working on it I promise Michael! 😢

[visual sapphire]

Google doesn’t like people with the name Michael.

[wooden helm]

Why

[rain cipher]

Thank you very much, but is Google still accepting people for this beta version?

[wooden helm]

Cool never heard of this before.

[nova ferry]

any advice to look back in Library for thousands of videos? Search doesn't seem to help (it looks like missing some) and scrolling down is way too slow (for the bast 5 hours I'm slowly scrolling and now I'm only a week behind).

[narrow dove]

How nice it would be if you could make videos with sound at once. 👀✨

The videos would be even more lively! 🌌

[unique abyss]

Or sound to video with prompt, like the (uploaded) image to video

[knotty flax]

I got access today

[crisp basalt]

oh like just now?

i want to prove to these people who think google labs didnt read your form etc

[knotty flax]

I got it an hour ago. Applied a week or two after wait-list came out I think. Definitely the longest wait between this and other labs waitlists but that's kind of understandable. I already did a ton of testing through API when that became available

[crisp basalt]

ah i see, well come on in and join the fun @knotty flax ✨

youre in on the right time after they did many update to the videofx tools

[knotty flax]

I noticed for a brief moment yesterday that on freepik, the API tool I was using that choosing start AND end image was available for Veo 2. That could be coming feature.

[crisp basalt]

wait what

start and end is availalble in api?

ah no its still start image only

but, expect start frame and end frame coming to the videofx 👀

[knotty flax]

I noticed it yesterday specifically available for Veo 2 on freepik on my way out the door and figured I'd test later but it was gone. Maybe they need to do testing. They removed start image for a bit when Veo first came out there

[crisp basalt]

img2video was briefly available for api but google disabled it after few weeks, then later reenable again

[mental mulch]

I got a very rare error

429 Too Many Requests

Hopefully this gets fixed soon

[knotty flax]

Yeah from the Google form thing

[analog sedge]

We need more aspect ratios

Atleast 9:16 one

everything is on reels nowdays

Also who is the genius on google behind the new veofx layout?

10-20 videos running in the background making my PC slow and I have rtx 4090

[edgy pasture]

NO WAY

they actually released veo2 on Gemini ??

For Gemini advanced users

[unique abyss]

There are some reports that some users are seeing this. Of course, rollouts usually happen in bits and pieces. I have Gemini Advanced, but am not yet seeing this, on the app or on the web.

[wet kiln]

I have no idea

[crisp basalt]

codename robin was correct after all

@edgy holly

[crisp basalt]

this

[edgy pasture]

That’s cool, ig I need to resubscribe to it

[slim sorrel]

Sooo I got into VideoFX today. I keep getting this error, sorry if this has been asked before.. but am I doing something wrong?

[molten seal]

I htink the end frame option was showing maybe because you had another engine selected like Kling, before choosing Veo2.

[mental mulch]

the errors have resolved but Now im having trouble generating images it keeps saying im generating too quickly with t2i2v when i waited like 5-10 mins already

[paper inlet]

Is anyone else missing videos in their VideoFX library? I'm missing videos from 3/24 - 3/30.

[winged plinth]

You're not alone, for me I keep getting crashes after scrolling through a few videos, and I’m never able to view older ones in the library. On the generation page, the “Load More” button only loads a couple of pages, then stops completely.

[paper inlet]

I like the new minigrid that @crisp basalt pointed out. I prefer the scroll-loading vs. having to keep clicking the "Load More" button. But there are videos missing, in either library link, this one with the minigrid and "Load More" https://labs.google/fx/aitk-web/videofx-tt/tools/video-fx OR this other one that loads by scrolling https://labs.google/fx/aitk-web/videofx-tt/library?type=videofx

[winged plinth]

Yeah, the library’s still a work in progress, so it’s mostly basic functionality for now. My best practice is to download the gens every 8 to 10 to avoid issues.

[crisp basalt]

Also, is there a way to remove failed or incomplete generated videos? A few of them, including this one, are still showing on the main page even though they no longer appear in the library

[mental mulch]

Speaking of, also the error messages on the top right should at least have a clear all button

[crisp basalt]

@odd citrus

[torpid cedar]

Like I mentioned celebrities aren't gonna work always

[edgy pasture]

Do they allow it?

