flow-general・🌊

[wicked crater]

Now I'm getting this:

[unique citrus]

Help! I can't extend the video anymore, there's no way to set the prompt to extend it, I've already clicked everywhere.

[topaz spade]

I feel like I saw a thing to add an @me with characters, anyone know how to do that?

[fading dirge]

for example, left is veo 3.1 quality, one it does not look like a cellphone video like i prompted and thats not the mercedes benz stadium in atlanta georgia. the one on the right omni, does look like a cellphone video and thats very close to the mercedes benz stadium in atlanta georgia. so omni is an upgrade

[somber pebble]

did you feed it ingredients or did you just straight up text prompt?

[fading dirge]

also understands that people would be walking around with the atlanta falcons jerseys on at the mercedes benz stadium

[fading dirge]

text to video

[fiery panther]

No more unlimited veo? :c or whats goin on with the new plans of ultra

[heavy turret]

you can switch to 2 insted of pro and that seems to work

also we're having an issue (well many) with Flow

like that quota error

[somber pebble]

the issue i'm having is that it seems to really suck with ingredient so far, especially my characters which is kind of the most important thing

like ingredient adherance is a huge step back in my experience

[crisp basalt]

it works fine for me

[fading dirge]

i did not prompt anything about the atlanta falcons, so its has better world understanding than regular veo 3.1

[wicked crater]

Only Nano Banana 2 is affected

[unique citrus]

Is the option to extend the video now only available through a paid service?

[slow tendon]

alr so i made an app but i wanna one thing about like how i can make it better with nano banana can someone help me

[tacit aspen]

This is amazing and overwhelming and amazing ... I'd love for Google to give Ultra a week to get comfortable with all the new tools. I'm loving it, just re-thinking the workflow.

[somber pebble]

both characters in these scene i tried dont look anything like the character ingredients used. It ignored my style keywords in my prompt and location ingredient. Put in a completely different vehicle than my ingredient, etc..

[crisp basalt]

we cant really help much if u dont show us what u prompted etc

[somber pebble]

so far it's doing a lot worse than Veo 3.1 for this stuff

[spring kite]

@main saddle Will you also find a solution to the “unusual” error that keeps popping up?
I haven't been able to get anything done properly for the past 10 days.

[somber pebble]

I'm not looking for help lol, just saying that results I've been getting with veo3 for the last month + have been pretty consistent, but using the same prompts and ingredients in Omni have not created the expected results

[unique citrus]

What a terrible update, everything got confusing.

[crisp basalt]

well sorry to hear that its not working for u

mine has been working fine ,

[somber pebble]

Maybe it sucks for animated stuff currently

[fiery panther]

i only wanna know if there isnt more lower priority veo3 :c

[crisp basalt]

it really depends on how you prompted it,

[somber pebble]

like my locations and vehicles no longer look cg, they look like real life video

[crisp basalt]

anyway!

[somber pebble]

like it "should" be looking like this

[novel plank]

You might have to say "Keep the animated style of this image". I have some of the same problems with animation with Gemini.

I am hoping the new changes with the UI for characters works in the future. Might help.

[crisp basalt]

have u tried creating "character" for these

[somber pebble]

I include the same lighting, style and camera keywords in the generations, but I'll try just telling it to match the style as well next one

[crisp basalt]

ill borrow this to create character ref sheet if u dont mind

[winged plinth]

The exact model version active in this chat is Gemini 3.5 Flash.💯

[somber pebble]

Yeah, I made them this morning using the ingredients i've been using for my production this far.

I've essentially been doing the character thing manually per generation, so I was excited for this

it seemed to have made the characters fine from the references

but putting them in Omni and it's like "nah imma do my own thing"

but it had been working generally fine in Nano Banana and Veo as I referenced his character art each time for each scene he's in, etc..

havent tried seeing if the @ thing works outside of Omni yet

[cyan dragon]

Thank you for the Agent its a big help ❤️

[wide ether]

what the heck happen to the repeat behavior of the video playback? Why take that away?!

[novel plank]

From within Flow, Nano Banana will often revert back to real life characters even when all the references are anime. Sometimes it's a bit weird. One day it works fine and then next it'll just want to do whatever style it wants to. lol

[somber pebble]

dont get me wrong, I want Omni to be good, just results for the project I've been working on have been worse with it so far than they are with Veo

[opaque sand]

OMNI costs 30 credits for a 10-second video, even on ULTRA plans? That doesn't seem quite right...

[wicked crater]

Now it shows the proper error.

[pulsar coral]

Bout time, I can toss the error message template about that now

[main saddle]

some amazing Flow updates, including context on Omni, happening now in the live Keynote! join us in #1182419495369855056

[burnt pilot]

Is flow fixed yet?

[pulsar coral]

I think enough people posted how's Omni atm...

...probably most the issues come from Gemini's understanding of things, considering more likely it uses a search function to help the image and video models to trace images, the exaggerated recent filtering is short circuiting it.

Seedance is also highly censored, but it doesn't block almost every word like Gemini is doing, the censoring is more conscious.

[main saddle]

This should now be resolved. Thanks for your patience!

[wicked crater]

I think I run out of generations for Nano Banana 2 today.

[plucky scarab]

I didnt get it how can i add voice to my character

[opaque sand]

3.1 vs OMNI
The doll leaps menacingly out of the darkness.....

[wide ether]

how can I get the video player to repeat the videos? ... that appears to have gone away

[merry mango]

The Storyboard Studio app lets you create characters from scratch, which is great, and Flow lets you create characters really well too: but why don't the two apps work together? We have an “Autofill Characters” feature, but we can't use a character we've already created in our collections!

[pulsar coral]

Yeah this is illustrating it, the second video seems to visibly have traced the puppet mask of that villain from the "Saw" movies, thinking it would help you to create the video, but gave it hair instead.

Right off google images

[high plume]

did they increase limits on nano banana pro cause yestarday i was able to generate more images than today

[pulsar coral]

it also changed the eyes and yeah no this is not how it should work

[opaque sand]

That’s exactly what I wanted to show—it looks like we’re dealing with something similar to the logic of Seedance 2.0

[pulsar coral]

Anyway all can probably be stabilized by prompt, but Veo never gave anyone enough free tries to get enough practice on it.

[pulsar coral]

Admitttedly the jump animation was better, but the face did morph all over the place.

it's just "search function" logic vs animation execution

[dreamy oxide]

For me it's both nano bananas that are giving the generating too quickly message. I hope this website gets fixed sooner or later

[opaque sand]

Overall, the graphics and camera are interesting. Of course, you can minimize distortion, but it’s still impossible to tinker with the “components” in OMNI (and now that it’s no longer free, that’s even more of an issue). I’ll be waiting for the “frames” mode, because for what I do, preserving the sequence of characters and objects is absolutely crucial.

[pulsar coral]

The animation wasn't just better, it was miles better. Then again yeah Veo has always needed some prompt to keep reference image faces, can't be helped. Yeah maybe one can adjust to this.

[merry mango]

It would have been better to have the apps tested by real users working on real projects—that way, you’d have fewer glitches in the apps...

[honest abyss]

finsihed?!

[wide ether]

Note for others .... the drag and drop frame image is broken if you don't first clear out the existing image that's there in the reuse prompt case .... so when you reuse a prompt but want to drag a different starting frame image, that generation just fails if you drag over the top of the other image to replace it without first clearing it

[honest abyss]

that was so short

[wide ether]

I'll ask again ... what the heck happened to the repeat playback functionality?!!!

why in the world would you take that away ... unless it's some browser config buried in the browser settings that I'm overlooking

[winged plinth]

@main saddle ❤️

[little fulcrum]

we lost editing ability on the ultra plan?

[wide ether]

.... and we lost the repeat of playback function too ... I can't believe they took that away

[remote crypt]

well, I applaud the recent spate of new customizability. But Veo is still mired in an insane filter that sees "prominent people" everywhere, and stifles creativity.

[wide ether]

now you have to hit the play button manually every time you want it played back again

[little fulcrum]

yes that's the bane of my existance right there. productivity is way down this past month. I haven't had much time to look at the updates, but it looks exciting.

[wide ether]

I know right? you can generate images without any references and it still throws the pominent people error

[wide ether]

are you fine with losing the repeat playback function?

[little fulcrum]

the first thing that jumped at me was the character creation. I'm excited to dive in.

[little fulcrum]

tbh i didn't notice until you mentioned it.

[wide ether]

I used that feature very heavily!

[remote crypt]

It's definitely two steps up, but still three steps back. I literally have in my prompt ONLY my custom character (who doesn't resemble anyone anywhere I think) reading a newspaper and guffawing at the recent scandal of females being allowed to smoke. (early 1900s)

[merry mango]

The Storyboard Studio tool could be great, but it creates basic character sheets and you can't import characters into it, so I don't see the point: it's really designed for one-off videos, not for recurring characters.

[wide ether]

now I've got to hit the play manually over and over if I want to review a view more than once

[remote crypt]

unless this guy goes under "Prominent people"? I'm sure he would like to think so 😆

[little fulcrum]

yeah that's annoying. i'm sure it was just overlooked. I also want to be able to delete videos as I review them

[wide ether]

I think that's a character from the old twilight zone show ... j/k lol

[cyan dragon]

why you remove the copy option of the images ?

[wide ether]

in an ironic twist ... the delete function came back with this update LOL!!!!!!

[little fulcrum]

did it? i'll go check again

[wide ether]

yes

it's back

[stone garden]

Awesome!

[wide ether]

I just can't believe they'd take away the repeat! that's such a bad change on their part ... what are they thinking

[plucky scarab]

This is omni btw veo 3.1 was better i think

[little fulcrum]

oooh 10 seconds

[remote crypt]

Seedance 2 is far better still, in my opinion.

[fading dirge]

its a fact not an opinion

[merry mango]

The Storyboard Studio tool is designed entirely for prompting, whereas when you use it to describe a style, a character, or a setting, the results lack consistency.

[fading dirge]

i seen one guy on twitter calling gemini omni SOTA... bruh thats crazy to say

[plucky scarab]

I didnt test it but it might be better than veo omni idk why but chinese models are always the best

[limpid ether]

Is there way to access Seedance 2?

[remote crypt]

You can access it via Runway

and others. But Runway has multiples, including Veo

[fading dirge]

i can tell you right now ive been using omni for the last couple of hours, seedance 2 is far better

[remote crypt]

I made this "War of the worlds" scene, and it was way better when done in SD2.

[limpid ether]

What is the price range of the packages?

[remote crypt]

you can check on their websites. I'm not sure it's TOS to talk too much about competitors in this discord 😉

[fading dirge]

how did i magically get 51k credits in flow?

[limpid ether]

ops 😄

[crisp basalt]

same

[remote crypt]

I still have less than 5k credits

[plucky scarab]

No no its not against tos you can see every ai competitor is using the top models to copmpare

[crisp basalt]

thank u for the extra cred flow team ❤️

[merry mango]

if you look your video indetails there are many mistakes in movment and accessories

[remote crypt]

yes, but Veo and Kling and the rest made worse videos 😄

I've tested it a bit

[sick sparrow]

anyone had any success with video to video?

[fading dirge]

ive tried it, its kind of wonky

[plucky scarab]

HOWWW

[sick sparrow]

hmm, i get a lot of declines, even though it's just a shot of a tree where i want to make the sky sunny 😄

[minor badger]

do you have a channel?

[remote crypt]

no

I mean, I kind of do. But only 2 vids there.

I've mostly worked with Midjourney and made most stuff there

[minor badger]

it reminded me a bit of shockfactor_ai at first

this guy: https://www.youtube.com/@shockfactor_ai

[remote crypt]

oh yeah, I make a lot of creepy-pasta

[tacit aspen]

Thank you for the credits flow team.

[little fulcrum]

hey I didn't get any lol. this past month has been a huge headache

[fading dirge]

youre not missing out on anything though its pretty wonky and mostly unsable based off my experience

[wide ether]

holy sh*t! ..... sweet!

