#flow-general・🌊
1 messages · Page 32 of 1
This image create image thing is so amazing especially without the watermark and it’s HD
yea im glad that they didnt put the gemini watermark
And u get to do x4 at a time
On mobile (no add to prompt button)
Without going back in Gemini reupliading again and again and again new that etc
on pc
I heard nano 2 is around the corner
You have to click the button 3 line thing
actually theres no button for that
i see the issue now
Oh I see
No
@summer rapids so im guessing you want to be able to add genereated images into the lets say for frame to video start frame box
Yes, you have to download it and reupload it
For mobile on desktop you can add directly to the slot
ahh thats a bit hassel
i guess they want to make sure that the image is properly fit to 16:9 or 9:16 before using em as ingredient or frames
hence the need to download and upload it
Just a missing button, not a big problem, it's just out some hour ago
actually
@summer rapids have u tried do page zoom out on your mobile browser
cuz the button is there actually but its hidden due to screen size
this is how it looks like if u do scale zoom out
yes just a resolution bug
they need to change it cuz i know that same button placement , is also for "add to scenebuilder: button which they purposely made it not visible on mobile browser
Yes that works turn your phone horizontal and touch the area and there’s a arrow
^that too
so i guess that would be temporary workaround for now, but until then /./. ill go ahead and flag this in #1319424187730755636
Is anyone else having their images generate in a square aspect ratio since the last update?
My first few images generated correctly in a landscape ratio; now they are all square, and so they do not work properly when trying to generate the video, since the video outputs in landscape mode
@summer rapids yo i made a new thread in #1319424187730755636 , u can go ahead and bump it so that the flow team can see it
I totally randomly asked it. I'm so shocked. I didn't know this had already been worked on 😲
Is there a way to add a sample audio file to use a custom voice?
ElevenLabs does this
nice to see the flow team implementing community feedback, i remember mentioning during one of the studies how helpful it'd be to have a way to make static images directly in flow rather than gemini before moving to video. definitely speeds up the workflow
is this nano banana 1 or 2?
They all get stuck at 99% and then fail.
Constantly redoing is a waste of user time
Videos generated by Ultra subscribers will also be in the wrong order due to generation errors, and it is very inconvenient to not have one-click downloading.
Hello guys. I generated a lot of videos and asked for it to have voice but it didn't add any generation at all. How do I claim the credits back for it?
Aşk
Hello @rough jolt we wanted to remind you of our server guidelines, stated in #start-here・📍 : “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now.” Please keep in mind that non-English discussions and content may be removed. Thank you.
I think it would be a crazy good feature but not sure if veo supports it
i heard about nano banana 2 what the case
what's this do? No action when i click it
do you have any generations? it should make the view more dense!
Hello. I need help creating videos for my product. I want flow to use my product only that is ring, and create a good video. I tried giving it my product image and few designed images of my ring like girl wearing it. But it didnt exactly create a scene, it is changing my product in the middle of the scene. What is it that I am doing wrong here?
yes i have many videos, nothing happens when i cick it
ohh it looks like u are cooking something.. did u taste it?, if it good , can i know how did u cook it? , like looks like kinda a new type of urgs..
Prompt Like a Pro, a recurring live event series in the Gemini Discord, will be taking place tomorrow
with a section dedicated to Veo & Veo prompting! Hope to see you there! 
im sorry, has there been any announce on veo 3.1 fast and quality as to when the accent for other than english is gonna be fixed? since veo 3.1 is only usable thru labs, as fars as i know, and if i can use veo 3 please do tell me, it is been a waste of time and moeny, no matter if i do this in labs or any other service providers, even use the same prompt expander from veo 3 text to image, it can just accept even if simple dialog lines are being specified for each one they language and accent and then in the end "all the characters speak [language] [natiional country dialect] accent from [country.] i had to force this on veo 3 every time and it worked. but now with veo 3.1 nothing works. NOTHING. not one single line of dialog for one single character. completele ignores the directive. call it from multiple service providers, same. waht is the issue? someone knows if google has said something abotu this, like a fix? anyone , who is not speaking in enlighs has been able to actually use their own language with their own coutnry in veo 3.1 withotu any issue? how? any help will be EXTREMELY APPREACIATED
anyone having problems with audio? in flow, using veo3.1
yes
no audio generation since the update
no audio generation at all, i have a deadline for the ai filmfestival, this is unacceptable, i payed over 100€ for that crap?
this is extremely annoying, i have it too since the update
just for curiosity, you are tryng to get dialogue in english USA and music or sfx does nto work properly? becase if that is so, then we are a little bit mroe ...more troubled than i tought. i just wasted a lot of money to have 1 or 2 dialogue lines with almost no context , all of the context goes to reinforce the "spanish argentinian accent from Argetnina". i have repeated that in a simple prompt like 4 times ,and then for each dialogue line, 1 before and 1 after . so in total it goes way beyond 10 times for a single line of dialogue for only one chracter to speak in such accent.
Yessssss
hi all! seeing a few reports of audio issues in Flow - if you haven't already, please submit your experience in #1319424187730755636 so we can take a closer look. please include as much detail as possible!
it has no structure, it happens on some accounts apparently. my friends has no issues. when i open another project it kinda work 1 out of 4 generations has audio, but my shortfilm for the festival project doesnt generate sound at all anymore. what scorsery is this? we have a deadline 6 days left, this is unacceptable. i payed for this.
i tried creating a new project, doesnt work. as if the ai wants to tell me "hey ur short movie sucks, dont work on it anymore"
totally understand the frustration here 👍 please add these details to your report in #1319424187730755636 if you haven't already, so we can take a closer look. thank you! 🙏
Hi, may I ask you something?
Is there any way, when creating a video with Flow, to send it an image of the product that I want the person in the video to show?
I created total 14.2M Reels views in a month using Google Flow in Korea.
Based on this experience, I’m thinking about creating a free YouTube course for beginners.
What would be good to teach them?
Tell me in this thread 😁
Can Chinese-speaking users participate?
paro
veo 3.1 is something every day. in labs fx same image + same prompt, all good. in a service provider, without any type of prompt adition, just add the image and the text,that's it, same image, same prompt, failed more than 7 times. why?
I also tried that actually. But still no luck ☹️
I want to create videos for my work on flow. I need tips and tricks to make good prompts and improve my videos. Information on how to edit videos, tell stories, and use visuals effectively would be very helpful. Thank you!
Could someone please help by providing ways to create effective prompts?
There are hundreds of videos about all of this on YouTube - they would be more in-depth and helpful to you than a chat here.
is Veo back to bugging again? Can only queue 1 animation at a time instead of the regular 5, it just claims I am requesting too fast lol
How can i use the MIXBOARD, anyone can help please
I have content being flagged...its an image of a female tennis player...generated by me
is this common?
update effed up everything, i cant generate dialugue anymore, already used 10000 credits, what is this nonsense?
is anybody here working for googlelabs ? this is extremely infuriating guys.
no solution at all since 18 hours, no credits refund, nothing, i payed 120€ for nothing.
i wont use google profucts anymore, i will klingai to finish my movie.
go ahead , nobody’s stopping you from chasing your goal
another case of weirdo with too much money but lacks critical thinking skills ..
if there is no dialogue, the sound generation works fine. or if i promt: oldman and friend are talking", then it generated some random dialogue which works. BUT, if i want to generate my script dialogue for example: old mab says:"hello", then audio generation wont work. what is this crap?
you poor little guy, go find a work, then yuo will have money too.
couldbt handle not being my buddy lmao
must be a democrat
still no solution, how are we supposed to finish our film now? i want my money back or immediate solution guys? this is unaccaptable, your update effed up everything
You do know that getting things exactly how you want isn’t in the cards when it comes to Gen ai?
Why not just pay traditional artists to get it done lol
I see that veo 3.1 is 0 credits for ultra subscriptions (really nice!). How long can we expect that to last? (With the sticker shock of that subscription, just evaluating how long I should continue that). The 0 credits on that model makes the ultra sub worth it!