[crisp basalt]

nope

[edgy pasture]

I’ve never tried

[torpid cedar]

At times I might get one with Gen-4

[crisp basalt]

anything involving public figure seems to not work

[torpid cedar]

I did get this cool Bart simpson one though

[crisp basalt]

wait on veo or

[torpid cedar]

Veo

[crisp basalt]

WAIT is spongebob still working on veo

i do know someone actually posted al ot of spongebob from veo

i forgot who

[crisp basalt]

OOH THATS DOPE THO

wait did u prompt it bart simpson in the input prompt

[torpid cedar]

No it's image

[edgy pasture]

Was it I2V or just text?

[torpid cedar]

I2v

[crisp basalt]

the cropped part at the beginning , the model does know what subject it is to be able to generate a zoom out like that

[torpid cedar]

Yeah the intelligence

[crisp basalt]

through api or videofx?

[torpid cedar]

Fx

[crisp basalt]

oh thats interesting,

[torpid cedar]

Since me working with Runway. I've gotten better with certain image to video options

[crisp basalt]

also i kinda understood how Image2Video in videofx works

The user uploads an image → The system uses a multimodal model to analyze and understand the contents of the image → It generates a descriptive prompt based on what it sees → Then it reads the user's text input → Merges the image-based prompt with the user's prompt → Creates a final combined prompt → And finally, generates the video output.

[torpid cedar]

I heard veo 2 is coming on Gemini as well

[crisp basalt]

yeah theyre working on it for a while now

[torpid cedar]

Didn't release for mine yet

[crisp basalt]

its a slow roll out i think

[torpid cedar]

Yes

[odd citrus]

It already released to some people? I remember they've catched it looking at the logs that they will implement it but it wasnt out to anyone

[torpid cedar]

Yeah it's out already

In waves

[crisp basalt]

well we are getting close to "public" release , maybe later this year

[odd citrus]

If its already out in gemini Id say its one month out

[crisp basalt]

oh yeah the gemini one , u can get veo access through google one ai plan

[torpid cedar]

Yeah not an official global release

[crisp basalt]

so if anyone havent redeem the "try gemini advanced" save those up for when they release veo in gemini 👀

[torpid cedar]

I've had Gemini since day one

Lol paid for everything

Considering I'm an android person

[dim glen]

Wrigley how accurate is it with people's Images

[crisp basalt]

well it can definitely detect if the image contains IP, or public figure , celebrity

hence it will reject the generation right after user submit

[torpid cedar]

Midjourney has it's own uniqueness just like imagen 3 does. But if we are truly honest with ourselves. Native gen is the top tier of image generation thus far.

[dim glen]

Oh wow but I'm just curious. Does it like alter the original look of the uploaded image? Like if I upload my own picture, will it maintain the original look?

[crisp basalt]

Ever since the ChatGPT image tool dropped, people’s expectations are totally changing

I hope google cookin something good for imagen

[dim glen]

If you don't mind, can I send you a picture privately to help run some prompts? That's if you don't mind. (I don't wanna share it here)

[crisp basalt]

it actually keep it consistent

also yes you can dm me

[dim glen]

Thanks.

[torpid cedar]

Google already has this just not fully implemented

[crisp basalt]

definitely

[torpid cedar]

Probably will upgrade some stuff too

Veo is still top video model then it's Runway Gen-4

Don't sleep on them either

[dim glen]

Messaged you

[torpid cedar]

A future idea for you guys at Google. Being able to upload an image while it describes the prompt to generate anew. Would be cool for a video 2 video output as well

[paper inlet]

My videos of the mural, the blue volvo, and the teal VW rabbit, etc. are all gone. Everything from 3/24 - 3/30 is missing from my library. cc:ing @iron matrix

[crisp basalt]

actually yeah i wanted to mention that too, i think i realized some , not many , but few of the vids were gone from my library

[paper inlet]

Also, videos that I could see yesterday that I created on 3/31 (the anthropomorphic otter in the tropical market) are now all gone.😱

[crisp basalt]

wait

youre right , mine shiba one, is gone too

[fierce jackal]

wheres the best plave to add sound effects and audio?

[rocky matrix]

does anyone know what the TT Slow is?

[fierce jackal]

Nope haven’t seen that

[rigid venture]

trophy thomas mode

[near gate]

Question is Veo 2 coming out on Gemini on Monday???

[iron matrix]

are you on the correct url? https://labs.google/fx/aitk-web/videofx-tt/tools/video-fx

Also when you say gone, what do you mean? Can you see boxes, or are the actual generations non existent?