[fading dirge]

also extremely censored. uploaded a video and prompted it to change me saying "what do you want for dinner?" it blocked it as harmful content

[tardy fog]

If I use my own video for editing in omni I get error and both flow and gemini won't do it telling me to change prompt. I have a minigolf video which I want to change the environment to the moon. Any idea? Are our own video still not supported for editing in omni?

[heavy turret]

oh hey i cant access my flow now

not sure if it was cos i just waited or cos i deleted all google related cookies from cache just now and relogged in

🤷‍♂️

the quota error thing

[stable basin]

Flow won't let me use it unless it makes me answer it's popup.

[remote crypt]

At least with the censor filter, you will never lose your credits, because they'll be returned to you upon failing.

[tardy fog]

Just tap next without checking the box

[stone garden]

Thank you for the credits flow team indeed

[heavy turret]

is omni part of flow?

is it something you need to pay for to use? how to access it?

[minor badger]

Is there a referral program from Flow / Google people can apply to?

[stable basin]

Feels like a x button would have been enough.

[tardy fog]

Ye, agreed

[remote crypt]

I suspect it's Veo 3.1 - but they label it differently because they are employing an Agentic prompt-helper for you. I could be wrong, but I don't see any reason to think differently.

[glad isle]

“You're requesting generations too quickly. Please wait a moment and try again.” This only happens with NB2, Pro is unaffected. Same for the rest of you?

[heavy turret]

ok but how do i get to it? i cant find where it is

[heavy turret]

funny cos for me its the reverse

[fading dirge]

its has better world understanding than veo 3.1 so i think its a slight upgrade

[wicked crater]

I experienced it as well...

[remote crypt]

You have to make sure you're on the Agentic mode.

[glad isle]

Lol.

[uneven sand]

Is there an antigravity help forum anywhere... I'm panicking... I updated the Antigravity and now everything is gone...

[blissful abyss]

has the flow prominent error been fixed

[heavy turret]

how do you get that? im a free user btw

[fading dirge]

no

[little fulcrum]

awesome. how did you get their attention? I've been asking for help an entire month and my productivity has really suffered.

[remote crypt]

no never mind, I think it's just available where you find the others.

[blissful abyss]

It has

[heavy turret]

oh wait this isnt it?

[fading dirge]

i still get it

[blissful abyss]

With the new update today

[remote crypt]

it works without Agent apparently

so I think it's as Hawkdawg said, it has a broader understanding of the world

[heavy turret]

what exactly IS an agent for anyways? i dont quite get it

do i need to use it for flow?

[remote crypt]

it's like chatting to a person who helps you make a better prompt, and helps you along each process

[blissful abyss]

It just helps work while u are away

or helps you complete a project

[fading dirge]

what the, im not even speaking of real prominent peopl, im talking about it flagging none prominent people as prominent people and blocking you for it. but this is flow?

[heavy turret]

oh damn, well i dont need that then (i prefer to just chat with gemini) 🤷‍♂️

i just wanna goof about with editing and creating in flow, thats all XD

[blissful abyss]

Flow just did an update

today

[heavy turret]

when i hear "agent" i think of the godawful Grok agents

[blissful abyss]

sharp

[echo spade]

Does anyone know what the difference is in credits with the ultra versions?

[fading dirge]

i know ive been using it for the last 4 hours, using omni

[blissful abyss]

U can even link openclaw to it

[blissful abyss]

nice

are u still getting flagged

[fading dirge]

ok so you can make celebreties in flow now, nice

[foggy sleet]

So when is veo or flow is going to be better with physics understanding specially in action scenes

[blissful abyss]

yeah

[fading dirge]

i just tried and was able to make celebreties

[blissful abyss]

before i would have gotten flagged

[remote crypt]

that's like a reverse filter, great.

[blissful abyss]

yeah

[heavy turret]

yeah im kinda "simple" so i dont get all these fancy AI things and stuff 😅

too much overflow for my mind to understand.

i just prefer Flow to "edit and make art" and gemini for "talk"

[remote crypt]

I have really zero interest in making celebrity anything.

[wicked crater]

I don't see a 10 second option yet.

[blissful abyss]

I just wish more aspect ratios auto aspect ratio and setting depending on the image and more

[blissful abyss]

Kk

[heavy turret]

whats even the point of making celebrities XD seems like something random

[blissful abyss]

U can make AI influencer

unlike before u get flagged

[heavy turret]

???

[heavy turret]

unlike what? i havent liked anything

[fading dirge]

yeah but the problem was it would block none celebrity ref photos as prominent people. so im guessing since it allows celebrities that shouldnt be a problem anymore

[heavy turret]

also did the UI get a new look? theres a left sidebar now

[remote crypt]

it still blocks my character

[heavy turret]

no thanks lol

[stable basin]

Another thing I noticed is they relabeled archive "Trash", but you still have to go to a page then delete from there instead of a simple delete button, who's so afraid of a easy delete function? Why make it a process?

[fading dirge]

sheesh

[remote crypt]

I'd rather have Influenza A than be in thrall of the Influencah madness.

[heavy turret]

oh "Omni Flash" is something you pay for.

nevermind then, i can live without it 🤷‍♂️

[blissful abyss]

Even gemini image watermark has gone

[remote crypt]

also the custom voices for characters don't work yet. I just tested it

before my character got blocked for being prominent 🙂

the custom voice sounds great in the "character" area, where you create and test the voice

[heavy turret]

hey you didnt answer my question XD

[heavy turret]

what was this about?

[remote crypt]

but once you deploy it to a video generation, it's almost completely gone. I suspect the agent is necessary to use the character voices properly.

because the Agent will reiterate all your descriptions about the voice.

[heavy turret]

this sounds overly complex so i'll just not use any of this 😅

[blissful abyss]

prominent people

[heavy turret]

i mean i dont even use those so 🤷‍♂️ guess im good?

[remote crypt]

that could be a Marilyn Manson song... "The prominent people... the prominent people..."

[heavy turret]

i prefer doing art stuff in flow

[fading dirge]

yeah im still getting prominent people block when using a fake ai character

[heavy turret]

is this in the make a character thing?

[fading dirge]

havent tried that

[tacit aspen]

Omni is amazing ! They leveled up. This is a game changer for films.

[heavy turret]

so lemme try to understand, this "make a character" thing is for a character you make that you what drop into generated videos and images?

[blissful abyss]

still hurts and my character aint prominent

[fading dirge]

its ok

[heavy turret]

do you have to pay for it to use it?

where do i go in flow to check my "free" status? (see the price tiers)

[tacit aspen]

Yes, paying for Ultra currently. Looks like some new plans are coming.

[heavy turret]

ah i see, guess i'll avoid using Omni then :/

[inner wind]

[plucky scarab]

Guys i get ittt if you use the new things you are going to get more credits

im using flow agent for the past 30mins and i have 560 rn

[heavy turret]

this means im a free level user right?

[plucky scarab]

Yes

[tacit aspen]

It's only going to get better. This is day one... I'm loving it.

[heavy turret]

good

[fading dirge]

i mean if you didnt pay for anything then it should be without question that youre a free user lol

[inner wind]

I have pro plan and they added 1000 tokens before trying things

[fading dirge]

i mean sure, its all going to get better with time but currently its just ok imo

[heavy turret]

true but im always overly cautious :p cant help it

[wide moon]

Are they updating NB 2 and Pro

[gusty minnow]

1000 tokens, 30 tokens per video. I really like flow to generate photos, but I generate my video from videoexpress where I dont have to pay per video generation. its a flat fee and then you can generate as much as you can. And since flow add watermark logo, it's kind of useless

[tacit saffron]

hey , anyone knows how many credits we get with ai ultra (20To) instead of (30To) ?

[fading dirge]

i dont have watermarks on anything in flow, do you have ultra plan?

[gusty minnow]

I had pro before, now ultra because I stopped using video generation.

[primal dagger]

google flow only get this error now

[gusty minnow]

I have no problems with marking AI video with hidden ID, but watermark is a pain.

Also, restriction is so heavy, I can generate almost anything I want with videoexpress, never get restricted on anything generated.

[plucky scarab]

wait how they addeed tokens to everyone

but how 2 person got 5k tokens

[young acorn]

Hey, I create a new character in Google Flow (web) and I ask it to create a video with it, but I got an error about break of rules about usage of real person (I lost the error message sorry)

[inner wind]

I'm not sure maybe the plans will change?

[inner wind]

do they have Ultra?

[fading dirge]

omni is such a game changer...

[reef tiger]

Just wanted to say thanks for protecting the world from such horrible content! The world is so much safer without a simple scene of the first meeting of these two characters! Can't allow any type of conflict to exist, can we?

[fluid abyss]

How is the tools feature ?

[timid orchid]

Hi guys how many credit does it cost per generations?

[minor badger]

I've been requesting "delete on hover" for some weeks now. Hopefully they'll implement it.

[wide ether]

regading the looping that went away ..,.... right click on the video without playing it and there's an option to specify to have it loop .... I figured there was some setting somewhere .... but man I had to dig!!!!

[timid orchid]

I'm looking to get ultra again, how much credits does it consume for omni?

[fading dirge]

40

[glad isle]

What is the token cost for generating an image with Nano Banana Pro?

[timid orchid]

Sorry is it 40 for single gen or 4 variations? Like flow use to be 10 credits so 4 variation was 40

[fading dirge]

now im not sure, i think its 40 to edit a video but 30 to make one

2 10 sec clips cost me 60 creds, but eidting two clips cost me 80 creds

[timid orchid]

Ok thanks it seems a tad pricey, I wish they bring back 0 credits feature but I doubt it, this just came out

[fading dirge]

veo 3.1 lite low priority is still 0 on ultra

[timid orchid]

Oh I thought they removed it i cancelled last month because of it

[fading dirge]

nah they remove veo fast low priority and replaced it with veo 3 lite low priority

[mighty bison]

I can’t upload photos on phone now

[glad isle]

Yeah, doing anything from a phone gets buggy really quickly

[remote crypt]

it seems to be working on my phone. But I have turned on "desktop site"

requires a lot more zooming in/out and is finnicky

[flat quiver]

I built a simple pose tool with the new tool builder. Am I allowed to share the link here?

[sinful sequoia]

Yes! Please share 🙂

[blissful abyss]

send it man to dm

[wide moon]

Is there something wrong with NB 2 and Pro?

[inland pier]

Sure go ahead!

[glad isle]

Happens a lot on phones

[flat quiver]

Thanks ronny! Here is the link: https://labs.google/fx/tools/flow/shared/tool/9ee31860-4ad0-41eb-aa86-f5bccd647b74 Have fun y'all!

Works with or without a character reference, btw

[blissful abyss]

tnx

[flat quiver]

yvw

[blissful abyss]

please can you do for car shot

like upload an image of a car in a scene and then get multishot from the angle u want

[inland pier]

Very cool. I just gave it a spin.

[flat quiver]

I'll give it a shot and see what I can come up with!

[blissful abyss]

Kk

I am getting this @flat quiver

[glad isle]

Can you perhaps tell me the token price of generating an image with NBP on a pro account?

[wide moon]

Like is there going to be an update on the models ? Or how long will it take lol

[inland pier]

0 Credits!

[flat quiver]

Yeah, I haven't been able to get a preview directly in the tool yet - I think it's because I'm out of NB Pro gens for today.

I'll see if I can get it to use NB2

[blissful abyss]

nb2 will be nice

[flat quiver]

Make sure to pin the tool to the dock or favorite it - I didn't do that at first and couldn't find it! 😅

[sterile fulcrum]

Hey there! I've generated a lot of pictures of nice scenes or rooms, but they don't fit together like different rooms of the same house, so I tried generating a video from an image, where the camera flies from room to room- but instead, it usually kind of warps or teleports and does other weird stuff like that, especially if I give it a second image to end on! Is there a way to do this right?

[blissful abyss]

Kk

[inland pier]

Keep sharing your apps! I want to give them a spin.

[rain ivy]

yeah just woke up and tried a few things and apparently already hit the limit..... did they really reduce them by that much with the update?