Exactly
They never mentioned whether the 0-credit fast mode is limited-time or not, but yeah it’s definitely the best benefit you can get out of ultra plan
Video
On the main flow page it states "Try for free now" ... to me that implies a trial period then the charges kick in. If charges kick in, then that makes the ultra plan being worth it kinda borderline.
Sorry in advance for being new.
It seems like most of the issues I'm having are fairly common, makes sense, it's a new thing.
However....
Im getting a lot of violation notices, something about prominent people, when I try to generate clips of myself.
First, the idea that I am prominent is laughable, but I'm unable to name my own Youtube channel, show my own logos, even using my own name can present a problem.
Considering most of my content contains characters that are different versions of me....
Is there a fix for this?
Google knows I am me.
Hello everyone. I've had the opportunity to test Flow through the free Pro trial, and although my journey was a short one due to how fast the credits were consumed, I have to say it was a pleasant one.
I've managed to get a few shots for my film (the ones after she gets out of the car), and here is the result.
Hii. Maybe you are a prominent person? 😄
I agree this is a problem, I haven't seen it that often. But as soon as I try ANY logo, even a logo I made, it does something really wanky with it. It's useless for animating logos so far.
If I might add a few words, this is more of an experiment to see what can be done with the tech at the moment. Currently, the first minute of footage should be cut, as it does not belong. I have sort of an idea in mind for the type of story I would like to tell.
Secondly, there's some shots inconsistency between her and her father in terms of the environment and lighting which should be solved. I probably would've solved those if I had more credits.
Third, the film would probably not start with her in the park, this scene would probably happen later.
What you're seeing is basically footage which could pass for production, and the only reason it starts this way is because of the lack of footage I was able to generate.
Overall I am impressed with what is currently possible. If someone had the time and the budget I think it would be possible to put something together.
Also be aware that doors have major issues. It will open doors at the wrong side (rotating around the side with the door handle), and will hallucinate a new door handle on the other side once the door closes again... etc etc.
so avoid animating doors if you can
From what I seen Grok too seems to have issues with car doors.
In terms of opening an closing them at least.
I've never used Grok, shitter-account deleted 😛
Not tried latest Kling either. But then again, talking about other apps in here is against TOS I think.
I think each one has their own issues. And it's still kind of early-days.
What I hope from Veo (or others) is more user-friendly GUI, with a much more powerful timeline-slider (and better designed) that allows you to step frame-by-frame, instead of second-by-second.
And in the far future, some integration with audio AI systems, or at least better in-house audio-system, so that there is more audio aspects one can control.
There is no 11Labs integration for example. And what you can do with voices so far is pretty weak.
I've used it, they are very generous with free accounts, they give you daily videos. The issue is that the resolution it generates at doesn't work too well with people from a distance when you scale up the footage, it sort of produces undesirable results.
But, it is the only tool I know of which allows you to generate new footage daily for free, which is awesome.
ITS AMAZING APPRECIATE IT...
Veo 3.1 still does impress. With 3.1 I can (more easily) for the first time make super long videos. I'm currently working on a video animation that will be more than 6 minutes long. I will do a complete rework of the audio and splice it together in Capcut.
Even if I won't be using the audio as-is, it is super helpful to have the audio there, for voice lip-sync purposes for example. I can remodel the voices generated by Veo using 11Labs, and keep the pacing, and just change the voice itself.
Thank you very much 🙏🏻
Indeed, it's useful to have an audio reference, but I agree, most of the audio it produces can't be used for anything "serious", mainly due to continuation issues.
Once Google fixes some of the frame-bugs for blocks (beginning and end-frames), I think smooth transitions will be easier to get. There's currently a bug with the "extend" function operating from a different frame than the actual final frame of the previous block.
You can easily spot this bug yourself by just "selecting" or sliding to the "final" frame of a block, and then press "spacebar" - theres more video after that frame, even if just a few hundred miliseconds.
Oh, I have a solution for that, not sure if you're aware of it, but all you need to do is to cut the first 2 seconds or so from the begining of any "continued" clip and it should fit right in.
Ah, what is the best way to accurately "select" or "trim" those 2 seconds?
I found this solution myself, but it rarely works very well, because of the imprecision of the timeline slider.
What I think it does (which is very smart) is it repeats the first 2 seconds or so of the ending of the previous clip in order to ensure stability for the continuation. Grok for example has exactly this issue, which is the first few frames of any clip suffer from what I call "calibration" issues. I hope they will fix this.
Yeah I'm sure that is a good thing to do - but only behind the scenes. 🙂
Not for us, the end users.
I did it manually haha, just slowly cut from the ending of the previous clip and play the footage, repeat the process until the continuation is smooth.
I think I noticed that too, but at least now you know what to do when it happens.
yeah I kinda did, but I just keep generating until it gets a smooth one. The problem is, sometimes the generation I get is so good I don't want to scrap it.
so I will be forced to do that. But I didn't know it's as much as 2 seconds into the next clip. That sounds like a lot.
Doesn't that mean you only get 6 seconds of "new material" for each extension clip?
Oh no no, I wouldn't do that, you're wasting credits. No need to do that because the footage should fit, it's just that it takes a bit of work on your side.
You have unlimited generation ? 👀
It doesn't do anything better
yeah, Fast generations unlimited
it's my hobby, so I am only wasting money. Still 50% off for the first 2 months. I'm still on month 1.
we'll see afterwards
it is quite expensive
Oh wow. If I was able to afford that maybe I would've been able to continue my film lol.
By the way! One more thing! I do not recommend using upscale when you download your videos! It sort of smudges the footage!
ah yeah ok. I haven't used that much. 🙂
good to know.
I usually download the whole footage. Like the first part of my animation is 2 minutes 5 seconds long. There is no option to upscale the whole thing. Only per clip.
I will find another tool to upscale
I mean, I don't know if it does that every single time you upscale something but I've seen some footage which looked worse after upscaling.
One thing at a time, first just generating it, and get it right. 🙂 It's a claymation in the style of Norwegian children's TV from the 1970s. In this animation, they speak in "trønder" dialect (Trondheim), but it's for grown-ups. The main character is sitting inside there's thunder and lightning outside - he says to his cat "Well Albert, there's not much to do about the weather" - and then he drinks the coffee cup - except it's not coffe, it's the trønder-drink known as "karsk" (moonshine in coffee). And his hair turns into a gaudy slick hair-do. Then a mailman comes knocking on the door to deliver a letter, they harangue each other, about the hairdo and the mailman's moustache. Then he shuts the door, opens the letter which contains a stamp with a smiley-face on... and written next to the stamp is "lick here" - which he does. Then the world turns upside down as he's thrown into a wormhole into another world. Now inside the belly of a giant snake.... and that's where I am now. 😛
only funny for people from Norway really
but amazingly, Veo can do the dialect.
partially..... sometimes it turns into Danish and stuff.
Well, midjourney for the reference characters and stuff.
which I then feed into Veo
mail-man
Why do my comments keep getting deleted?
Hans-Bjarte after he drank the cup
"ingredients to frame" has been extremely useful for the change-ups in the scenes
I love ingredients but wish they would incorporate that with 9:16 as well
even if only 3, I can still do a lot with 3. For example, when the mailman is holding the letter, the mailman is object 1, the letter is object 2... but once you have an image with them together with it (you can use Gemini Nanobanana for this) - you can turn those 2 objects into 1 image.
so a good tip is to join objects together, for each scene. And in one scene I only needed the character's arm/hand to show (as it picks up the letter) - I used Nanobanana to erase everything other than the arm, and used the arm as its own object as ingredient
Veo's ingredients to frame solution
that way you can avoid inadvertent scenes like - if I had used the whole body of the character with the arm, I could risk him showing up in the scene.
and makes it a lot easier to prompt
I didn't know that it had issues with 9:16, that sucks.
Absolutely can't be done yet
I have no use for 9:16 yet at least.
Maybe if I was making something like... youtube shorts?
Making lots of shorts myself so it would be useful
ah ok
what kinds do you make?
there's a lot of fake cute animal shorts out there 😄
Historically correct ones lol
historically correct in what way?