[paper inlet]

The videos are missing from either link. 1. From the prompt input page: https://labs.google/fx/aitk-web/videofx-tt/tools/video-fx
2. From the library button: https://labs.google/fx/aitk-web/videofx-tt/library?type=videofx

For #1, there are two empty boxes, which I think are failed gens and don't account for the number of missing videos, as there were many.

For #2, there are many empty boxes, but I still don't think it accounts for the number of missing videos.

[near gate]

did you try it?

[near gate]

I just got Gemini advanced so excited 😆 😍 I still don’t know when Veo 2 is coming out on Gemini

[silk hill]

has anyone figured out how to make image to video NOT cut to random different shots after less than a second of the videos it generates?

[wooden helm]

Will there be new registrations open for video FX in future?

[rain cipher]

Did Google stop accepting people on the waiting list?

In video fx or whisk

[crisp basalt]

👁️

Wait is this the same person that I knew

[nova ferry]

same here. did you get it solved?

[dim glen]

Could someone please help me run i2v? Pic and prompt #flow-tips-n-tricks・✏️

[paper inlet]

https://discord.com/channels/1172568727942860810/1358113875555516513

[iron matrix]

yes there will be at some point!

Sorry for the issues with the library! It will be one of the first things we take a look at on monday 👍 thank you for letting us know!

[crisp basalt]

https://www.youtube.com/watch?v=PEUcCF-jVCI

breath taking 👏

[glossy path]

gm gm

took an image of me and a buddy

got us dancing

[teal oak]

something funky going on with videofx today?
Keep getting this error: There doesn't seem to be anything here. Try a different prompt and check out our content policies here."

[dim glen]

I'm still working on the short movie I'm creating about an imaginary country. With ChatGPT, native image generation and editing, Gemini, and now VO2 for image-to-video, I think my work is going to be much simpler. So, could someone please help me animate this picture? Picture and prompt #flow-tips-n-tricks・✏️

[whole veldt]

i signed up for videofx, how long did it take yall to get accepted?

[crisp basalt]

Mine was .. a day

That because the registration for discord community was open for one day , and most of us got in!

[whole veldt]

ahhhhhhh

[whole veldt]

i didnt know there was a discord until today but i reregistered like a month ago

[crisp basalt]

It was this

[whole veldt]

sad days

is there any way at all to still get in?

[crisp basalt]

But from the looks of it I think Veo 2 is getting closer to public release

[crisp basalt]

Well , nothing you can do but to wait for now sadly

[whole veldt]

ok, thank you for you helpp

[crisp basalt]

Recently Gemini got this

[whole veldt]

are you able to make videos from it?

[crisp basalt]

So maybe lookout for any update from Gemini

[crisp basalt]

Apparently yes you can

aistudio website will also have veo-internal model ready so stay tune for that too

[whole veldt]

is this in google cloud? sorry i havnt used much of the gemini models

[crisp basalt]

Oh this is Gemini website

It’s not rolling out to everyone yet,it’s a slow rollout

[old mesa]

I've noticed that i2v fast mode makes cars go backwards - like, all of the time.

normal mode doesn't have this problem.

[odd citrus]

interesting, nice catch. will try

[sick hawk]

Thanks dude

Been waiting to test that, but it wasn't live again to test.

AM gonna teat that there

Now.

[old mesa]

Some examples

[crisp basalt]

Ahhhh i love how veo maintains those classic old video feeling

most of the ai video generation model that ive tried would straight up "enhance" the quality , erasing those old video feels aesthetic

[old mesa]

Yeah! I generated it in v7 from Midjourney. However, in i2v (not fast mode) it likes to change the quality ughhh

[crisp basalt]

oh god the 2nd one tripping me out

[crisp basalt]

one trick I’ve used is flipping the script with a bit of reverse psychology

Instead of saying “people running away from [subject],” I’d write “people running towards the subject.” And surprisingly, it works.

[keen compass]

@crisp basalt it's alright

[near gate]

Wait does that mean the full veo model would be available for free 🤯

[dark osprey]

Why do I not have it?

[near gate]

its legit though

[warm flint]

It's been a while since I've seen weird minecraft gens

[crisp basalt]

We can expect this during the upcoming Cloud Next event. 👀

[fierce jackal]

Trying to use my phone to make things and I’ve noticed videos never load, is this a common issues for people?