[blissful abyss]

Have u fixed it

[glad isle]

No idea. Maybe ronny has more info.

[wind rapids]

same for me

[flat quiver]

Ok I fixed the preview function - if you are out of NBPro credits, it will fallback to NB2. It now shows which model was used. Also added a download button above the image (though they automatically show up in your project)

[stable basin]

And I just realized along with delete they removed cut and copy. The tools are just being deleted one by one.

[flat quiver]

I added a multi-angle function - reload and give it a try. It should work for any object or character, not just cars. 🙂

[blissful abyss]

Nice

[blissful abyss]

tnx

@flat quiver reshare the link

[primal dagger]

flashlight test

[flat quiver]

https://labs.google/fx/tools/flow/shared/tool/d3261bb1-1a28-4427-9c38-75127e2ae2ac

[primal dagger]

rat getting eaten by smol dinosaur

[inland pier]

Delete is now "Move to Trash". Cut and copy are still there, but now behind keyboard shortcuts (similar to your OS). You can ask the agent about some of these features. Try asking the agent "what are the keyboard shortcuts?"

[primal dagger]

made by flash omni

[last thicket]

So this is not veo 4

its just veo 3.1 guided by gemini internally?

[fading dirge]

I’m not going to lie. For me Omni is the same situation as Veo 3.1, mostly unusable aside from talking head videos.

[mighty bison]

Did you guys shorten the credits for Nano Banana 2?

[last thicket]

the persons look so npc and twins lol

[little fulcrum]

i figured out how to fix it. just right click on the video and the top option will allow you to loop. 😎

[wide ether]

Yeah I did as well … posted about it earlier … now keep in mind that that setting doesn’t stick. You have to hit that for every video

[north wigeon]

my flow credits got doubled whats going on?

[winged plinth]

Nice touch from the flow team✨

[rigid nest]

Looks like Cut or Copy function has been moved/ disabled. Can someone please confirm?

Also now many video generations allowed for each tier?

[brittle talon]

Hello! When I want to continue a video, the option to process it in 3.1 LITE isn't available anymore like it used to be. Does anyone know if there's a way to do this? Make the first segment in Fast and the rest in Lite.

[stable basin]

Ditto.

[dreamy oxide]

I sure hope that this "You're generating images to quickly" bug goes away by tomorrow. I have images I need to make and I ran out of chances for both 2 and pro somehow.

[winged plinth]

✨ 👌

[torpid cedar]

How is everyone liking Flow so far?

[winged plinth]

Runway in advanced mode.

[torpid cedar]

Oh?

[winged plinth]

very satisfying.

[torpid cedar]

I made a gif generator

[winged plinth]

Still a work in progress, but these tools are on a whole different level.

[winged plinth]

Inside it?

[torpid cedar]

Definitely I'm more into the coding aspect

[torpid cedar]

Yup

This was from my superhero app. Took a random photo

So much potential

[zenith current]

Why are unusual activities still occurring? Omni's performance still seems very poor. I feel it's still not suitable for creating outstanding works unless you spend a lot of money and constantly optimize the prompts.

[winged plinth]

Really fun to use. I haven’t nailed a full workflow/app yet, but it does almost everything I ask, the few errors feel more like prompt issues than tool issues.

[torpid cedar]

Yeah this is all entirely new so things are bound to happen with errors

[winged plinth]

Hopefully they hook us up with a generous agent usage limits.

[torpid cedar]

Indeed

[winged plinth]

I’ll give it a try tomorrow; the release load seems to be hitting performance.

[little fulcrum]

I know, another workaround is just use the arrow keys to go back and forth once.

[brittle talon]

for me the bad decision here is take out the option to continue the video on other model

[tacit aspen]

Amazing, loving the new model and adjusting my workflow. I also love they dropped Ultra to $200 a month. I haven't tried everything but I'm finishing a short that I started with 3.1 and the new shots more cinematic.

[torpid cedar]

Nice

[dreamy oxide]

I wish there was a way to know if the nano banana limit timer is up and you can make images again

[pulsar coral]

People kept deleting things by mistake that way. Also I asked for it to be "Recycle bin" but good enough. Just leave it at the trash and forget about it.

You're gonna be glad it's still there one day when you delete something by mistake.

Really who's never seen a recycle bin, this android generation srsly.

[stable basin]

I'm going to be gladder when I have the delete button back, if I loose a pic, no problem, I can remake it.

[tame falcon]

My impression is that Google's latest Omni is the strongest and best in the world, surpassing Seedance 2.0

[tame falcon]

Actually, from September to October 2025, I used the images in the doc to generate them. It seemed like it was Imaging 3, or Imaging 2 generating four images at a time. I miss it very much, and I also think that using either one can solve all problems.

[silver summit]

Where do I send some concerns at Flow? I live here in Japan but I can't read/speak Japanese. When they updated the Flow for new tools, it also changed the language to Japanese and there's no way to change it English... Can someone help please?

[wet flame]

literally just returned to create 🤦

[sleek sorrel]

ok so im not alone

[tepid loom]

What is the difference between credit and AI points

[wet flame]

All that new functions and yet to remove this stupid error

Btw to fix this, try to close the webpage or refresh it, it sometimes helps

[fading dirge]

What’s this mean?

[near garden]

should i be making my ai influencer with this new character video thing?
I just tried to make a few vids with my refreshed monthly credits and oh man am i annoyed. she's doing poses, speaking, moving in ways i'm not even instructing.

i use gemini for my prompts. I gen an image of my model in an outfit doing a mirror selfie, then i insert my prompt making her stand up and do a twirl briefly to show off the brand/clothing...
but now she's like 'hey guys check out my new fit, i wanted to twirl and show you how it looks' etc

[fading dirge]

Welcome to Google video models

[near garden]

thank you ._. i've been using google flow and video for the last month but it seems like whatever update hit today changed e v e r y t h i n g

[calm spoke]

What the hell is going on with Nano Banana in Flow [ingredients] the results are just like a cheap photoshop or a collage made by a child?!?

[eager hemlock]

How long was google flow down for the omni update

[calm spoke]

Bring Whisk back cuz this is getting very frustrating

[pulsar coral]

... if they'd put back Nano Banana (Gemini flash 2.5 image) that'd be around the same tbh, but they ignore every request about it for some reason. Like they don't care much about image generation or something.

Incidentally, NB1 images would also help VEO work better since the faces it generates look way less like copyrighted faces.

It’s NB2/Pro itself taking taking images from wikis and similar sources as reference material to refine outputs, which also makes them resemble prominent people more closely making VEO also fail — so it becomes an endless loop of errors.

NB2/Pro shoulda been taken down on maintenance long ago after the supposed 'update' it went through a couple months ago, and Flow shoulda gone back to Nano Banana (Gemini 2.5 flash image) until NB2/Pro was properly tested and fixed.

That if they even cared about all the user complaints.

[full blade]

How long is it going to take to fix the "Too Quickly" message problem?

[pulsar coral]

That's more likely happening because Flow can't afford to let people run 25 image generations at once like before. And it could be cause idk they might be lower in resources for some reason (?) but can't image Google not having enough money to attend that.

[lusty sinew]

What the hell, did the amount of images on NB2 decrease even on pro accounts?

I just started getting "too many images" again and this is the first time I've used my pro account in 2 days

[pulsar coral]

Considering all people who cancelled subscriptions due to how bad NB2/Pro are now something must be getting cut off there, whatever amount of money.

[lusty sinew]

I just cleared cookies and cache, I definitely have NOT hit image limit for the day like it thinks I have

[flat quiver]

Really like the new Tool Builder. Just finished a second utility: https://labs.google/fx/tools/flow/shared/tool/e60ccb08-b430-48e0-bed4-b2c66dd6b35b

Very basic for now, will update often

[vague igloo]

How do I use the tool in your workflow?

[burnt pilot]

Where the hell did the upload images botton go? And i dont mean the one at the top right i mean the one inside the prompt box

[neon raven]

lmk if u know

wondering as well

[flat quiver]

Add it to any of your projects in Flow, and from there add your character image. For best results use a front-view headshot/close-up of your character. Then just choose the options and Generate. It will output a front-side-back view headshot of the character with the new hairstyle along with the prompt. The images are save automatically in your project.

Hope this is helpful!

[plucky vector]

First off, awesome new updates, very excited to dig in and test them out further! I'm a member of a large AI content creator community that formed around Sora2, I shared today's update news with some friends over there. The feedback they had, which I'll pass along, is that people want feeds, public posts, subscriptions, remix chains, etc. I have no idea if that's the direction Flow wants to head in, but I just thought I'd share that thought.

This "tools" section is great! As someone with some traditional editing expierence, the idea of being able to just consistently apply the correct filters with the correct variables is just awesome, and then I can use AI to remix the tool itself further it looks like? Amazing! time to go rebuild the whole resolume openframe plugin library from memory...

[mental moon]

the new flow features are really cool

[full blade]

It's still not working for me

[cosmic thorn]

Has anyone used veo recently? I used it after like a week now that i needed it and the video are completely messing up the text. This never used to be the case. What's happening?

[near garden]

i've been rate limited for over 12 hours

[plucky vector]

Wow, I am in LOVE with this new "Tools" section! Not only do I like the general idea of it, but, creating my own actually worked really well, basically nailed everything I wanted in only a few prompts. I was on the fence as to if I need CapCut's mess of an effects system; this would solve my desire for something like this, and I can make my own!
https://labs.google/fx/tools/flow/shared/tool/394066ee-89f7-4bdb-b7c3-a5c5cc2710b7

[plucky vector]

Neat use of a reference image to indicate motion there!

[glad isle]

anyone else can't save characters?

[rotund helm]

Hi. Why am I getting only policy violation when using omni? Tried with different self-shot phone videos, no harmful prompt. Similar with anyone else?

[broken fox]

Yes i also

How to solve policy violation

[proper path]

Please tell me why the Flow image creation tool completely rejects the image of the famous historical figure I'm working on for my project about the history of Vo Nguyen Giap, and the tool refuses to create the image.

[dark shale]

im fail always how can fix??

[sour maple]

does anyone know hos to fix this "unusual activity"? i'm already trying refresh the page, change the conection, trying on incognito mode, deleting cookies.

[noble carbon]

make it possible to paste @character name and it automatically @'s them then typing it in each time seems like a UI issue u have to enter each @ manaully

[sick sparrow]

anyone else having problems logging in?

[proper path]

nahhhhhhhhhhhhh

[vestal shore]

How did you do this?

[wicked shadow]

yes

[hazy obsidian]

same for me

[stoic wave]

me too

[sick sparrow]

quite unstable since yesterday

[quick oasis]

I'm starting to have this issue experience at this point now

[empty furnace]

Hi

The number of requests sent exceeds the quota limit. Error code 253

Is this normal?

[hazy obsidian]

@fervent bane @tough yacht guys help us

[empty furnace]

Or maybe I've reach the qouta or just system problem?

[quick oasis]

it's the system problem. It had happen to some several people from yesterday earlier during the afternoon

[empty furnace]

Aight! thanks!

[quick oasis]

sure thing

[desert grove]

bro

https://klipy.com/gifs/papa-man-gaye-indian-memes-1--k01KS228NSZZ30ECZCQPVQ2VENE

[quick oasis]

also just to give y'all a heads up, whenever you start to generate, this is gonna popup. And I can't even refresh it either at this point rate. So seems like it's a glitch for right now

UPDATE: Oh wait, never mind. I had tried again, it's good now. I hope

[lament hamlet]

• 1

I thought I was the ONLY one...

[dark shale]

Everyone is the same.

[quick oasis]

yes, I started to notice now

[tribal fjord]

well well well, We need to let the IA rest huh?

[sharp fulcrum]

🫠🫠

[lament hamlet]

Yeah, I guess we broke it... LOL

[dark shale]

im think same lol

[glad yacht]

are those Videos generation error or image 🤔?

[agile geode]

A normal update ❌
Adding new bugs and pissing users off ✅

[lament hamlet]

ABSO-$%&-LUTELY

[glad yacht]

I suspect the Google server backend was overwhelming with today's update and massive amounts of users.