Anything history, presidents, civil war, billy the kid, ect
Ah cool
I have done some Roman history stuff, I'm a history nerd for ancient history 🙂
Keep me going
I made Marcus Crassus in Midjourney based on his statue. And then I made him talk Latin in Veo. That impressed me.
Just finished one on the Sphinx
Beautiful. I personally use something called Story Planner for putting ideas together and planning. You don't need this of course, a simple notepad could suffice, this just makes it a bit easier to organize things. @remote crypt
I can translate Egyptian hieroglyphs (for real) - it's one of my hobbies 🙂 I got Gardiner's 3rd Edition dictionary here
Thats awesome
yeah I use Notepad it looks like crap but it works 😄
Hey, if it works it works.
I use chatgbt and save to Google docs
Haha
The palatalization on some letters is wrong though, this is not proper Classical Latin
but not eccesiastical either, so it's weird
Veo does a LOT that impresses me, even if there are really annoying quirks.
even 11Labs - dedicated to voices and sounds - can't prompt a trønder-dialect on the fly. Veo can.
Yes I agree I'm hooked.. I was hooked on Hedra, not anymore. Veo 3 for everything now
😄 3.1
Yup
I wonder how long the dev cycles are, I can't wait for the next update, or at least some hotfixes would be nice.
Faster and faster now
Also, Google devs respond to our bug reports in here. So be sure to report any you find, but first check if it hasn't been reported already.
Will do, nice to know
I got my "Doors are not being opened properly...." bug report there 😄 And they didn't find any issues with doors, so I had to prove it with videos. They are looking into it now.
it's like 99% bad with doors
Oh wow
just test it yourself, make a door image in another app, use it as frame 1
then say "person comes out of the door then closes it behind them"
Now that you mention that I had to do about 50 clips to get a passable door opening and closing
the door may rotate at the wrong axis (rotates around where the door handle is) - and when it closes, the door handle will have been teleported to the other side
yep that sounds about riight. Worthy of reporting, you should join my report then with your findings.
I was doing a video on Napolean had trouble with that
I tagged you in it
Ok great
you can see my test videos there too
I haven't tried generating a wholecloth (with no source images) video involving doors, perhaps Veo does well with doors generated by Veo absent images?
And that is why Veo devs couldn't reproduce it?
I mean, I never make Veo videos without source images. Unless I'm making something less serious.
I really like the new Image editor function in Veo though. It's a lot more powerful than Nanobanana in that it includes some handy sketch-tools you can use to mark things.
It is Nanobanana, just with extra tools.
And you don't get the stupid Gemini AI watermark at the bottom right.
I see hasn't tried that yet.. but I'm sure I will soon enough!
like, you can circle an object with red, and if you want to move it to a specific place, mark that place with blue
One of the issues I told you about with Grok, if you look at his face you can see what I meant earlier. I wonder if Veo can get rid of that (I doubt it can without a reference ?)
I am still on my Pro trial but I can't generate until a few hours later. They give you like 3 free videos a day, I'll probably try to see if it can fix it a bit later. Yesterday it consumed my third one without generating anything and it also deleted the previous 2 generations for some reason. @remote crypt
Yeah, it looks bad.
I mean, it looks okay if you don't scale the image at all.
But I do have to scale it to fit with the rest of the footage so yeah, it is an issue.
I mean, it's already scaled at the correct resolution. I just mean if you don't click on it to see it at full size it looks okay.
btw there's a livestream on Discord that started 8 minutes ago, about prompting in Gemini, Veo etc...
check Announcements channel for the join link, it's in a different server, the Gemini server
3 free videos a day isn't much when you consider I go through about 300 to 400 to get about 2 minutes worth for my shorts. Takes me about 10 times or more to get 1 good clip
I will never use Nanobanana to create new images though, Midjourney is far superior in terms of power and control over aesthetics. But, for AI-editing, Nanobanana is superior.
Nana b is great to edit with only
yep
MJ has been working on their product with a very strong emphasis on aesthetics from the very start. I was with them from the beginning, june 2022.
I'm with Leonardo, but very dismayed hardly use it now
Indeed, that's why I have to rely on Grok. Even with them I still can end up not getting anything I can use for days at a time. And when I do, well, sometimes it has issues like the ones I showed earlier. Unfortunately that one I don't think it can be solved so I might have to use the footage as is.
Though it does hurt the impact of the footage if it looks that way, It sort of breaks the immersion.
Leonardo is basically a pimped up Stable Diffusion.
Yup, was good at first years ago, now it's a memory
I was in the early alpha of Stable Diffusion also, in August 2022. When we were only 1200 users. The CEO Emad Mostaque was in the discord too, and I chatted with him.
SD hasn't done much lately, still stuck at 3.5 XL etc..
But I remember when SD did come out, it was better than MJ - for a time.
Hedra was the bomb for lip sync, still with them. I'm a mod on the discord
ah Hedra never tried, but I have heard of it.
I haven't tried much lip sync yet, but I should learn that.
Google flow veo3.1 does great lip sync, voice consistency not bad, but not great. I prompt for a familiar voice. Tends to work
I wonder how well third-party post-work of lip-sync would work on claymation figures. Different kinds of mouths.
like Kling, Hedra maybe?
Only good thing about Hedra still is up to 5 minutes lip sync.. good for songs or Narrations
I made a funny video of a rude Irish concertina-player on a pirate ship. If anyone wants to see 😄
Ohhh
flow-gallery
Yup
Ok
If I wanted to get around that limitation I could simply use close-ups instead, but is that something I want to do ? No.
Can’t wait for nano 2
what will be new in Nano 2? other than "better" 😄
Demis Hassabis is such a genius. It's awesome he won a Nobel prize for his work at Google Deepmind, for the 200 million proteins folding with Alphafold 2.
I grew up on his games, when he was a 17 year old AI coder for 90s DOS-games like Theme Park, Dungeon Keeper.
You sometimes have to find creative solutions to get the footage you want, that's dedication.
Yup
4k res and much better prompt adherence
No need to use an additional upscaler
Awesome.
Veo 3.1 fast is free with professional google flow subscription..
Seriously ? Let me try lol, I'm pretty sure it won't let me continue generating footage.
That's how I'm generating 500 clips daily
I am wanting to create a 30 second ad with a spokeperson talking but I am unsure the best way to go in order to create the video longer than 30 seconds.
Hmm, I tried and it doesn't let me generate more 😢
Ultra subscription
Do you ahve to go all the way to ultra to make a video 30 seconds?
how do I get it to use the same person for each frame?
Copy your last frame and start new clip with that
Gotcha
Or you can use Flow which is part of the Pro subscription and it will give you the option to expand the last generated frame.
Yes
That way is better actually because you get actual continuity, rather than just starting from a frame.
Nice chatting off to upload a video!
Take care!
Hi
@storm haven I'm in Flow but the subscription it is offering is the ultra
It doesn't let you use it ?
I was able to use it on Pro, so it should be working for you as well.
It states I ahve 40 credits ATM so if I hit the plus above the frame in the scene builder it will allos me to add to that fram?
@storm haven
Yes, but use veo 3 fast so that you can generate 2 clips instead of just one. Also auch, that's only 2 clips you can generate.
Keep in mind when you consume those you will not get additional ones until the next month.
Make sure to use these settings
Also the Plus you have to click is this one @rugged root
The other plus just saves that frame as an asset which you can use later.
Making an attempt at it now
How many credits do you get if you go to ultra?
25k seems to be the number but not 100% that they apply to this type of generation..
The other guys said you get unlimited generations if you use veo 3.1 fast.
Oh ok..
As in, it will not consume from your credits.
Ya this last add used 20 credits
Yeah, that sounds about right.
I don't think it can go lower than that.
You get 3 additional videos per day but not in Flow, for that you need to use Gemini.
And in gemini you can upload an image and have it make a video out of it based on your description. You don't get the continuation feature which Flow has.