[ornate kelp]

I keep one tab open on the video generation tab, and another on the library tab and just refresh the library tab when the video is done and it loads there.

[fierce jackal]

https://tenor.com/view/garthwaynes-praising-hallelujah-world-not-worthy-gif-8958089796881687621

[fierce jackal]

Genious

[rotund yarrow]

Am I missing something? Video gen is waitlisted?

[edgy pasture]

oh amazing, had no idea they were doing an event

[crisp basalt]

I think the clod event around mid April?

Clod? Cloud*

[mental mulch]

i actually reached my gen limit lol

now gotta wait til next month lol

[crisp basalt]

@edgy pasture heads up ,

[silk hill]

where's this from?

[crisp basalt]

Aistudio website changelog, it was shown this briefly before they removed it all of sudden

[edgy pasture]

Hoping there is a cooldown instead of monthly limit like Sora 🔥

Maybe they changed it

[hushed rain]

News is starting to drop:
https://techcrunch.com/2025/04/09/google-brings-a-music-generating-ai-model-to-its-enterprise-cloud/
Veo 2 video creation model has been enhanced with new editing and visual effects customization options. The company has also launched a voice-cloning feature powered by Chirp 3, Google’s audio understanding model, for “allow-listed” users. And the Imagen 3 image generator now delivers what the company describes as “significantly” better performance.

WHOA:
As for Veo 2, the model can now remove background images, logos, and objects from existing videos, and extend the frame of video footage (to convert landscape video into portrait, for example). It can also now adjust the camera angles and pacing in AI-generated scenes to create timelapses, drone-style clips, and more, and it can interpolate between specified beginning and end frames.

[crisp basalt]

CRAZY

[hushed rain]

The event is in a few hours...can't WAIT.

[crisp basalt]

AH I’m so excited

[crisp basalt]

@edgy holly Google actually cooked BIG ONE for us

[lusty nymph]

is it live for everyone to watch?

[torpid cedar]

So what about the 4k resolution we were supposed to get?

[edgy pasture]

AMAZING 🔥

[edgy pasture]

12 noon EST, so 2 hours and a half from now

[torpid cedar]

I have the Gemini astra feature already

Because I have the latest A.I Samsung Galaxy as well

[lusty nymph]

Thanks a lot!

[torpid cedar]

I may sound bias...but I love Google stuff. Android all day.

[crisp basalt]

Even tho I moved on from being pixel fanboy (I’m iOS user now) I still love Google stuff as well

[torpid cedar]

That's cool. I never looked back having android. I used an iPhone before and didn't like the applications setup.

[molten seal]

Is there a live link somewhere ?

[lusty nymph]

looking for it too

[edgy pasture]

https://cloud.withgoogle.com/next/25

[lusty nymph]

Thank you!

[edgy pasture]

Welcome!

There’s this one too https://www.youtube.com/live/Md4Fs-Zc3tg?si=cRL_U9od3rvOz910

[maiden depot]

4k 👀

video extend 👀

[wet ice]

thoughts?

[edgy pasture]

Are the changes only on Vertex or will they come to videofx and Gemini too?

[proven anchor]

I see it in ai studio now

[rocky matrix]

you do? where? I can't see it

[proven anchor]

https://aistudio.google.com/generate-video

Free too for now idk limits

[wet ice]

8 s max? 720p max?

[proven anchor]

yes

[wet ice]

well that's disappointing :))

[proven anchor]

says coming soon for frame rate and resolution

[fair gust]

Hey im not able to generate a photo of a shirtless man working out? does this break rules?

I mean image to video

@hardy halo

[edgy pasture]

I dont have it, did you sign up for it or smth?

Thank you

[proven anchor]

no never signed up for anything

[edgy pasture]

Did get an API key or put your credit card there?

Maybe its rolling out

[proven anchor]

nope

[crisp basalt]

Oops

OKAAY my first aistudio Gemini video generated

[covert urchin]

404 page for me

[crisp basalt]

https://aistudio.google.com/u/0/generate-video

I think not everyone get it 🥲

The negative prompt section is really needed in videofx !

so far im not sure if it actually work but let me try something

the negative prompt does work,

https://cdn.discordapp.com/attachments/1319424140028940358/1359612776305135716/Generated_File_April_10_2025_-_3_34AM.mov?ex=67f81d92&is=67f6cc12&hm=3b19c9c6705cd727c830709fb9161dfb74c257491407d9aa2e9a7fb02110bd66& cuz earlier it gave me this , when i didnt put anything in negative prompt

either Poe forgot to turn on something

[near gate]

I am trying to use veo 2 on my google ai studio account with gemini advanced how do I get to use veo 2 on google ai studio and Gemini as well?