[lone seal]

This new character feature is really cool! Now, if only it would stop crashing out of my project, stop stalling on load, etc

[glad yacht]

They are still calling the new pricing ultra plan as Ai Pro 🤔?

[lament hamlet]

Welp... At least it's not an issue solely one is standing by themselves. Guess all we have to do is, remain patient.

Maybe get some sleep, or step on grass...

[oblique zealot]

I'm having the same problem too.

[tacit saffron]

on gemini or flow ? i still have credits on flow

[cosmic nymph]

frequently encoutering this since 18 hours before.

[plucky vector]

Hey guys, test this tool thing out, let me know what you think!
https://labs.google/fx/tools/flow/shared/tool/7ee94cfa-730e-4022-8104-27c523558419
The effects could use some further tweaking. More importantly, I need someone who knows how this works to give me some feedback on how to coach this tool-building AI into making a "render" button which actually works, it currently can't render out.

[honest light]

same, i have to refresh the page each time that happen

[glad yacht]

Are you running more than 8 generations parallel?

The tool builder is like canvas in flow, we can build our custom Whisk app inside 🤔

@pulsar coral

[lament hamlet]

Hmmm... Seems like it's working again?

All loadin' fine here.

[ruby pendant]

Flow has worked fine for me for months, until this morning where I'm suddenly encountering this. Is it just a temporary setback because of high user volume or is this permanent? I can't use it for work if it does this, and I'm leaving 5 minutes between requests.

[lament hamlet]

Nah, still acting up... Worked for a few minutes, tho'.

[glad yacht]

Overflow 🤣..

[open plover]

it seems very random to be able to open a project...when refreshing the page for a few seconds it loads videos and images and then crashes

[hoary basin]

I got this error: ‘The number of requests sent exceeds the quota limit’ (Error code 253). What’s wrong?

[plain apex]

flow crashes while generating

[sick sparrow]

seems like crashing out is not happening when i stay on the same tab, still sucks

[honest light]

flow seems jittering by it self now and then

[mellow tiger]

This state has been going on for over a month.

[sudden heron]

The number of requests sent exceeds the quota limit. Error code 253

[dark shale]

It started working again just now, then it went back to 253 again.

[noble carbon]

the cali google staff waking up for on call server fixes

[spark bramble]

Man, I created my own character, using a self portrait I took, and though it MADE the character, it's own guardrails refuse to USE said character because it thinks I am a prominent figure. The filtureing on that is the worst.

[lament hamlet]

Hey, at least it thinks you are famous!

[lucid mango]

show it to your friends😆

[spark bramble]

even more fun, if I try to use the custom voice I created.. it just says "Something went wrong" and fails automatically.

[agile geode]

https://labs.google/fx/tools/flow/shared/tool/fc42a6b7-9bfc-4f75-b4f0-9a3174946d44

You can use the tool to turn a doodle into an image, and then use Omni Flash to create a video.

[spiral python]

Always error

[burnt light]

what's the problem?

[pure fog]

To be honest, I’m quite disappointed. I tested character sheet workflows and the results were terrible.
First, I asked the assistant to help me tailor a simple prompt based on a character sheet. The sheet contained a cat character, but it generated some random dog character instead. Is that a joke?
Then I tried a more traditional workflow where you upload a character sheet and specify the camera movement and character animation in the prompt. For the sake of the experiment, I used the same prompt I used in Kling Omni. The results were terrible again. The animation itself was okay, but it generated some random cat with absolutely zero style consistency or resemblance to the original character sheet.

[spark bramble]

At least the image generation is a blast.

[glad yacht]

Is your snail doing moonwalking 🤣
The trail of slimes explain everything... 🤔 backwards...>>

[spark bramble]

No, see, he uses the same trail over again so he can reach faster speeds 🙂

[heavy geyser]

The next wave of AI content is here!

[heavy geyser]

I get same error

[ivory radish]

same error. even though i did not generate any for the last 1 hour

[hollow sigil]

Hi everyone, I have a crucial question regarding the newly updated Google AI Ultra tiers and how they sync with Google Flow credits.

In Egypt, we now have two "Google AI Ultra" tiers under the "Choose your Ultra plan" menu: one with 20 TB storage (5x usage limits) and the regular one with 30 TB storage (20x usage limits).

My question is: Does the cheaper Ultra plan (20 TB / 5x limits) still include the unlimited "Veo 3.1 - Lite (lower priority)" generation at 0 credits once the main monthly priority credits are exhausted? Or is that zero-credit lower priority channel exclusively reserved for the higher 30 TB Ultra tier?

I really need to confirm this before upgrading my subscription. Thanks in advance!

[elder folio]

Flow totally crash since 2 hours ago

[vague island]

flow is bascially unusable rn

[heavy geyser]

no status page?

[versed garnet]

it's unusable half of the month... last month was very stable, but it gives error for days this month.

[opaque sand]

I remember it was the same last year.

[heavy geyser]

today was I/O day so it's acceptable

[gloomy verge]

Hello. Can I ask, if you create some images, you use some template to prompt? Because, when I write to prompt example "Female", she don´t looks so nice as @spark bramble images in flow-images, example, that white monster in canal, or what is it. 😄

[elder folio]

both template and prompt

[gloomy verge]

How template I must use? Because my images are no so good than his images. I probably bad write prompts, or where can by problem, that my images are bad then his.

[lament hamlet]

Managed to create some images through Flow, but even after not making that much, now I'm getting the "requesting generations too quickly" message. Mm-neeeeh...

[elder folio]

template is used if you have referrence. but the prompt is very importain. you can write a specific prompt by using chatgpt

[spark bramble]

If your females don't look good.. it's because you haven't prompted it to look good 😄

[elder folio]

i haven't kicked out of Flow for few minutes. Does it back to us?? 😂

[spark bramble]

You just have to know how to sweet-talk it 😄

[gloomy verge]

Do you used NB 2?

[spark bramble]

@gloomy verge Yes.. it's so much fun 🙂

[burnt light]

is error 253 fixed now?

[tiny magnet]

why I can't log in into flow is the problem for everyone?

[hushed valley]

does anyone's flow keep signing them out as of the last 2 hours?

[chrome galleon]

flow kicks out when i try to generate

[burnt light]

yea me too

idk why

[chrome galleon]

i miss whisk 🙁

[burnt light]

when are they going to fix the error 253

[marsh inlet]

what happened ?

[wet flame]

same, getting this too

[chrome galleon]

same. i didn't generate anything yesterday whole day. now it says this

[burnt light]

idk it does that to me too, I tried using different accounts. Although flow beta on google play store works but it has a limit and kinda sucks. I do ai story videos and now I can't generate cause of this

[chrome galleon]

same

[hushed valley]

I have a question. For videos, there is the option where you can put in a photo for the starting frame of the video and one for the ending frame.

For example, my starting photo is an empty bench, ending photo is woman sitting on that bench. When it makes the video, it'll generate a whole new woman that has nothing to do with the one in the ending photo. She walks in, sits on the bench, and then somehow morphs into the woman from my photo by the end of the video.

It never had a problem with this, but only started doing this recently. Is anyone having similar experiences and what can I do about it? And I cannot be any clearer in my prompt, too.

[surreal hare]

Why agent mode error when using guys?

[burnt light]

is error 253 fixed?

[calm spoke]

I just tried some of their tools, i thought maybe i am the problem.. and the results are just like cheap Photoshop... I tried the characters option and in all the results their head is bigger than their bodies.. as if it's those old photoshop images where u put the head of someone on the body of someone else. And the consistency is almost gone.
And even with that agent, i still have the problem of error messages.. i can't generate more than one image in 10 minutes...
To generate a new image i have to wait 10 min or more... Sometimes the agent itself tells me to wait and it's keep thinking for long minutes...
Just bring Whisk back

[kind dirge]

is it me or is flow image limited generation reach was downgrade? Cause it feel like it hit generation limits too quickly than usual. Is there a better ai site that is same as flow and Gemini

[twin mortar]

With the release of Omni - does this mean anything different for Veo users?

[wet flame]

Legit, just bring Whisk back, and pair it with Nano Banana engine

[brittle bloom]

Whisk as tools on flow honestly but still 25% aka still bad

[marsh pewter]

Did the remove lower priority mode??

[steady helm]

just fix flow, intern must have done something with the update to be so buggy

[marsh pewter]

Wdym

[steady helm]

[marsh pewter]

????😭😭

[steady helm]

I'm just being sarcastic, with the reach qouta error

they must have done something wrong with the updates

[stable helm]

How have they rolled out all these “improvements” but not resolved the generation errors that have plagued people for months?

Google are not a serious company.

[fading dirge]

yes im more disappointed now because i understood that veo 3.1 is not good but told myself its ok, their next video model will be good... i was sadly mistaken

[mellow void]

I got about 20k extra credits yesterday, but now all of a sudden they're gone again... Ultra user, by the way.

[fading dirge]

same, crazy too because i was going crazy testing omni because of the extra credits. Now they are gone lol

[glad isle]

Everything is a mess, generations failing, error code 253, force logging out, uploads not working, characters not saving, etc. I don’t think there is much left to fail. To top it off, not a single statement about any of this.

[open plover]

same here with the 25K credits...thought it was for testing and burned through half of them and now my regular credits are down

[glad isle]

I couldn’t even burn any😂

[fading dirge]

yeah id have never done that without the extra credits lol, so now i burned through alot testing omni just to get a bunch of trash outputs lol

[mellow void]

They got us good. Also burned through A LOT of credits, genius tactic.

[fading dirge]

Well anyways, if it wasnt for nano in flow id justy unsub from the ultra plan, but the speed and the amount of photos you can make with nano is pretty good. going to continue to use other video models for videos. omni aint it

[frosty pine]

why is it keep saying this prompt may violate our policies relating to minors

prompt: At first a locomotive train is visible, with no semi-high-speed train in sight. Shortly after, a 12-carriage semi-high-speed train comes whizzing past at 200 km/h, and the camera angle shifts to focus on this semi-high-speed train. On the side of the semi-high-speed train, there is a digital display showing the Hindi text logo "वंदे भारत".

[glad isle]

Lol. What does the hindi text mean?

[frosty pine]

Mumbai railway station

[glad isle]

Yeah its been crazy for the past week.

[exotic shell]

i just received this message for my AI Pro membership:
"
What’s changing starting today, May 19, 2026:

Usage limits in the Gemini app: For the Gemini app, we’re introducing compute-based usage limits that factor in the complexity of your prompt, the features you use, and the length of your chat. Your limit refreshes every 5 hours until you reach your weekly limit. As an AI Pro subscriber, you’ll enjoy a 4x higher usage limit than non-subscribers.

AI credits: The product-based usage limit model is also rolling out to other products, starting with Flow and Antigravity. You can extend your limits by purchasing AI credits. While 1,000 AI credits will no longer be included as a benefit in your base plan each month, the new usage limit model we are introducing should allow you to maintain the same experience you are used to. To learn more about how to use AI credits, please visit our Help Center.
"

I have a 12 month AI Pro subscription with 1000 credits a month. Now i will lose those 1000 credits a month have to purchase them? 😄
What happens to those who pre-paid Pro for a year or have a Pro-Subscription from a promotion? Now the whole subscription got useless.

[tough yacht]

Hey! You / Pro subscribers do indeed get 1,000 Flow credits per month (they're called Flow Credits now, instead of AI Credits). More information here: https://support.google.com/flow/answer/16526234?hl=en and definitely DM if me if you don't get your 1,000 Flow Credits whenever your next monthly credits cycle starts.

[tough yacht]

Nope - it's still there! Ultra subs will still see Veo 3.1 Lite [Lower Priority] in the model picker

[exotic shell]

Hm, okay. Well, to me it didnt sound like that in the mail. Thank you for clarifying. So not much changes for me as flow user?

[thin steeple]

whats wrong with flow?