There's also ten or so gens a day via Google Movies
Oh yeah but I don't think that works with images, I think I've tried it.
It sucks to lose the credits to bad text generation
I have unlimited fast generations with ultra, about 25K credits for quality. But qualiity is mostly worthless in my experience.
It does now, there was an update to allow image uploading, which is awesome because there's a character limit there for the prompt
There's a new Image editing tool in Flow yep
It's NanoB but with extra tools
better than regular Gemini NanoB - because it doesn't add the stupid star watermark at the bottom right
I've had to use MS Paint to remove the watermark 😄
It would've been nice if you were able to share that with someone else lol. Too bad they don't have a teams feature.
yeah
I'm sure I've wasted a lot of GPU-hours on my many failed generations
Once prompt-understanding and image-understanding, geometry-understanding, etc... when ALL of that improves, hopefully we will need fewer generations to make something approximating what we WANT. 😄
We are the early birds still, kinda.
When Midjourney was in its infancy in June of 2022, its image generations were awful. I would often have to "Vary image" a hundred times before it even approximated what I wanted.
Does flow works in Spain?
I am not 100% sure, but I think so. I only got access to Flow in Norway this summer.
I assume it got rolled out to Europe in general all at once.
I've been to Spain 🙂 beautiful country
i have been trying to enter even with others accounts, but it says me that i am not enought age, my account age is more than 18 so o think it's by the country
tysm
well, I can't help with that. Maybe someone from Google.
But you shouldn't try with someone else's Google account anyway.
A Google account is kinda sacrosanct these days, like your personal ID.
i know, i created another account for it, don't worry
ok
There's alot of useful things which they could implement indeed, such as a way to store characters as assets which can later be used consistently throghout your scenes, one level further would be to also be able to store their clothing.
And most importantly Locations as an asset (though I suppose this would be much more difficult).
Perhaps have the AI generate a whole scene based on a description, let's say a park. Then, it would store that as a "memory" asset, which would store the way the whole park looks at 360 degrees. Think of it like rendering a full location, then simply "reading" from that memory whenever the scene uses that location.
This way there should be far less inconsistencies when it comes to locations with things such as camera turning back to something or someone, and then changing places.
For now, I just want the basics to be good. Timeline-slider is absolute...... !"#¤
I want them to prioritize right 😄
Of course, basics first, the foundation.
I'm currently trying to animate the bile of slime that comes from within the belly of a giant snake. I am having trouble getting the flood to carry with it the 2 objects in the scene.
parts of skeletons and trash, and then it is meant to push the claymation doll and an amcar towards the camera's foreground
Have you managed to get the objects into the frame/slime ?
yeah, they are there, it's been a smooth transition to this point
a "meat-door opens behind, and an explosion of transparent beige slime with parts of skeletons and trash gushes forth into a river that fills the cave flowing towards the camera..." etc
It hallucinates clones of the car and the character too sometimes.
and sometimes, "meat-door" generates some stuff that I didn't ask for 😅 oh well
Okay then I would I would try is to do is:
-
Let it generate what happens after the slime comes you, let's say 6 more seconds of footage (or as long as you want the slime to flow).
-
Then pick a frame from the last clip it generated which you like, get it inside something like Photopea, remove the objects from wherever they are and move them where you want them to be.
-
Remove the clip(s) you generated until you get to the clip which shows the bile coming out. Pick the last frame which is good.
-
Import that frame which is good as the first frame, and then the frame you edited with the objects where they need to be at the end as the last frame.
-
Generate the clip, it might work out. The downside is you'll end up with a shorter version (if you wanted more clips than just one) but if this first clip is actually good, then you might be able to pick a frame midway, repeat the process, and this way you'd get additional footage which should have sort of the physics or at least the positions of the objects you're looking for.
maybe I should try "esophagus"
this involves trimming the blocks as well right?
Yes, especially if you do your editing inside Flow directly.
I will edit everything in Capcut after, but for the continuous shots, I can't do that in Capcut very easily.
It doesn't allow you to trim footage ? Like that's a very basic thing.
Otherwise you can send them to me and I can take care of that for you, I'll send them back to you with continuity.
You've been nice to talk to so I don't mind helping you out if you do need help with this 😁
oh I mean, of course I can do that 😄
just for the general final edit, will be done in Capcut
like, splicing the videos together, etc.
extract and replace audio
and maybe some other edits too
Yeah, and I'm pretty sure you shouldn't have continuity issues as long as Capcut allows you to trim out the clips.
Flow need a upgrade
just "a"? 😄
Yes i have a lot, but it doesnt show anything - whats wrong?
Am I the only one who noticed that after this update, the more videos you make, the more mistakes the AI starts making?
And the "n" an upgrade... but ?
Grammatical Articles, a, an and the.
is flow down right now because it's not generating images or videos <@&1173646484311384114> <@&1366831844502274180>
Hello, unfortunately this is not a moderator issue in the future please ping us if you see something that goes against server #rules-and-guidelines・📋 Thank you!
I see, sorry first time posting in here on something like this
No its totally okay! however since you are experiencing issues with Flow you can leave a detailed report in #1319424187730755636 channel with as much context as possible. I t helps to keep the team organised and it’s also where all the technical issues are logged!
Our #server-directory・☎️ has all the information about our different channels so if you ever need guidance on where to start, there would be a good starting point.
Hey! Does flow always use a Veo watermark for downloaded videos or am I missing something? I'm on pro
Hi google labs team,
I have a few video-generation related modifications to request. I’m looking for a physiotherapy-focused video generation tool similar to Hyperhuman.cc. Could you help adapt or develop something like this for our marketing campaign?
Only Ultra Plan generates video without watermark
why google took so long to integrate gemini into its searches whilst brave and bing were doing it 2 years ago
cuz this is Flow
Really? Wow
Hi. I logged in yesterday and got the message about the latest update and additional credits being granted. It's supposed to be 500 credits but up until now I did not receive any. Am I missing something ?
Is flow down now?
I have the same problem
me too
It's working for now
i got this too
they really be adding quota even on the page itself?
😭
Maybe there’s an incoming of an update
it better be huge one
this is probably the first time that labs google page got hit with... quota limit
Looks like something's gone haywire with their rate limiting. Since Veo 3.1 is free for ultra subscribers, rate limiting is definitely important to have in place (I'm a software engineer, so I can tell you, they need rate limiting in place), but looks like something's gone wrong with it.
I find the ingredients to Video section along with the image generation section, altogether a bit confusing
I think this interface is the best for image generation and editing but this only appears while I'm in the editing section. But not in the generation section
Yes I can’t access it either wow
Never mind
if you can get it if you keep trying .... their rate limiting is broken at the moment
well i can understand if they do rate limit on video generation but, this is the whole page of labs.google
yeah its back up again
..... which is definitely not good
???
does anybody know if there's a way to download all videos in one operation like you can with whisk?
got the same error message
Wondering if they're hitting hard limits from the Google cloud platform (Assuming this is all running under that cloud provider since it's all google lol).
not as familiar with google's cloud platform as I am AWS (I'm WAY too familar with AWS as my software runs on it)
same problem
oh I thought it was just me getting this.
Hi, sorry for these weird errors, our best guess is there was a large-scale denial of service attack, which caused that issue. We believe it's better now, please let me know if you're still seeing it.
😂 no, not intentionally adding quota just to load the site now. I think we can afford that part without any credit system or rate limiting.
Attack?
The Google staff probably after finding out about the service attack :
But yes thank you for keeping us up to date @thick oriole !
Yes we had a fun hour tracking this down and fixing it. We're honoured that labs.google/fx is interesting enough that people are really really trying to abuse it in various ways 😆
Who in their right mind would do such a thing ..
We do have a bunch of improvements coming out in this area, which will stop the bad/abusive traffic from our end, and then hopefully help us to have a better experience for you all ('good users'). e.g. being less strict with our rate-limiting controls.
We got called good users! That's basically headpats! 
does that mean you guys already chose whos the bad users
jk
speaking of bad users
<@&1366831844502274180> this man right here
Someone out there writing a script that hammers away at the service, image generation and video generation to infinity, eh?