???

[molten seal]

Forgot to tell you thank you.

[nimble glen]

No veo 2 there. Use it on Vertex AI

[edgy pasture]

You’re welcome!

[near gate]

😭

[edgy pasture]

What about MattVidPro AI on YouTube?

Or this right here?

[hushed rain]

Nah-- Matt was working off the same info as all of us. I borked that this morning to. Admittedly, the way the model notes were written, it SEEMED like we were getting it in Gemini.

[mental mulch]

I hit my generation limit last night do I gotta wait til May 8th now?

[brisk panther]

Yeah unfortunately

[brisk panther]

Damn. Veo 2 via Google studio only gives you a few gens then stops working? That sucks lmao

[brisk panther]

Yeah, it's just that I've been trying to get access to this since it was first announced. Everyone around me got beta access from the site, and to be honest, I'm really the only one among them that haven't yet. I'm sure there are alot of other people in the same situation, because there's only so many you can accept before it becomes a problem considering it's a free test. But yeah, makes sense. Atleast I was able to generate 2 videos

2 is better than none 😂
Hopefully there'll be a way soon that I can generate more without needing to pay a fortune

[fierce jackal]

I mean it was obvious that it wouldn’t be completely free once released unfortunately, being its video gen, cost a lot more

[brisk panther]

Oh yeah definitely. That's why I was hoping to get into the beta access while it was free, but nothing worked out 😂

Yup. Surprised they had the free beta open for so long considering the amount of people that were in it. Must of cost alot. I was gonna get a plan on an external app to get access, but the prices were crazy 😂 I think like 20$ for maybe 5-10 videos? 5 seconds each

[fierce jackal]

That bites, when are they gonna stop the beta access anyway?

[brisk panther]

Have no idea. I was hoping to get into it even for a little before it ended, but it seems they are already planning the pricing and access if I'm not mistaken

[fierce jackal]

Ah

I better use what I can of it the

Then*

[brisk panther]

Yup. Any idea exactly how much the ai studio mode gives you? It seems I'm able to generate one video every 30esh minutes. Wish they said exactly what the rate limits and quota was

[edgy pasture]

The free tier is based on the paid tier’s usage, so they could never give a rate limit or quota. As for the paid tier, there are 3 types and tier 1 (only add you credit card, no chargers needed) gives you 50 RPD

[brisk panther]

Oh yeah that makes sense. I forgot Google ai studio had an option to put in money like that. Any clue on how much money per video?

[edgy pasture]

No idea

[edgy pasture]

35 cents per second USD on the Gemini API and 50 cents per second on Vertex AI USD

[brisk panther]

Damn yeah that is quite a bit understandably 😂 and the api is what the gemini ai studio uses?

[edgy pasture]

Yes, pretty expensive

I wanted to see the price for Veo2-fast, no idea what it would be

[brisk panther]

Probably cheaper. But for my use cases, I wouldn't even bother to try it. Does whisk animate use veo 2 fast?

Oh yeah. That's what I got access to. Whisk animate. The veo 2 on that is really different

[crisp basalt]

It’s still the same Veo 2 model, just fine-tuned to animate an images more smoothly without jump cuts or glitches.

[cobalt junco]

o I still get some funny glitches 😉

[brisk panther]

So "Veo2-animation" almost. It lacks alot of movement I find, but I guess that's it's purpose to make cool animations

[cobalt junco]

I can get some pretty cool stuff out of whisk

[crisp basalt]

If I recall correctly , whisk animate uses veo 2 distilled model or something I forgot

[brisk panther]

So almost like the fast model? Not entirely sure what the distilled one is lol

[cobalt junco]

it's nice and quick and easy for a mockup to tune in VideoFX hehe

[crisp basalt]

It seems like a lot of people are misinterpreting what Whisk Animate is actually meant for. (Don’t worry I’m not judging 😄) While it does use the Veo2 model, that doesn’t mean it’s designed to output the same kind of cinematic results you’d expect from something like VideoFX or other AI text-2-video/image-2-video models.