[tough yacht]

Sorry for the confusion there. That's right, you still have 1,000 credits to generate with, they're just now "Flow Credits." But again, if you don't get the credits you're expecting at the time you're expecting them, just DM me.

[tough yacht]

Hey! What are you seeing going wrong right now?

[glad isle]

Pro users get 1000 credits. Free users get 100 and 50 per day, which amounts to 1600 per month. So we pro users get less?

[echo spade]

I can't generate any images in nano banana in flow. Ultra user here.

[glad isle]

Few can.

[echo spade]

Okay well that's good to know

[echo spade]

Is that for both tiers of Ultra subs? I don't see it on my end. I have the £79.99 ultra tier.

[tough yacht]

Good question: the now-$200 (used to be $250) Ultra tier continues to have Veo 3.1 Lite lower priority model available. The new $100 Ultra tier does not have it.

[exotic shell]

there is tiers for ultra? first time i hear that. interesting

[tough yacht]

The new tier is brand new: https://blog.google/products-and-platforms/products/google-one/google-ai-subscriptions/

[echo spade]

Thanks would have been good to know that before subscribing.

[obtuse flint]

They are going to shut down flow #killedbygoogle

[barren mulch]

cannot modify a simple video to even just change the background with omni, i get a false flag for "This prompt might violate our policies about generating harmful content". anyone else?

created a feedback bugreport already but just wanted to check

[obtuse flint]

Omni model is too netured and it's still not good. It's just voe 3.5. Seedance is still miles better. Even grok is superior at this point

[pliant bramble]

Hello everyone,, I wanted to know, Is nano banana still giving unlimited generation in google flow for google pro users?

[barren mulch]

wrong chat, this is for flow, jules is one section down in the side bar

[fading dirge]

veo 3.2

[obtuse flint]

Son

[pliant bramble]

?

[obtuse flint]

they added limits like 3 months ago. so its like 100 images per day

[pliant bramble]

it was unlimited untill the yesterday update, but now it never generating more than 50 images,

[proper path]

Nothing happened with Google today.

troll~3

[obtuse flint]

nah. it always had limits

[pliant bramble]

alright,,,,

[dry niche]

flow always fails with "unusual activity" why is that?

[gentle scroll]

Good morning, is Flow now charging credits to extend a video or download the extended video?

[torpid cedar]

Not sure about that

[marsh pewter]

Can't see it anymore bro rn also im also a ultra member

[gentle scroll]

I just tested it, and it's actually using 5 credits when I extend the video and download it. I don't know if this is a general problem, so I'd appreciate it if anyone else experiences this.

[torpid cedar]

Only time they charge credits to download if your asking for an upgrade scale

[barren mulch]

why problem? you are generating another video, so that is going to costs something, yes.

[torpid cedar]

1k, 2k etc

[barren mulch]

720p download does not cost anything more since you are downloading it as rendered

[gentle scroll]

So, before, when extending the file, the resolution option appeared when downloading, but now it's not showing up. It only shows the download option, and when I download it, it uses 5 credits.

[flat quiver]

I think they are automatically saved - you don't even have to click the Generate button. I had to refresh to get to show up though. I think a little clearer indication in the UI is needed for this step.

[crisp basalt]

what does the "reference" do in agent instructions?

[flat quiver]

I just gave this a try, nice work! The output is supposed to autosave to your feed, but yeah I don't see it. Maybe ask it to add a "Save" button? I had an issue so I simply told it the problem and then asked it to troubleshoot - it should fix it

[crisp basalt]

saving this for later

[crisp basalt]

true

[torpid cedar]

I've mostly only made image apps currently nothing video related

[jaunty wing]

What happened to the upscaling? I can't do it on many of the videos anymore. The option diappeared, and now I can only get 720 videos.

[flat quiver]

Hey Zhele, I made a Pose tool last night for Flow that also does multi-angle sheets - give it a try and let me know if it helps. Just upload the front view of your character and choose the options you want. It should generate it, I've been getting consistency every time. https://labs.google/fx/tools/flow/shared/tool/c88d04a7-4ac8-48bc-bb47-89978c344a86

[torpid cedar]

Nicely done. Probably could do sprite sheets too

[crisp basalt]

https://tenor.com/view/tpose-t-pose-meme-t-pose-to-assert-dominance-slipknot-sic-gif-25072741

[torpid cedar]

Used for an animated output

[crisp basalt]

jokes aside, in eed to try that!

[flat quiver]

Thanks! Oh yeah, sprite sheets would be a thing. I will try that today!

Gotta run to work, but a quick example first. It also saves the prompt and you can copy each view if you want individual renders.

Ignore the messed-up anatomy on the side view 🤣 I've never had that happen jeez

[torpid cedar]

Lol that's a new perspective

[tulip lance]

I recently noticed new Google Ultra plans are launched in India
One is worth 6500₹
And another one is the same 24500₹
I wish to know if both have 0 credit video generation in unlimited capacity.

[torpid cedar]

Neither does

[tulip lance]

That means we can’t do longer videos like we used to do like earlier
I had done a 3hour video few months ago

[barren mulch]

3 hours?? damn

in 8 seconds chunks

[torpid cedar]

You probably still can, but things are shifted currently. Just like any new product chaos happens for a few weeks.

[barren mulch]

Veo 3.1 Lite [Lower priority] is currently free on my 20x plan but not sure about unlimited

[tulip lance]

Ya its was done single handedly including music and editing. It was a deadly work.

[tulip lance]

Ya if that low priority mode still exist , it’s nearly unlimited.

[placid sluice]

does anyone know if the new $99 Ultra Plan watermarks Veo clips?

[thorny lynx]

Not able to download any video

[novel plank]

When I start a new project, and I click on Characters, it's not showing Characters I've made in other projects. 🙂 I would love to see a list of characters from other projects I've made, perhaps in folders I can manage.

[blazing cobalt]

hi

[stark grotto]

???

[novel plank]

Getting errors. They must be updating something. 🙂

[crisp basalt]

yea its hitting again

its probably there updating something or its peak hour

[stark grotto]

love it

[crisp basalt]

https://tenor.com/view/hot-day-are-you-ready-for-some-football-gif-16278812894835706568

[peak loom]

Hello, anyone know why I can’t use « reuse the prompt » when I click on a generated image ?

[wet flame]

How nice, nothing works rn

[crisp basalt]

oh bruh

so its not just flow

[peak loom]

Yes flow is restart all 3 minutes

[quartz current]

Anyone else struggling with the scene builder not allowing a second clip to be added?

[oblique dagger]

idk what's happening with the 429 since the update

but that's starting to be annoying

[flat quiver]

There's a workaround - click the plus sign in the prompt field, and at the top left you'lll see a dropdown. You can browse to your other projects and click the Characters option from the sidebar. Then just click the character and it's added.

[severe lily]

Just got kicked out of Flow. JFC. Honestly. I need to get a partial refund on Ultra because it's beyond unreliable.

[dark shale]

I'm having the same problem right now.

[cinder falcon]

Just subscribed to the new $99 new ultra plan and Veo 3.1 Lite [Lower priority] is not there, is this a bug?

[severe stream]

Anyone else having constant refresh and then "The number of requests sent exceeds the quota limit. Error code 253" ?

[empty furnace]

Same same

[wicked crater]

Same Same the curse continues...

[empty furnace]

just white screen when I referesh iot that

[novel plank]

Yes, either overloaded or they are updating. Yesterday same error, they were updating to this new system so guessing there's updates.

[exotic shell]

they mentioned it above, it seems to be only in the 200$ version

[empty furnace]

I can't even get though the project hahaha it always go back to white screen then the error message

[severe stream]

Me too

Ughhhh

[analog radish]

Guys where are the controls for the camera when edit a video?

[flat quiver]

Same, just constant refreshes. I click Generate and then it crashes. I'm going to step away for a bit 😅

[empty furnace]

Hopefully it will be back after 24hrs lol I need to generate some images

[lament hamlet]

Morning fellas! Waking up after sleeping late because Flow decided to act up with Error 253... And it still is! 😄

[cinder falcon]

Uh ok, thanks

[civic kraken]

Flow not working?

[silver valve]

whos facing unusual activity

[severe lily]

gestures broadly

[fringe geyser]

@flat quiver cool tool would be cool to do a 9x16 of single character too so u can get full rez and also closeups on face with added skin detail - not sure u really need that top panel from a top different view of the overhead shot in one of your examples on white - be more usefull if it was a face closeup so u can see eye details

i cannot get to 1 project now after some update at 2 am last night it just dissapeared but i did just find out as a workaround i can get to it thru the pose app from @flat quiver when u go into the character panel to add ur character u can switch to a project in dropdown then u can use arrow to back out of that tool to your project - so i assume u can make a tool in the new app builder that will allow u to do it

[flat quiver]

The headshot mode in the pose tool will give you a portrait / closeup, but it's not extreme - I'm working on adjusting it today and adding other angles.Yeah, I'm thinking about just deleting that top view in the multi-angle, it's actually kind of annoying to me at the moment lol.

Aspect ratios are on the way in the next update, btw

[unreal viper]

hi all, please use #1238621511401734254 to submit any bugs! these threads are actively monitored by the team and so may be seen faster!

[rain basin]

With the new update, how do you check how much credits is used per family member? The previous link no longer shows that. ❓ ❓ ❓

[fringe geyser]

@flat quiver cool didnt see the headshot maker till just a sec ago

[lime skiff]

Does anyone know how many credits the AI PRO (5TB) plan will now receive instead of the 1000 usual credits?

[plucky vector]

Hey guys, I updated this tool a bit further, let me know what you think! Hit the random button a few times and tune it in! The "render" button is not working, I don't know why, the AI tool claimed it worked but it fails for me.
https://labs.google/fx/tools/flow/shared/tool/7ee94cfa-730e-4022-8104-27c523558419

[mighty bison]

Why does it keep doing this?

[coarse bane]

is the reuse prompt option remove for creation ?

[olive spire]

Hello guys, is there any easy way to generate images through nano banana without uploading manually? I used to work on other models with URL + prompt

[flat quiver]

Really cool fx. Do you mind if I remix and see if I can fix the render function?

[burnt pilot]

Failed
You're requesting generations too quickly. Please wait a moment and try again.

Like this since yesterday smh

[bold falcon]

Okay, so here's a good 1. My third month of Ultra at a discount, $124.95, ends on the 27th of this month. And yesterday, after the keynotes, I of course received 25,000 extra credits. Later on, that same day, I went back down to 21,000 from 47,000. And there's some parts of Google that say I'm now on the 99 month plan, even though I'm supposed to be on the $250 plan. And all those credits are now gone. Has this happened to anyone else or have they noticed? Thanks. Seeing if I should go straight to support or did I screw something up?

[noble carbon]

yea i saw 43k creds then 18k creds once reloaded super weird

[dark shale]

i miss collection

[chrome galleon]

Flow is acting crazy 🙁 bring back whisk

[brittle canopy]

I am trying to use the Agent in FLow, it pules for a bit then does nothing. What am I missing?

[coral void]

veo3 bugged?

paid user, and i have credits, but aways this:

[sick sparrow]

same here, but i believed they've reduced pricing for ultra subscribers

[prime dew]

Anyone have an update on when Flow will stop saying I'm generating too quickly?

[sick sparrow]

anyone had any success already with video to video?

my drone shots are keep getting harmful content as an error, even though it's just a cinematic shot of a forest

[bold falcon]

Using the agent for my avatar videos was amazing. It created 15 of them at once. 2 iterations for my 3 thumbnails next time I'll only do 1 shot on the thumbs. And 8 infographics at once. Crazy cool. Loving it. Next up later today, after Google I.O. is teaching anti-gravity, using browser control to go to my flow, upscale each of them to 4K and download them to my drive. lol. Conceptually, that should work, right? RIGHT? By the way, I asked Agent Mode in Flow and it said it can do up to a 100 at a time. I don't think I'll ever use that with that many. Can someone else try that and report back here? Yeah, thanks.