This is great news! I assume you're adding controls (If you don't already have in place) to shut down users, IP addresses, etc. that exhibit abusive request patterns?
I missed out on all the DDOS'ing earlier. How is everyone's weekend going?
Veo working fine here thankfully.
we are doing well now 🫡 how about you?
Can't complain 🙂 Outside though? Weather is awful. Sleety snow, wind. Hey, it's Northern Norway.
And tomorrow? +4 to +6 celsius and rain. But I'll stay in doors.
at least the snow will melt away
it's so rainy here in CA too! no snow though 🙂↕️
😄 yeah snow in California is probably rare. My brother studied at Monterey IIS
and my relative - ex police chief of Orange County lived there - he passed away two days ago.
weren't close, but met him 3 times
But hey, rainy in California is a good thing!
sorry for your loss 🙏
thank you
Are you at the Google HQ building?
or is it like dispersed among multiples in like a campus
Google has quite a few offices across the country!
of course 🙂 But the main one I know would be California, since it all started over there.
Silicon Valley
difficult for me to grasp the quantity of people - I live in a town of about 5000 😄
Hi everyone, I have a question, is there any way to keep "the same voice" on each video generated with VEO 3....?
Two techniques I use to get as close as possible: Either 1. Use extend from scenebuilder. This will only work if the voice you want to keep using is in the last second of the clip you're extending. or 2. Use something like AI Studio to provide an extremely detailed description of a voice sample you provide and then use that text description in your prompts
🤔 I see.....interesting, I´ll try.
thanks a lot 🙌
Let me know how it goes for you!
unfortunately no. I am having the same issue. I think my solution wiill be to use 11Labs, and simply stripping away all the audio, and replacing it with custom made audio, and voices.
But it really helps having the base audio of Veo - as it has already done the lip-sync for you. You can then generate a different voice following the exact same pacing of syllables of that Veo audio.
Why does it keep refusing my requests ?
How exactly is it breaking the policy, I'm confused.
Yeah, I even tried without a prompt, same thing happened.
I tried with a different angle too, same thing.
maybe it suspects it's a real person.
He doesn't actually exist, I generated him.
yeah, but Veo probably can't know that, right?
it sees a photoreal human, maybe that does it? idk
Yeah, maybe that's what it is. It might think it's a real person.
unless Veo has access to all the faces of every 8 billion human on Earth, it can't know for sure.
which - obviously - it doesn't, and shouldn't.
Anyone tried the new 3 references to video as yet
Apparently you can now do it in Gemini app
yeah, but useless in Gemini Nanobanana.
Only useful in Veo
Well, that is a bit harsh of me. What I mean is that the "ingredients" can't be used in a "boolean" way to subtract or replace parts from image 1 to image 2.
Oh and you can’t use portrait mode either with it
I'm sure you can use Gemini NanoB to generate an image with 3 ingredients. But, I have Midjourney...
meh, I can wait 😛 It will take a lot of heavy lifting to get NanoB 2 to get even close to MJ's functionality and artistic power.
But prompt understanding and editing? Yeah, NanoB wins that hands down.
NanoB for me is only as good as its ability to understand editing prompts. That's it.
All of the UX work to be done, to give more control to the user, all of the aesthetics-training to be done, and that has never been an emphasis from Google on this.
Maybe by NanoB 3
depends on the development priorities of the team
if they keep down-prioritize the UX & ControlNet functions, like most do... I don't see me using it any time soon.
What's up with VEO today? Has he decided to outdo Runway 4 in terms of the amount of crap generated in one scene?
no its just you
This is just crazy.
A typical prompt like (to simplify) “the hero stands still and speaks, no other sounds, no music” turns into constant background music and a constantly walking character fifty times over...
P.S. And don't tell me how to write prompts, I know how.
thats a bad prompt tho
but since u know how to write prompts then i wont comment any further
go touch grass lol
weirdo
I was testing and I got something useful
you can add something like this to your prompt:
"Voice: Female, slightly deep contralto tone, warm and confident, Spanish Latin neutral accent, tempo 0.95x, natural breathing and expressive pauses."
request ChatGPT or any other tool, to prove you with this kind of details to keep a very consistent voice
Will be great if Google adds the capability of using our own voice recordings with videos.
Not in UX and creative control. I've tried them all since the beginning. Prompt understanding is a different matter.
Omni reference, mood boards... Etc
Negative prompting usually never works. Prompting needs to be positive.
This lol
I was Test the prompts with perfect Details with ingredients that's Flow make the Video amazing
I mean it makes sense why negative prompting doesn't work, when you consider the process of diffusion. It starts with noise, and gradually whittles the noise away towards something. Not away from something.
Accidentally bought 20k credits before realizing I could’ve just subscribed to AI ultra for $50 more, is there anyone around that could help me make that swap by chance? If it’s possible, if not, it’s just a dumb mistake on my part 
Hello, a friendly reminder we cannot generate images or videos in the server! If you are looking for help or guidance please visit the #server-directory・☎️ and use the appropriate channel to do so. Thank you!
hi. why i can't save and download project video made using flow into my laptop? anyone can pleae help
hi guys
Hi guys where do I report Nano Banana issues with flow. I keep on getting this response. could not generate image - try again later. Is it my prompt or is it a known bug. Appologies if I am a noob and posted in wrong place
hey guys when i use ingredints feature its not letting me create portait video how to fix that
Have they added credits to the ultra subscription now??
or is Veo 4 coming? It says that Veo 3.1 is not available atm
yes its coming
in 2026
emil veo 3.1 litereally released like months ago why would they release veo 4 right away
idk, but obviously they're doing some update as 3.1 is currently not available
its available for me
where does it say this?
same
lol
Is it dead? Trying to put an image and saying it violtes the policy
hi everyone
Yoo
I'm so glad you got something useful out of it! Yeah, using an LLM to provide rich details on the voice (either from scratch or from audio you upload) is a good stop-gap. We'll be making voice consistency a lot better in the coming months.
Definitely need help with transitions and audio if anyone can help ❤️
That’s the first full min video I’ve made with flow and CapCut
Definitely need help with transitions
Forwarded your video to #flow-videos・📹 , the channel for posting your creations. Thank you.
Hi! Is this the right channel for asking general questions about Flow?
Yes, this is for general flow questions and there is #flow-tips-n-tricks・✏️ for prompt help and tips.
hey guys when i use ingredints feature its not letting me create portait video how to fix that
i want to make drive thru mcdonald video. using veo 3.1 quality version but car going backward it has glitch I lost 100 credit again. overall video quality is best but prompt accuracy still not better than sora 2.
Yo guys, is there a way to download receipts for Flow credits? I have topped up twice but don't seem to have any record of it in my account for billing purposes
thank you for Flow! Our clients love these video's
Hello @hardy hedge , a friendly reminder videos and images are not generated within the server, you'll need to head to https://labs.google/, pick a tool, and generate it there. Thank you.
Hello, as a friendly reminder we cannot generate images or videos in the server! Please do not post your prompts in the channels as it will not generate. If asking for advice or tips on your prompts do so in the appropriate channels.
When nano banana pro...
Hi, how can i take control of the camera movement? i want tho draw the movement, but i can not find the option
Hello~
This is a question that arose from taking sales information from Google spreadsheet and performing monthly comparisons.
-
Is there a way to make a workflow or created app run at a specific time, schedule setting?
-
Overall, I was very satisfied to see that they used no-coding to design, design, and even create an app.
However, is there a way for this workflow to run (results) faster? -
Every time I run it using this organized workflow, the design changes, like insights, is there a way to fix it and print it out?
sir this is a channel for ai video generation
oh... sorry man! thx noticed that
#flow-general・🌊 I’d like to share my idea: there should be a dedicated channel in Flow where users can share of idea for advertise products such as perfumes, clothing, kids toys, industrial materials, and more. Do you think having a separate channel for business ideas or product promotions would benefit the community?
lmao no
ok
man, not sure i should file that as bug/feedback, but the camera really has a huge bias towards frontal view and slow zoom in, despite what you prompt
also working on a wuxia inspired tai chi scene (inspired cause i play Where the winds meet right now) and damn, the AI craps itself with a lot of particles (cherry blossom petal and water drop swirls) and mangles faces that are a bit farther away
Hello @real gazelle please submit any feedback/ideas in the #1238621511401734254 . Thank you.
did you resolev this?