Whisk Animate leans more toward casual, fun use cases.

For example, turning your design into a simple turnaround or giving life to a static blueprint. It’s more of a visualization tool than a full production engine.

[brisk panther]

The new models on suno use to end up becoming free for everyone, but it's been months and I've only gotten the 10 generations they give when you download the app 😂

Yeah, I wasn't really asking about it here necessarily, our conversation just headed in the direction for me to say it

What apps have it that much?

[brisk panther]

I'm aware.

[brisk panther]

That's a cute video haha

[blazing plover]

Thanks 🙂 My favorite part is that it made the bumblebees wobbly and clumsy, just like in real life 🙂

[brisk panther]

Haha yup 😂 was that whisk animate or regular veo 2?

[blazing plover]

Veo 2

[blazing plover]

There's another one in the Whisk gallery I made as well Lol

[analog sedge]

How do we get access to the new veo2

[nimble glen]

easy. go to google ai studio.

[vapid stratus]

DO we already have access to the in paintying in Veo 2?

[crisp basalt]

nope

[vapid stratus]

Mone doesn't look like that

[nimble glen]

its not rolled out to everyone on google ai studio.

[crisp basalt]

Probably best to shift this convo about AIStudio over to#off-topic・🐢 or #general・🧺 since this channel’s meant specifically for the Google Labs VideoFX app.

actually no we do have a thread https://ptb.discord.com/channels/1172568727942860810/1349795590988365846

[torpid cedar]

Google definitely cooked based on the current LIVE STREAM...

[nimble glen]

Peoples asked where to get veo i just responded

[rocky matrix]

what live stream?

[torpid cedar]

Google live stream yesterday

[crisp basalt]

i was gonna ask that , theres a live stream ongoing?

ohhh

[torpid cedar]

They have one up to April 11th I think

[crisp basalt]

but yeah they cooked so much for us, and the labs team are actually cooking something big for us too as of right now 👀

[torpid cedar]

I need that vertex layout in labs

[crisp basalt]

the one from aistudio is it?

[torpid cedar]

Yeah vertex has all in one

[crisp basalt]

its a really clean ui ngl

[torpid cedar]

Yeah labs needs that but their img2video feature still buggy

Plus I don't like the crop adjustment as much. They just need ratios added to use for whatever output is needed

I'm loving Project Astra on my Gemini app too. It can see and remember anything

[crisp basalt]

i have a huge need for that one ugh

[cyan flame]

All of my video generations are gone from my library. Is this intentional?

[torpid cedar]

A bug

[charred dock]

Hello @everyone! In the future, with Veo 2, we might get a feature called "V2V" (video-to-video).

[torpid cedar]

It's not far-fetched

[wet kiln]

I haven't created a visual story in a whiiile

[viscid fjord]

Anyone else unable to select the normal Veo 2?

[crisp basalt]

same

its stuck here

[rocky matrix]

Same for me

[outer aspen]

Just got Veo 2 through AI Studio! Not special anymore haha

There's negative prompting as well as generation length selection of 5-8 seconds, and even a 9:16 aspect ratio option

Frame rate and resolution cannot be changed yet but they're shown as coming soon

[crisp basalt]

Just a little heads-up, it’d be awesome if you could share your output in #1349795590988365846 . The VideoFx channel’s mainly for the Google Labs VideoFX tool. 😄

[outer aspen]

Makes sense! Haha, interesting how it's the same model but through different systems. Hopefully we can get the same features through VideoFX soon

[outer aspen]

lol

At this point surely AI Studio warrants more than a thread

[crisp basalt]

I mean, a Discord server just for AI Studio would probably do the trick, but no worries, that was just a little reminder 😊

[outer aspen]

There's one for Gemini in general

[raw swan]

hey everyone, for those who got access to veo 2 on google ai studio, what is the rate limit and every how much time it gets reset?

[mental mulch]

How do I use Vertex AI for Veo without having to pay

[crisp basalt]

You have access to videofx already tho,

[mental mulch]

I hit my limit 😂

[crisp basalt]

Well veo on vertex ai also have much higher limit than videofx

[mental mulch]

Ok and does it cost money

[crisp basalt]

Soon to be yes

[mental mulch]

Screw that lol

Guess ill wait til next month lol

[mental mulch]

A month is seriously too long :/

[mighty bramble]

Are people still getting early access to videofx? I submitted the waitlist months ago now but still no dice