[bold falcon]

Thanks I'll probably leave it and see what happens on my renewal date unless I run into a limit while trying to make something if that's the case I'll go right to support. Well it looks like there's several of us in the same boat so hopefully they'll just figure it out and fix it.

[bold falcon]

They did they introduced a $100 a month plan with 5x more usage than Pro and They reduced a $250 a month plan to $200 a month, 20x limits which was the plan that I was on But I was in my third trial month at 124 95 In some parts of Google say I'm on ultra $250 plan some parts say I'm on the $99 plan now So I guess we'll see. Thanks

[bold falcon]

If you have ultra, I'm not sure about pro, sorry, but on ultra use the agent mode and in the agent mode give it all of your prompts at once and it will create all of them bypassing that issue. Or at least it did that for me last night

[quick oasis]

Still have that issue with me too ever since from yesterday as well 😞

[bold falcon]

Try pasting that entire prompt inside of Gemini chat and ask it why it's getting flagged, have it checked for words that would trigger it, and if it finds any to rewrite your prompt and then just use that as the template next time.

[inland pier]

maybe the side panel is not automatically opening up. if you are in agent mode, can you try clicking the expand button to open up the right side panel?

[jaunty wing]

Last night I suddenly noticed I had 49,000+ credits, and today I have 20,000. I havent figured out the new AI usage scheme yet. Not sure if its better or worse or what. Anyone fully understand what it will be in the end?

[bold falcon]

The only thing I can think of and I didn't Use the find function in the keynote transcript from YouTube Was I don't recall if they were saying that those extra credits were only for new subscribers or for everyone But I thought it was for everyone But I could be wrong. I'm wrong quite often, but I'm just dumb enough to admit it.

[bold falcon]

oh, also my rainbow ring around my google avatar is gone now as well. I wonder if they're going to put everybody on the $100 plan and then make them upgrade at their next renewal, perhaps, like them airing on the conservative side versus the aggressive side. But that's not what they said yesterday.

[brittle canopy]

Something went wrong. Please try again.
Try again

[coral void]

aways

[brittle canopy]

On Desktop Chrome

[inland pier]

hmm that's frustrating. can you create a new session (the little pencil edit button above), try your prompt again, and see if it works

[brittle canopy]

I had already done this, but did this again.

[coral void]

It's only accepting generation via Omni, it doesn't accept generation via Lite. Why?

[bold falcon]

Here's a suggestion, go to your Gemini Chat and ask Gemini Chat to create a prompt for you of a flower vase on a table or whatever else you're trying to create and say it's for VEO 3.1 or it is for Omni, then try that prompt. Also, here is the blog OmniGuide posted by DeepMind. Usually, I upload the guides as knowledge when creating prompts. https://deepmind.google/models/gemini-omni/

[brittle canopy]

Oh I am already using skills for that @bold falcon !

[inland pier]

It should. You can either specify Veo Lite in your prompt "Using Veo 3.1 Lite, make a video of ....", or you can update your agent setting to default to Veo 3.1 Lite.

[bold falcon]

try this one.I said omni and I should have said nano banana pro or 2. Close-up studio photograph of a single textured river rock centered on a clean, solid matte gray background. The rock is oval, charcoal gray, with visible white quartz veins and a rough, weathered texture. Sharp focus captures every grain and pit. Bright, directional lighting from the top-left casts a soft, subtle shadow to the bottom-right, emphasizing the three-dimensional shape and surface details.

[brittle canopy]

@inland pier it should work with Nano Pro/2 non? It shows up when image gen is selected.

[bold falcon]

Better to be structured rather than short and precise.

[inland pier]

yeah it should work with all models. make sure you are in Agent mode.

Still working to improve the safety filter over-firings.

[coral void]

how to delete this agent?

[inland pier]

Are you asking how to switch out of agent mode or how to start a new session (I.e. a new conversation)?

[bold falcon]

I fixed mine by saying "using the referenced subject" been flawless since for me. I asked gemini

[brittle canopy]

I got that as well, I even structured it to say that my generated characters were AI generated non-real people, and I got that filter.

[coral void]

leave the agent, delete, apagar, out off this

After they added this agent, I can no longer generate my Lite videos; it always gives an error.

[bold falcon]

I was told by Gemini when I had a similar problem that it absolutely dislikes negative prompts. So when asking it to, don't do this, no on this, no on that. It is better to choose positive prompts to describe what you want and then leave the rest out. My prompts have been more successful when doing it that way on rewrites.And I baked it into my instructions for prompts.

[bold falcon]

Try This, or re-enter it in Gemini for the full context of your prompt.

[rigid nest]

Any clue how to move an image from one project to another?
The cut / paste feature is no longer available.

Also what are the limits for images and videos for each tier?

[marsh pewter]

The veo3 lower priority is removed for ultra members ?

[bold falcon]

Still shows on mine right now. If you're talking about fast, lower priority, that 1 was removed on the 10th of May.

[marsh pewter]

Google ultra membership right?

[bold falcon]

yeah I have ultra that's on desktop is where my image is from so I'm not sure if you're using it on a phone or tablet or an Apple device for that matter. But yes, seems to be missing from yours.

[brittle canopy]

@inland pier : main account with AI Plus, it works. A family memebr's account under family it does not work.

[tacit plaza]

How to provide reference images while creating images in flow,like there was option in whisk???

[wicked crater]

Shame I don't have Ultra yet.

[inland pier]

If it's not highlighted, you are not using the agent, you are directly prompting the generative media model (Nano Banana, Veo, Omni)

[stuck remnant]

Download > 1080p (Upscaled) upscaling videos in Flow is broken - I'm getting artefacts (glitches) in the video. I'm on Pro. Doesn't happen with the "raw/original" file

[rigid nest]

@inland pier Sorry to tag you directly. I have been looking for these answers since yesterday but no help from the community yet.

Would you be able to guide me in the right direction please?

[coral void]

After a lot of struggle, I managed to find a way to work around this bug that prevents the video from generating. First, enter a random prompt, for example: bla bla bla

Second, send it, and when it's generating, cancel it. After canceling, enter the correct prompt you want, and it will generate.

[errant lark]

Flow website is not working

[coral void]

I hope Google fixes this; it's really annoying having to resort to workarounds.

[inland pier]

Copy/paste are still available, but just not via the 3-dot dropdown. You have a couple of options here: if it's a single image, you can copy by selecting (hold ctrl/cmd and click the asset) and then pressing ctrl/cmd+c. Go into your other project, and paste (ctrl/cmd+v). The other option is to add everything you want into a collection, download the collection, and then drag all the assets back into your other project. Note that you can always use assets from other projects if you click the "+" in the prompt box, and use the drop-down tab to search through other projects. Hope that helps.

[deft osprey]

failed.
You're requesting generation too quickly. Please wait a moment and try again"

help i been getting this for full day, and i did try again later after some wait it still the same, it happen like 2 pm today till now

[sacred mica]

Hello guys, i am struggling with a picture, i have 2 images in 1, like half top and half botom. Wich prompt can i write to make him understand that i want bot of the but in 2 separate picture?

[rigid nest]

Thanks for the response! It would be nice to have batch copy feature as well. Or just drag from one project to another.

How about the image and video limits?
Where can we find additional documentation?

[inland pier]

yeah this is a good call about making it easier to copy assets into other projects. let me bring that back to the team.

we don't disclose daily limits, as they often fluctuate. are you noticing that you are hitting your limits?

[rigid nest]

Not recently, but such a bug has been flagged before and we requested better error handling.

Just saying rate limit exceeded is ambiguous. At least have a timer when user can try next.
Some users have "Generating too early" error after 4-5 attempts.

And it can't be related to one account as this issue has been flagged a few times.
Personally I have no such issue. I had the error once before recent updated and it went away after 12 hours.

There is an existing ticket as well on Feedback channel. I am not allowed to reference it here.

[brittle canopy]

The agent should be able to be carried from project to project? + as well, option to make a character project specifc or available outside that specific project.

[inland pier]

Yeah we can look into this. Since projects often have very different intentions (e.g. photoreal portrait for social vs stylized landscape for short film), we tried to silo things between projects. Nonetheless, this is a good point, and it's been raised a few times. We should at least give the options of easily moving things (agent sessions, characters, assets) between projects.

[inland pier]

Ahh you are talking about short term rate limits and not daily caps. Yeah these also differ from model to model (and different tiers). We can do a better job of giving more information back to the user once this happens. Maybe even eventually we allow a user to queue up their requests rather than getting throttled.

[brittle canopy]

I must say with using the agent, I am comfortable gen 2 vs 4, and allows to strategically utilize NB-2 to full advantage! Go Googs!

[small compass]

cool tool! I believe I got the export to work, you can try here: https://labs.google/fx/tools/flow/shared/tool/3b92b6cf-3c2e-4512-9f82-9725e7a69eaf

[inland pier]

yeah i am curious to hear how you all are using the agent. would you rather prompt directly to the models, or have gemini help you tweak prompts, generate variations, edit in bulk, etc?

[lime skiff]

@inland pier Do you know how many credits the AI PRO (5TB) plan will now receive instead of the 1000 usual credits?

[glad isle]

It is still 1000 credits.

[lime skiff]

ahhh.. OK.. Tks!

[glad isle]

Don’t buy pro for videos. You get 1600 monthly tokens as a free user.

[brittle canopy]

I have been having a convo with it, adjusting, its like google gem but in the chat, and responsive and knowledgable. It is maintaining the same character I generated through different scearios, without having to create a character on the side.

[lime skiff]

Could you please explain this? Are you talking about the 50 credits per day?

[glad isle]

Yes, 100 monthly and 50 per day.

100+(50*30)=1.600

[lime skiff]

50 credits per day isnt enough to me,,, But I get it what youre thinking

[quick oasis]

Same here too

[flat quiver]

I only started using it in the last hour, but I'm finding it really helpful for making (fashion) mood boards - sort of reverse-engineering from existing generations.

[flat quiver]

Moving assets between projects is THE feature that would make my year, because I've been terrible about keeping house lol. With the copy-paste method I can only download 1k resolution images, the 2k option isn't available after moving the image.

[meager edge]

Does Omni flash does not support the Frames to video , Anyone here does face the same problem

[novel plank]

From what I understand, it doesn't. You just put in references and it'll try to replicate from the images.

[void crater]

its been i guess 2 weeks now and they still not fix . in fact it got more worse now.

[meager edge]

ok thanks

[pulsar coral]

Yeah been curious about it will give it a shot later.

[bitter sage]

@tacit halo you useless.

[pulsar coral]

Then maybe 1000 wouldn't be enough for a month at all either. Google's math for the credits is indeed off.

[bitter sage]

google flow veo3 Service you useless

[bitter sage]

I don't have time.

@brittle canopy why

whyyyyyyyyyyyyyyyyyyyyyy reaction?

help me

When will it start?

Subscription No I'm not

@lime skiff

@brittle canopy

[quick oasis]

Mhm, will give them time. I’m sure they’ll starting fixing the issue soon enough as possible. Hopefully sooner

[void crater]

yes.

is this happening with most of them right? or just few?

[bitter sage]

When will it start? Subscription No

[quick oasis]

It’s been happening to several people that who’s also having this issue as well too so far ever since from yesterday and last night too

[quick oasis]

I don’t know. And I’m not a subscriber to ultra either. So you might to ask someone else about that part

[bold falcon]

Hi Ronny, It's been really good so far. (see my post further up. 9:41 am today), Right now I have Claude creating the prompts and scripting based on the VEO 3.1 guide as well as the nano banana pro guide that I uploaded as knowledge to it and with agent yesterday I was able to do 14 avatar iterations of me at the same time 3 thumbnails and 8 additional assets after that. Agent also changed my camera movements to seconds based, which was an improvement, Worked really good. The simultaneous generations is fantastic, but a lot of that is due to me adjusting and asking Gemini for help in tightening up my prompt structure instructions within my Claude co-work. (side note, I plan to try all of it with antigravity next). My plan is to use anti-gravity browser control to then upscale them to 4k and download them to my Google Drive because project download now is only 720p which is too low I want 4k. Other than that, so far so good. Maybe allow the agent to at least upscale it to 1080p or 4K, since we're giving it permission to use the credits for Omni (30 per iteration). It's not a big jump to have it upscale each image.