Is there some specific phrase you have ot use in order to get it to use the same person in a video when you are trying to add more to the video? I have given it additional script for it to use and it just keeps reusing the script from the first prompt
Hiii! I've been using flow a LOT recently. I've noted down some of my thoughts. Really wanted to share them here!
- I would love it if the ingredients won't disappear or get un-selected after hitting generate. I use ingredients to video feature mostly when I'm generating a story board so more or less in each prompt, the ingredients are the same. Right now I need to select the ingredients each time for every individual clip. As a workaround I use the "re-use prompt" button so that the ingredients are auto selected from the previous prompt!
- Confusing interface when putting the image for ingredients to videos. I wish we were given the option to choose if we want to edit the image before confirming them as an ingredient.
- Aspect ratio: direct extension or cropping of the current image into the required aspect ratio image using nano banana. If possible, make it accept images of all aspect ratios without needing to crop them.
- Ability to select multiple ingredients from the menu at once.
You can just click on reuse prompt tho?
Yes you are given choice , if you hover your mouse over the image the edit button appears
I am noticing that when I do extend a scene the pitch of the speaker changes to where it loses all of the low and mid end of the audio spectrum
Hi guys, Flow is performing poorly here regarding scene creation. Every time I try to export the entire video, which is about 8 minutes long, it always gives an export error. Is there anything I can do? Thanks for the help.
Man I wish they would update Ingredients to allow 9:16 generation. It would make shorts so much easier.
And in quality mode
Yeah I honestly wonder why ingredients are not currently compatible with quality mode. Hopefully the feature will be in the next few updates
Does the Ultra plan get rid of the watermark? Also any talk of when we can use our own voices?
I'm completely hooked on this Tech, awesome job devs!
i dont see some project in flow and cannot create some project, wth google lab
please help me
are the 3 generations a day in gemini the fast model or quality model?
https://www.cloudflarestatus.com/ maybe because of this ?
Welcome to Cloudflare's home for real-time and historical data on system performance.
for me nano banana does not work - both aistudio, google flow (ultra plan)
Hi guys, Flow is performing poorly here regarding SceneBuilder. Every time I try to export the entire video, which is about 8 minutes long, it always gives an export error. Is there anything I can do? Thanks for the help.
Nano banana is still 1k on google studio…
this isnt flow related, but i saw something interesting on aistudio app section 😂 Google answers to Sora 2 Cameo feature
well technically Flow can do "CAMEO" features too,
aka ingredients to video 😅 well except flow has more customization control i guess
This is cool I don’t see a sign of nano banana pro yet
same
Hmmmm
i dont htink they announced anything about gemini pro 3 either
probably later today
Yeah I’m thinking that too
oh they updated the gemini page btw
I’d like the Banana Pro.
🍌
how to use flow in free
how to make ugc video 1 minute for free ? credits are too expensive
Audio is all messed up on Flow right now (Fast and Quality) - distortion, pops, clicks.
whats this about ultra having to pay 10 credits for veo 3 fast now? just got a popup when I loaded the site
Really, I hope that's not true
3.1 fast is still letting me gen for 0 credits, wish I properly read the popup, only noticed the "10 credits" part as I was closing it
I guess its only during peak hours?
I appreciate that. I'll keep track of peak hours
I need to monitor my credits as well
it's not a very clear announcement, i hope that's what it means, but it could just mean cost increase is coming and isn't implemented yet. idk
I'm in florida, so I guess I'm safe after 10am to 8 pm
Correct, I don't know either
It sounds exactly what it stated, during peak hours, no free fast..
for 13 hours of the day lol, I mean they're off-peak hours for north america but still, they seriously need to update the wording there
guess we'll find out in 3 hours if credits suddenly start being consumed at 8pm pacific
Haha, yeah it's not a very professional announcement..chatGBT could have done better
also just to put that into perspective, 300 * 30 = 9,000, 9,000 * 8s = 72,000s. at api price, 72,000 / $0.15 = $10,800 on the api
chill my dawg lol
er 72,000 * 15 w/e
Wow
dont feel bad, I've generated that amount before, it's just wild to think how much we really get as high throughput users on unlimited plans, seems totally unsustainable
Well, it takes me 10 clips to get 1 I like.. sometimes 20 for 1 eight second.. I'm a perfectionist
I hear you. especially when the platform is set up to batch videos with huge concurrency, spitting out 20 in a minute, it feels made for it
Spend all day on 1 minute of clips
Sometimes it flows and I get 2 minutes a day! Yes!
Always changing the prompts
But I believe it's the best so far
it's also genuinely just kinda addictive, like a bizarre video slot machine
I enjoy taking a thought into form. Making it real.. it's super addicting!!! I can't stop!
But then AI will replace us soon maybe. Haha.
I use chatGbt for my scripts, then I get a real image of subject and AI it... so it won't get flagged, then Google flow veo
Did fast 3.1 get 10 credits?
I got a pop up that says its now 10 credits during peak hours for ultra users.
Ahhhh it was only during peak gotcha
sorry to see fast credits are not free anymore. I hope they won't raise the price to 100 credits next.
that's too bad. The decision was to inspire creativity (and get more subs...) when they made it free for Ultra. Usually it's an upward spiral 🙁 10, 20, 30, 50. It's not a good move. More people should flag this.
Sad thing is the next step is 20, 30, 100,
hi
The reality is that in veo3.1 I find myself having to run many generations to get what I want especially across consistency. when it happens it's magical. a pity that it^s costing now 40 credits per generation
It stills free for me
Probably was 10 during peak hours
so is it just during "peak hours" . then they are gently rolling this back in.
It’s 0 credit for me
Hello
Why would they raise the fast model to be 100 credit per video , doesn’t make sense to me
When you go from a “Free model” to a paid model it’s not a good thing. “Free for ultra users paying 250/month”
Why would they hike it to 100 credit per vid lol , it’s a fast model , not the quality one
Plus they already said it’s 10 credit during peak hour , cuz recently bunch of goon- I mean ultra users probably generating vids like non stop , in span of 30 min or something
Hi, for those who have used Veo on both VertexAI and Gemini API, do you also encounter more prevalent prominent figure issue on VertexAI?
But then for those people they should throttle it. Why should every one have to pay for only some of the people?
Idk why u ask me lol im not a google staff
weirdo
Oh because you made an assumption. No need to start calling people names here.
Google should gift everyone a free month of Ultra for all the issues they have had lately, no excuses not to
@burnt gorge Hello, we wanted to remind you of our server guidelines, stated in #start-here・📍 : “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now.” Please keep in mind that non-English discussions and content may be removed. Thank you.
So we're actually in the official google discord server? 
"Ultra users will see cost rise from 0 to 10 credits" for Veo 3.1 .... I knew it was too good to be true! I knew it wouldn't last! You pay $250 a MONTH for the subscription and you STILL are charged credits for video generation!
Exactly!! We pay $250 a month and STILL have to pay credits on video generation! That switched the subscription from being worth it, to no longer worth it!
I think we need to watch what happens next. It's a great model but it does take many generations to get what you want. Often the wrong character is speaking or the voice is off. etc. so those credits get used up fast
Am I reading that peak hours timeline right? They are in terms of pacific time which in eastern time that makes it through the middle of the night
it looks like from 8 pm new york time to 9 am LA time...for now.
so in terms of eastern time zone that's 8pm to 12pm the next day. It's weird that peak time is through the middle of the night. My guess is other countries are hitting the platform harder than the US.
they can easily solve the problem by making sure their rate limiting works ... it's not a hard problem to solve (I know as I'm a software engineer that designs and develops software that runs on cloud platforms)
instead they just bump up the cost 🙄
any news about nano banana 2?
They need to implement rate limiting per account .... if they haven't done that, then that's not a good way to roll out a cloud based service. One account should not affect the entire platform. If they implement rate limiting per account then they wouldn't have to go through such drastic measures like hiking cost for ultra users who already pay $250 PER MONTH!
If it was a small start up i would understand. This is Google.
I’m not google. It’s a personal expense
It's not rocket science to implement per account rate limiting ... that's how I have my cloud based services implemented.
My VEO 3.1 Fast keeps failing when I try to create with ingredients. Is anyone else experiencing a similar issue?
Same here
so it's been peak hours all night and credits are still at 0, im guessing that means the change isn't tied to peak hours and it's upcoming permanent increase.
google really needs to say something, clarify this, I've already cancelled ai ultra over it
was a bit hyped for gemini 3 but I wouldn't have been if I knew this would be the cost lol
i have them free
Maybe some upgrades going on -- havent been able to create much since last night.
The frames to video seems to work, but ingredients to video not so much. Perhaps there's a heavier lift on the back end for that latter operation. Again, don't want to sound like a broken record, but if they implemented rate limiting properly on a per account basis, then none of this stuff would be an issue
ingred to vid working fine for me:👍
did you get a lot of failed in ingrediants -- or everything normal?
normal
cool, thats a good sign. I got a lot of failed but maybe working itself out
we're moving out of the peak hours, so that makes sense
None of its working, ingredients, frame or text. I'm on ultra, too.. just hangs at 99% and fails. It's 2.42pm UK time, I'v been able to make 1 video since 8am this morn..
maybe try to log out, thats what helped my situation a little
I mean I'll try again for sure
ppl be acting surprised about bad service after logging in getting a banner saying 'warning: service is gonna suck, and it'll cost you now' lol
there's no workaround. wait until the magical hour of 9am pacific
looking forward to the magical hour -- little less distractions at night, but, hey can make it work
rate limiting ... rate limiting .... rate limiting .... per account ... per account ..... per account .... per account ..... just sayin' hoping google folks cruise through these messages
currently they must have the rate limiting per service 🤦🏻
you mean limiting to 4 at a time -- instead of people gunning in 6 windows?
i mean surely spamming here would help lol
meh its not really the end of the world, perfect time for me to take a break then until everything is good to go
i got all failed
yeah, it could be worse .... the peak time is through the middle of the night for me ... so no big deal from my perspective, until I see that cost increase (If that cost increase really happens outside of peak hours, then the ultra sub is no longer worth it)
Hi
Ingredients to video not worrking, please fix it
Hello Flow Discord crew! I wanted to clear up some confusion and apologize for the growing pains as we try to handle the current load. First, you can ignore the concept of peak hours; we were just trying to share when you could expect Flow to have the most traffic (and thus most possible disruption). There's no change in what things cost during peak hours. Second, for our Ultra subscribers: yes, Veo 3 fast gens are changing from 0 credits to 10 credits. Veo 3 quality generations are staying at 100 credits. Your Ultra subscription comes with 25,000 AI credits each month, which would be 2,500 Veo 3 fast generations if you spent them all on that model. We will restore some level of 0-credit fast gens for Ultra subscribers as soon as we can handle the load. Thank you so much for bearing with us on this!
ouch
Hey Elias, what kind of time frame are you expecting for us to be able to generate again? I've not been able to generate anything in the past 6-7 hours (2 clips came out, one without sound and one randomly appeared after saying it failed) despite trying again and again which is getting a bit concerning as using this for actual work projects?
" We will restore some level of 0-credit fast gens for Ultra subscribers as soon as we can handle the load" .... so that means the cost increase is temporary?
"We don’t know when. Could be tomorrow. Could be never."
that's how I read it
"We don't know when" is right, apologies - it's a function of demand and how much serving capacity we can get, but we're doing everything we can to deal with it as gracefully as possible.
thanks for clarifying
@wide ether My goal is to restore some level of 0-credit generations as soon as possible, but again, I'm not sure yet when we can safely handle the increased load that brings.
so I'm guessing there's a massive increase in new accounts ... that's the only thing that makes sense. Even with per account rate limiting, a massive/dynamic increase in the number of accounts would be hard to deal with.
I've expected it since the moment it went unlimited lol, it can never be truly unlimited, even as it is you have users pushing 300+ videos a day, hitting equivalent of $10-$20k worth of monthly api usage on a $250/mo sub. need per-account limits, or permanently spun up pool of gpus that is shared by all unlimited generations
Yeah, per account rate limiting really mitigates this, but if they have a very massive/sudden increase in accounts then that becomes a challenge, even with per account rate limiting.
is that why i keep seeing random ppl trying to "sell" flow acc with ultra plan /45k credit here
even when you consider the 2500 video a month cap, the economics still look dire for google lol, each of those 8 second fast videos with audio would be $1.20 via the api. 2500 * 1.2 = $3,000 a month in equivalent api usage for $250 sub
your speaking in terms of what they charge ... not actual cost from google's point of view. For all we know it actually costs google .0001 cents per 8 second video generation.
Thanks for the update. Appreciate it.
I agree, we don't know the actual price for them, I just assume api pricing is as competitive as they can afford to make it
I'm thinking the actual cost is very low ... i.e. high mark up on the charge ... so per account rate limiting should be fixed at google for this service and just enjoy reaping the revenue from the ultra subs
I speak from deep experience as a cloud software engineer ... but I work with AWS ... but I'm guessing the things are similar on the google cloud polatform
google did have per account limiting. more than 24 videos in a minute, more than x in given timeframe and you'd start seeing concurrency drop to one 4x batch. if they're in this position despite having those kinds of limits in place, the economics of unlimited can't be working out
im sure they already hav what u suggested in their roadmap plan or some sort lol
.... or there's a very sudden increase in accounts
all of which would be paying the full subscription price for unlimited because pro plan already paid 20 credits per 3.1 video
Well the cloud platform does or should have limits too (I know AWS has hard limits like that too) ... then perhaps they were hitting the hard limits
Damn… I subscribed to the Ultra plan 48 hours ago because it said it was unlimited, and I simply couldn’t create even 20 videos of 8 seconds. Now the “unlimited” suddenly stopped being unlimited and it’s charging credits. Absurd. If I were a heavy user making 300 videos per hour, okay! But 10 videos a day and being charged for something that was supposed to be unlimited is a joke… It seems like they decided to launch Gemini v3 and ran out of infrastructure. I thought Google prepared for launches, but apparently that’s not the case.
On top of that, the support is NON-EXISTENT.
Aside from the SceneBuilder, which doesn’t export videos longer than 2 minutes, it’s been incredibly frustrating to work with VEO/Flow.
Those who subscribed to Ultra now have to pay to generate ?
This could lead to a legal process here in Brazil.
Yes, during peak hours. 10 credits per video generated.
wait til yall find out nano banana 2 is credit based system in labs
ultra user here, fast mode is not unli anymore it looks sucks ill cancel my subs now
what is flow
its time to unsubscribe lol sora is better than them
I wouldn’t mind that tbh, once it’s much better
And gives 2-4k res I wouldn’t mind
What is the alternative? Runway? You try the unlimited?
Hi guys, new here. Nice to join the community. How do we publish our own app to Flow?
Like Salesforce is listed there
Runway really sucks if you dunno how to composite
What alternative?
Nothing is really unlimited bro
"t2v now costs 10 credits"
What the heck? I'm paying for an ULTRA subscription and you changed the modalities after we already paid the subscription?
10 credits is still 50% discount over pro account, and it's temporary. my guess is that gpus got suddenly reallocated to meet demand on gemini 3's launch and the flow guys panicked. it's phrased as a service disruption. it's not permanent, they've already said that if you scroll up a bit
Ive read that statement but still not sure what it means. its not clear if thats permanent or for how long. Also: " We will restore SOME level of 0-credit fast gens for Ultra subscribers as soon as we can handle the load." (some != all)
not u lacking critical thinking
clarified in DMs "we'll post an update soon about this, it was due to heavy load and will NOT be a permanent increase."