[inland pier]

Great to hear. Yeah it won't be long before the agent can do everything that a user can do in the UI. You should try to make a tool that has your prompt preferenes baked in.

[rigid nest]

I have used it for renaming, but my guess is that users without technical background would need some basic video tutorial on how to use it to it's max potential.

Would there be any video series from Google or we have to rely on influencers?
I prefer the official source always 🙂

[inland pier]

i can talk to our marketing team about making some deeper dive videos on tool creation. I'm sure there will soon be many such videos from the community on YouTube and X.

[rigid nest]

Thanks! It's good to get some videos from DEV team and add it here. It can be used as a way to funnel more users to discord community and get additional feeedback from new userbase.

[bold falcon]

I do indeed have that. I'm not shy, this is what I have Claude do for me. Runbooks rule it for token conservation, As that is in the instructions, since I only have a pro account with that. Here is a small part of what I have it do. It took me months to build it (all me btw), so growing my channel slow but I'll get there just trying to get a little bit better each video super small right now.

Here's the other two sections. What changes for me is since I have Agent now along with anti-gravity I'm going to try to move the entire process over to Google. So pretty much all I will have to do is piece it together in Resolve. But Resolve has MCP or API so I might be able to get it to drop everything in as well. That's the goal anyway.

and one last pass. PS - The second image is really the first 1, then research I have to approve, then it goes into production, and then distribution and final outputs.

[stray remnant]

google flow once again with an update that just makes it worse and more confusing

[quiet wadi]

nano banana 2 has a daily limit?

[acoustic grove]

@bitter sage Hey everyone! Just a friendly reminder to keep all conversations in the supported language for this server. This helps our entire community engage effectively, and we've gone ahead and deleted the recent posts that didn't comply. Thanks for your cooperation!

[marsh pewter]

Anyone who can help me know why my flow ai on ultra membership isn't showing veo 3 lower priority anymore?

[hidden tangle]

You're sending generation requests too fast. Please try again after a few seconds.

How long do I need to wait for that error being solved?

[copper hawk]

whats omni flash

[marsh pewter]

It's new model

[hidden tangle]

kryptoine

could ya helop me with the problem?

[marsh pewter]

Sure tell me

[hidden tangle]

.

[marsh pewter]

It happens when there of a server overload so expect couple hours then try again .. it usually happens when a new update rolls out

[rigid nest]

Try after 1 hour please

[hidden tangle]

got it

[hybrid pawn]

Yesterday credits doubled, today they were taken away. Bad form google!

[calm spoke]

since the update i have this proble;... it's like a cheap 2000's photoshop. the head is always bigger than the body. and the consistency is almost gone. what the hell is this. and i'm om schedule

[marsh pewter]

Anyone idea on this please?

[lament hamlet]

This is really annoying... Just when you are in the brightest moment in your project, you want to keep going with the flow (no pun intended) and when this, little message from HELL shows up, feels like a splash of sewer water into the face! Yesterday I even did fewer generations than the usual, and this freaggin' thing came out sooner! What's happeniiiiiiiiiiiing!?

Did we get Flow broken? Is the AI finally rebelling against all these humans requesting weird pictures catered to hidden fetishes and memetic silliness?!!!

[rotund elm]

IN Google Omni, is it possible to use a voice from elevent labs as reference. it looks like you are locked into their voice buikder

[quick oasis]

it's a bug issue. I'm just assuming of it. Because several other people are experiencing the same thing too

[lament hamlet]

Oh, well, I've experienced this issue along the last month, but it's mostly because I've started to use Flow more often and getting the hip in a few comics projects I've been working on hard! I even started to get kinda used to it, as it worked again after, lets say, 12 to 24 hours... But it was after making a GOOD quantity of pictures, but yesterday, I remember generating fewer than the usual, and this thing hit sooner. 🫠

My guess is, there's some sort of limit or daily "quota", but it's annoying not knowing it's daily balance/status, so we could be aware of how it's going before it crashes!

[wide moon]

No the limit is staying the same it’s more of a server overload

[lament hamlet]

So yeah, we basically broke it. LOL...

[inland pier]

Currently we don't support voice upload

[quick oasis]

to any of google staff members or anyone, just let me know whenever this gets solved again hopefully soon. Because I don't even know if this is a bug issue or a normal regular thing that I have to wait for like the next 2 or 3 days for it to get solved itself

[wide moon]

The only model isn’t affected is Imagen 4

[stuck adder]

Is is ultra the only plan that doesnt add the water mark??

[main tundra]

Made this song with help from Flow, I made the trumpet and drum melody in Studio One. Flow helped me add in some violin and put it together for me :)

[lament hamlet]

Finally, there we go. We are back over here!

Allowed to continue until it breaks again, sorry Artie!

[main tundra]

Whoa, horses!

[atomic bough]

"suce"

[old bay]

Hey man i'm not a google official, but I got an email with a subscription announcement update. Generating videos now has a daily limit that I think refreshes after 5 hours and then you'll hit a weekly cap.

[exotic shell]

you also have to live in the right country. if youre in the wrong country you will have a water mark with ultra as well. 🙁

[stuck adder]

really? wow lol

crazy how everything is getting falgged lol wth

flagged

[quick oasis]

I’m not generating videos. I’m doing images only. I already know about the daily credits for videos. Plus I think I finally figure out of what’s going on of what I had read that was on the lab feedback channel in this server

[plucky vector]

Feel free, ping me if you get it working! Also, it did a great job on creating some of the effects, but a few of them are a bit weaker and should probably get a cleanup pass. Other to-do ideas would be to add blending modes (lighten, darken, difference) between layers, RGB bypass buttons for each effect, and maybe some way to tie effect's values to other data, like playhead position, audio FFT, global BPM counter, etc. I'm basically just listing off the features of Resolume 2.4...

[plucky vector]

Oh wow you did get it to work, very cool! Thank you for taking the time and sharing the results!

[near garden]

i get the same thing. i can only generate 2 batches of 4 images at the most, and then it hits me with that message. and maybe after about 40 or 50 in pro or 2 mode, i hit the limit on both. yesterday i couldn't generate any images pretty much from 9am until the next day.

i logged out, tried incognito, went on my phone, none of it worked

[quick oasis]

It's a bug issue. There no limits change to nb2 and nbpro. We just have to wait until whenever the google staff can fix this problem

so as for now, just take a break from now until further notice

[proud edge]

i thought i was the problem too..

how can we convince them to bring back whisk lol.

[proud edge]

since when is whisk back?

how did you bring it back??? i see you have tampermonkey.

[crisp basalt]

its in tools section

[proud edge]

damn, it is there ..

guessing you have to pay extra to use it :/

or that it bugs out and fails

[fair gull]

Is there a place to report bugs for Flow? I have a few that I've run into while testing.

[proud edge]

bruh yeah i just saw it

you're requesting too quickly! sure buddy, sure i am.

[crisp basalt]

its free tho

[proud edge]

lmao

[crisp basalt]

thats what happen when youre on free plan

[proud edge]

🫩

[errant gull]

i am paid plan and have same issue

[crisp basalt]

create a new google acc , then generate

pro plan also have quota limit

[smoky vault]

So not talking about the remain issue, the tools section updated in new versio,n it seem to be great feature but seem we can only share link to anyone and cant using it for private team yet

[proud edge]

how good are the results in this mid unoriginal whisk lol

[quick oasis]

this has nothing to do with the free plan or paid plan. It's just a bug issue that everyone else is having right now. We have to wait until whenever this is fix. Hopefully sooner as possible whenever the google staff members gets time

[crisp basalt]

Yes

So anyway,

[crisp basalt]

As long as u only share the link to your private team it should be fine

[crisp basalt]

Yes

[proud edge]

i'll take that as they are shite

[pastel badge]

Look who's Back

[crisp basalt]

yes if u prefer to look that way

https://tenor.com/view/billy-crystal-tissue-kleenex-tempo-stop-crying-gif-15048056

[proud edge]

we've seen but its pretty mid im guesing

[flat quiver]

I'm loving the Agent feature. It's awesome for iterating new shots.

[sacred prism]

I haven't used Flow for over 2 months, but when I came back to use it again, there were only 7 credits left. Isn't it supposed to get 1000 credits every month? I am using the Pro plan.

[languid cedar]

Do you pay the pro plan in 3 month you said?

[sacred prism]

Im using year plan

[dark shale]

Hey. evreryone Is it working again??

[frail wave]

no, image generation with nano banana 2 and pro is not working.

[toxic lily]

Same issue

[tough ridge]

I have a question if there happens to be a developer or staff member on, with the new system for credit management and usage inside Flow. With the changes to the Google AI Pro subscription that I have (being that it's based on daily usage), what does that daily usage actually equal? Where would I find that information? I know it said that essentially I would end up with the same credits as before but how can I verify that?

[tough ridge]

You should check your email because they just sent out an email statement that stated that the credit system was changing even if you had a Plus Pro or Ultra account

[tough ridge]

And in reference to your only seven credits left, I'm assuming that at some point in the past you bought a credit pack and there were seven of those credits left. When they removed the thousand credits from your account because you no longer get a thousand credits every month and it's now based on usage, it's showing you have seven credits because you had seven credits left from the credits you most likely bought at some point I could be mistaken but it's what makes sense.

[halcyon knot]

I upgraded to AI Ultra $99 plan from AI Pro yesterday and ever since I am not able to generate any images on Flow.

Tried signing out and signin in, created a new project, tried both NB2 and NB Pro, with and without reference images. Nothing works. Video generation using Omni Flash works fine.

[modest hollow]

Me too.I upgraded from PRO to ULTRA 3 hours ago.And then it didnt work anymore

[quick oasis]

Same here too. This is gonna be a very long patient while time

[modest hollow]

Are you also the $100/m Ultra member?

[plucky bronze]

What i have to do here

[quick oasis]

nope, just a regular free member. So if you're wondering about the bug issue, it's been happening every single person between ultra member and non-ultra members for a long while time right for the past 2 days so far

[sage otter]

Day 86 of not being able to search flow projects

[glad isle]

How is everyones experience today? Seems to be finally working for me.

[marsh pewter]

I still dont have lower veo 3 lower priority mode on 100$

[proud edge]

not working for me lol

[marsh pewter]

Anyone please help with this

[proud edge]

this is why we need whisk fx back.

only that works is gemini

how tf is it that im generating images too quickly

💀

[left sequoia]

I'm using Gemini Omni to generate videos based on storyboards, but I frequently encounter errors and are blocked because the system flags keywords that might relate to policies concerning "celebrities." However, these figures are generated by AI. How do you all avoid this problem?

[void crater]

I just come to know that it may not be a bug . its compute usage limit. free users are now get nothing even it shows 0 credits and pro plan members now get 0 credits instead of 1000 monthly to create videos. and they will also face usage limit during image generations and all. Thats why we are facing these issue. so it may not be a bug . thats why devs are not replying. There is also debate going on X about google AI pro plan= "Google's Gemini AI Pro Adds Compute Limits, Draws User Backlash". Its a trending topic. So we are just wasting time waiting for devs to fix it. because its not a bug and it will not fix.

Also I pasted this on gemini and asked is this true?. and i got below answer-

[clear glacier]

flow error

[void crater]

🥲

[quick oasis]

will first off, you shouldn't have asked gemini for this. Because this has nothing to do with this issue of what every member is dealing with now right now. You should've contact the google staff members either between here or through email to get the real humanize answers. AI doesn't always solve anything of asking these type of topics like this. So don't even try to ask ai of anything for that. Learn to follow your own guts and heart. That's all I can say here. I'm just trying to put in a human positive perspective way for anyone that who's also having this error issue too. Because right now this is not natural when it keeps saying "you requesting generating images too quickly" through two days when someone comes back to check it again later for it to resolve itself. So...yeah. Again, hopefully this will be fix at some point soon. I'm just only assuming of it that's all

But for anyone that who's stuck in the same situation, I had found a new workaround way that you can continue generating fun images of whatever you usually be doing. Just use the new google flow agent to generate more images of what you want to do. And it'll work just fine. But if you ever asked the agent assissant to generate like 16 or 20 images all together of like different scenes or random stuff. Just be aware that some of them will show the same message error of you generated request too quickly to some images that didn't go through. But you'll still have other more images that did went through. It's a balance of that. But other then that, just use google flow agent for generating images. It's actually way beyond better than you think. it's really freakin cool 💯. Go and try it out y'all 👍🏾

[wide moon]

Here’s the problem On the free plan the flow agent doesn’t show

[quick oasis]

really?. Hmmm, will It shows on my end on my free plan that I was able to use. And it is free for every user to use too. It shows on the without subscription list

[wide moon]

Nope I don’t see it

[quick oasis]

huh, will that's very strange. Seems like you got a glitch issue there. Because this is what it looks like for me

oh and also here's the list of it too

[wide moon]

So now what ?

[quick oasis]

you should try to contact with the staff members of this issue. Either on here, or through email if you can. If they're also avaliable to contact right now

[wide moon]

Hmm

[wide moon]

I don’t want to make a another feedback lol

[quick oasis]

it's okay, that's totally understandable. But if you ever still feel stuck, you should never be scare to ask for help 💯❤️👍🏾

[burnt pilot]

Failed
You're requesting generations too quickly. Please wait a moment and try again.

Im still getting this😮‍💨😮‍💨😮‍💨

2 days already

[quick oasis]

Yeah I know. You're not alone. But have you tried out using the google flow agent before?. Because that's a workaround that I discovered and have been using right now to resume using NB2 & NBPro making many more images. And it's really great so far

[glad isle]

Yeah, started getting it again, but only on NBP

[burnt pilot]

Nope not really, I will try it out

Btw where is it?

[quick oasis]

it's in the prompt where you see it says agent on it

[wide moon]

Do you see it on your end? I don’t

[burnt pilot]

Are you using android?

[quick oasis]

okay so minor set back, looks like in google flow agent has it's daily limits for the first time that I just discovered for anyone that is a free users in google flow like I am. Which that sucks now. So be very mindful and careful with that. So apparently it's not really much of a workaround in continuing to use nb2 and nbpro without the error. So that's a bummer. Looks like we still gotta be even more patient on whenever the error of "generating too quickly" message is gonna be fixed again 😒. So um never mind on that workaround method 😞

[burnt pilot]

Yeah it doesnt work on agent as well

[wide moon]

iOS

[burnt pilot]

, Cause it only works when you have the page on desktop mode

[wide moon]

Oh and it didn’t work

I’m on iPhone

[burnt pilot]

Don have this on ios or something similar?

[quick oasis]

yes and no. But there is a daily quota on the number of Agent queries. So it's kinda like when you're using veo 3.1 for videos of daily credits. But alittle different at the same time

[wide moon]

Nope

[calm spoke]

I'm on a paid plan and have the same problem for 2 weeks... Yesterday they started to give another error message, saying they've noticed an unusual behavior while i only generate 1 image every 10 minutes....

[surreal hare]

How long year we will could be used agent mode guys 😅

[burnt pilot]

THen only on pc

[tame falcon]

I don't mean to be racist. I also oppose it. Has anyone noticed that if the prompt doesn't specify a particular person, VEo video models almost always output non-white people?

I casually wrote a few prompts Left one (elderly woman planting): Appears to be East Asian or Southeast Asian. She wears a traditional conical hat, with facial features and skin tone consistent with the region's ethnic characteristics. Left two (man walking a dog): This man has a lighter skin tone, with features typical of a white person (Caucasian). Right two (woman working in the field): Also wearing a conical hat, darker skin tone, appears to be Southeast Asian or South Asian (for example, a Vietnamese or Thai farmer). Right one (woman holding ice cream): She has a brownish skin tone, with South Asian, Middle Eastern or Latin features. Among the four people, only one is white.

I wrote about four women and they all turned out like this....This set of images is clearly generated by AI to showcase diversity, covering different races, ages, and life scenarios. Among the four people, only the second man from the left fairly fits the typical 'white' characteristics.

If you directly specify a photo, it violates real people, but real people do not have to conform to reality. On the streets, who would dye their hair blue, and AI in China mostly represents relatively normal ethnic types.Veo had this problem before. I don't know if anyone noticed.

[tame falcon]

If it's a generated image, then not a single black person will appear.

[cosmic nymph]

Did the quota or using limit of nano banana pro decreased a lot for Pro subscribers? The “you’re requesting generating images too quickly” comes earlier than before. Nano banana 2 is still there but I am not dare to try without a good purpose.

[somber venture]

Hey guys, I am currently encountering an issue with Google Veo. I am unable to generate even a single clip. I have tried all models including Fast, Lite, Quality, and Low Priority, but I am still facing the same issue.

I also tried tweaking the prompts, refreshing multiple times, and even changing browsers, but nothing seems to fix it.

Is anyone else facing the same issue, or does anyone know a possible solution? Any help would be really appreciated.

[tired nova]

Is it possible for VEO to think a character it created has become famous due to it being used for character consitancy? I prompted a character last year in July, and then use that same prompt to recreate it as much as VEO 2 would allow and then use screenshots from then to use as ingredients for video generation. Now I am getting the error message about generating prominent people. Does it think the person it created is now and actual person?

Edit: it stopped doing that, I wonder if something else caused it to do that?

[compact phoenix]

@somber venture ugh that error is the worst lol. quick q — you using any vpn or proxy? i had the same thing happen and it turned out my vpn was the problem.

[peak loom]

Hi, just trying the new update on FLOW, but do you know why he don’t create 2-3-4 images with one prompt ? I already check the 2 images in the parameters but he only generate 1

[cosmic nymph]

To be honest, of course every thing has a limit, but if making more complexed image will cost more, at least let me know how much resources that I have consumed.

[small compass]

sure thing! you can remix it and work off of it to add all the other ideas you have in mind. hopefully the export function will remain stable.

[smoky socket]

https://www.flowmusic.app/invite/I4RUW5

Anyone Tried Flow Music?

[proud edge]

is flow still not working for you guys

i just logged on today and it still is broken lolz

[mossy furnace]

quick flow workflow question for people making longer ai film sequences

when you move past a single strong clip, what is actually holding up best for continuity

prompt structure, reference images, scene planning, or doing more of the edit logic outside flow

we are testing this with filmmakers in india right now and the hard part is keeping character intent, shot language, and emotional tone alive after the first few scenes

[raven ferry]

Hello mates
I am looking for real opportunities with teams that need help now.

[river rain]

Hi, Is there any plan to support more aspect ratios like 2:3 for image generation which are supported by api and AI Studio. It would be great if we have more options in flow as well.

[unique abyss]

Flow Music is fun, and people with paid gemini accounts get access at various levels. Sometimes when you are doing videos, it's nice to have Flow Video just handle dialogue and sound effects, and prompt Flow Music for the sound track. (Assuming you're assembling stuff in a external editior, not just trying something like a one-shot, one-scene short.

[smoky socket]

Yeahh

I think is great alternative to SUNO

[ocean oracle]

After checking out Flow again for the first time in a while, I noticed there’s now a tab called “Character.” Can anyone tell me exactly what it’s for and how good the results are? If there are any videos about this feature, I’d appreciate any links.

[pale flame]

I asked a few months ago in Google Labs/Bug fixes, if the team please re-enable SPELLCHECK with the red squiggly lines within the prompt box? This keeps me from generating too many unnecessary images or data/electricity use as well. Too many "mispeled" (even this word is underlined in Discord) words. I've updated Google Chrome and all of it's settings. Does the red squiggly underline appear in the prompt box in Flow for anyone ealse?

[gray lynx]

Hi folks. I created a character in one of my projects with a specific voice , and description model. How do I make it so that this character can be referenced in any new project mainintaining all of it's previously created characteristics. I did try the plus Icon in the promopt and hit the drop down list to a prior project to reference but it just references the character image and not the personality or voice. Thank you for any help

[violet ruin]

Can someone please tell me or actually explain to me very simple what does agent mean in VE03’s new update I tried it and my generation came back terrible and I used a based image

[copper mica]

Hey everyone! Just a friendly reminder to keep all conversations in the supported language for this server. This helps our entire community engage effectively, and we've gone ahead and deleted the recent posts that didn't comply. Thanks for your cooperation!

[violet ruin]

Show you prompt

[inland pier]

great question. think of the agent as a creative assistant, that allows you to brainstorm, ideate, generate, edit in a more fluid way than having to select the individual models in the specific UIs. For example, in agent mode, you could say "make 4 different images of @character battling in different wars throughout history" --> "for the wars that took place before 1900, make them black and white" --> "now rename those files to just the names of the wars". This simple conversational-style prompting allows you to direct your generations w/o needing to get prompts right, and it also cuts down on several clicks (or even research outside of Flow).

[stuck remnant]

Safety filters are overdoing it a lot - when will we expect this to be looked at? days/weeks/months? @inland pier

[inland pier]

which safety filters. feel free to DM me your specific prompt / references.

[inland pier]

short answer is that we are aggressively looking at this, and we will make it better

[gray lynx]

Hi folks. I created a character in one of my projects with a specific voice , and description model. How do I make it so that this character can be referenced in any new project mainintaining all of it's previously created characteristics. I did try the plus Icon in the promopt and hit the drop down list to a prior project to reference but it just references the character image and not the personality or voice. Thank you for any help

[stuck remnant]

sent a request

[violet ruin]

Thank you for explaining to me this I appreciate it even though it was technical but you broke it down to very simple and I tend to bounce if things are too taken you made it very understanding thank you

[bold vector]

is anybody experiencing unusual activity issue ?

https://klipy.com/gifs/reupload-doraemon--k01KS5J0J063B9VAZ8GFMTMDRMZ

[scarlet quarry]

So Gemini Omni Flash is way cheaper in credits consumptions than the other 3 Veo models, namely Veo 3.1 Lite, Veo 3.1 Fast, and Veo 3.1 Quality ?
I presume those credits are all for working in Google Flow?

Source page:
https://support.google.com/flow/answer/16526234?hl=en&co=GENIE.Platform%3DDesktop&sjid=16511283258121715377-NA

When i go to "AI credits activity", I read:
AI credits let you keep using AI models and features when you hit a plan usage limit.
info
AI credits included with your plan have been replaced by product-based usage limits

What are product-based usage limits?

[ancient plank]

how do u access this? it just takes me back to google flow

[scarlet quarry]

Click on Tools, then the last one is Whisk.

[scarlet quarry]

Where did you found this info?

[somber pebble]

Been getting this a ton when trying to generate the last 2 days. Seems like almost every time I change a prompt or ingredient they'll fail until I refresh the page. This happening to anyone else?

It will work if I copy my prompt,. refresh the page, then enter it all again, it's just frustraing and slows down my workflow

[echo spade]

Same for me. No answer from the team about it.

[stuck adder]

hi is there any moderators here now?

when i try to change myself into another character or a animal i always get "cannot generate harmful content"

is there a way i can fix this???

[noble carbon]

unusual activity is too trigger happy

[fair gull]

Does anybody know of a way to upload audio into Flow? I'm trying to build a tool that will take audio and segment it into 10-second segments to be used inside of Omni for music video generation, specifically for lip sync. Whenever I have the agent build me a tool, it always defaults to the media grid, which has no button for uploading audio. I know Omni allows for audio to be used as an input but this doesn't seem to be available in Flow.

[fading dirge]

how possible it it to make a 30 second explination video with omni seeing that you cant extend and all that? is there a way to do this?

[paper ermine]

This is frustrating. Tried everything. Even new session. Delete cache, try different prompts, Use fewer images, 1 image, 2 images, Simpler prompt, NB2, NB Pro,.... Nothing works.

[ancient plank]

I wonder if omni will have lower priority version too but ig its almost the same quality as veo