I’m gonna check in a couple of minutes but after I saw the update notice a couple I wasn’t being charged 10 points
oh nice
This is what I figure was the most common interpretation of the update statement
Thank you for sharing that info
Hello guys,
Why am I having this problem with Flow? My Google account settings show that I’m over 18 years old. I have Google Pro AI, and I used a US VPN when I created my Gmail and Google account. If anyone knows how to fix this problem, could you please share the solution or DM me? Thanks! 🙂
Hi, guys does anyone know what the process of publishing your own application to Flow looks like? Like Salesforce has done
I think it's fascinating that their peak hours are basically the time NA is offline. explains why sora 2 had such a limited rollout. by confining themselves to canada and usa they basically limited themselves to the equivalent of global off-peak hours while the build up infra
nano banana 2 is already up ?
Hello, does anyone know why today some pictures/videos are generated and others not? I nonstop got a message "No image created. Try again." And alway for the same images to video..
Given that the physics in VEO are already terrible, forcing players to replay scenes, it should be free even during peak times for ultra subscribers. It's everyone else who should be limited.
They just changed the credit for 0 for 3.1 veo to 10 credits now?
If the model was better with identifiyng who's talking, with correct lipsync, with correct and consistent voice, better with physics, better with text, without unneeded black bars on top and bottom, whitout wrong policy rules. We could do with 1 or 2 generations what we need to do now with 80 iteration to get something usable -> no heavy load. Veo 3 is good but could be great with some tweaking.
If sora ever gets a ingredient feature like veo, it would kill veo. Only reason why I like veo more is the ingredient feature. It makes for consistent characters
i so feel you. in 90% of my generations, speaking people are having a stroke. 80 generations for 1 successful generation is the low average
Same...just saw that ultra tier is paying credits now when we're generating 100s of videos for the perfect 8s lol
Hi Elias is it possible to receive a top of of credits as this change occured when I used most of my credits on quality gen
me when credit

@tough yacht can you tell how much time are required to slove this issue of credits for ultra uses?
Veo 3.1 Fast seems to no longer be Unlimited for ULTRA users, has there been an official notice on this?
I just saw that, there is popup : "Each generation uses 10 credits" ... 🙁
"We will restore some level of 0-credit fast gens for Ultra subscribers as soon as we can handle the load. Thank you so much for bearing with us on this!"
"We will restore some level of 0-credit fast gens for Ultra subscribers" Well that's it
it was fun while it lasted
did anyone else mention tho that ingredients to video doesnt work at all, just the default one is changed to 10 creds.
Time to move onto sora
How many gens do y'all make 💀
I'm done. Not subscribing any longer.
How ever many is needed to generate a coherent video with this trash ai
to be fair its only temporary and u have 10,000 vids worth of creds as an ultra member
u wouldve hated being an ultra member in august
25000/10=10,000?
my bad
the math aint mathing
I wasn't an ultra member in August because it did not provide me the service I wanted. Now I am an ultra member...and they provided me a service I wanted for like a week, then changed it to a service I did not want again.
yea that does suck tbh
what's the best site to access sora? It's on like a million sites.
the official one
I purchased the Ultra plan because Veo 3.1 Fast was unlimited, and I needed that for a work project. It allowed me to make multiple generations (I don’t run them simultaneously; I usually wait for the result so I can refine the prompt) until I get the desired outcome.
It’s bad news that 3.1 Fast now costs 10 credits, not to mention that “frames to video” was the feature I had planned to use instead of “ingredients to video”… But since “frames to video” now has issues with using the initial and final images as context—for example, character consistency, etc.—which is something I had already mentioned as a feedback
it depends on what you work
okay, post that here if you can
not when u consider the amount of gens it takes to get 1 good one, 2500 a month u can take 75% off that straight away that are wasted gens.
true
eh, it was worse before. 12,500 creds and each vid was 20 creds
but ig since u guys prob recently bought it, expecting 0 creds fast gens and then losing that does suck
it does suck but it was a worse offering before
We want to clear up some information regarding some recent updates to <@&1374455577375412234>.
- Peak hours doesn't affect what things cost
- For our Ultra subscribers: Veo 3 fast gens are changing from 0 credits to 10 credits. Veo 3 quality generations are staying at 100 credits.
Your Ultra subscription comes with 25,000 AI credits each month, which would be 2,500 Veo 3 fast generations if you spent them all on that model. We will restore some level of 0-credit fast gens for Ultra subscribers as soon as we can handle the load.
Thank you for continuing to share your feedback with us in #flow-general・🌊 and #1319424187730755636.
okay got it
damn, these rug pulls on pretty much all ai are devastating. I don't think they're gonna stop anytime soon either. Not until someone finally files a class action
imagine paying $250 for a service and then the next day they completely alter that service to a shell of what it was and you're stuck with a $250 bill
it does suck but its all temporary
First it was OpenAI, then it was Grok, and now it's flow
what did openai and grok do to you
It has nothing to do with whether you purchased it recently or had it before. The unlimited generation was something that justified the high price of the plan, and it’s also useful for projects with complexity. It wouldn’t be a problem if what it generated matched what you asked for on the second try, but that’s not the case.
In my situation, I don’t generate “easy” content because it’s for work and the results need to be professional, and that’s not something you get instantly you have to keep refining your prompt, even if you follow the “prompt guide.”
its time to leave the Flow and find some new tools. this totally Ridiculous.
yea but you're forgetting it's not a permanent change. it's prob to adjust to overload of gemini 3 temporarily
Surge pricing on AI credits you've already paid for is crazy
And they’re not considering that I’m in the middle of work, where if my client asks for changes, I have to deliver them as well, and that requires more generations… They should have kept priority for Ultra, then Pro, and so on.
No I think it's not a temporary go and notice the wording of the announcement. They said they will restore some level of 0-credit fast gens. What do they even mean by some level?
They just took my sub for ultra literally just before this was announced, so I'll give it another month and if they haven't reinstated the unlimited gens by next month, I'll be moving on...
prob 0 creds to ultra users on a roll out basis
same here!
There just giving themselves breathing space, the wording comes directly from a politicians mouth means they can go either way...
I got complaining with Gemini it always has the limited words when a sure new updates and I don't like it I want him to fix it
same
welp what r u guys gonna do about it. its 10 cred generations,
just pray u dont run out of 2500 vids before they give us back 0 creds
Lets pray. 😂
Are you ultra by any chance?
ive been ultra since june
Both introduced a specific service with a laid out subscription and credit plan, then the developers drastically overhauled that credit plan and generation quality not even a few weeks after they introduced said plan. So people got stuck paying hundreds for a service that changed almost immediately. It's like if some guy introduces a monthly subscription for unlimited organic lemonade. Then after you and your coworkers put your money into it, the maker decides to limit the amount of lemonade to 5 lemonades after a week and a week later he replaces the organic lemons with piss.
That’s what I do and I assure you that for the amount I have generated it is not enough because if the equivalent is 2,500 I am sure that I have already generated more than that
theres no way today u generated more than 2500 vids
As i already told you…
In my situation, I don’t generate “easy” content because it’s for work and the results need to be professional, and that’s not something you get instantly you have to keep refining your prompt, even if you follow the “prompt guide.”
When I send today? Since I renewed my plan I am guaranteeing you that I have more than that amount
My credits are for jobs like edit, camera movement, other edits functions
Oh well, not sure how legal that is in the EU, changing stuff like that AFTER the payment without having to agree to it first, but that means i will unsub until you restore the 0 credit fast gens again
What about already use quality mode before the announcement ? I mean about the back to 0 credit for fast 3 when you’re out of ammo ?
saying "it comes with 25000 Credits" is not an excuse. Those are used for quality and other things that cost credits, plus its super easy to accidentally waste them all on quality, cause retries save the model you used
I cannot wait to get paid out during the class action
someone has good promt for restaurant food¡
yeah, amazon and netflix both had to pay every subscriber when they changed their monthly without warning
hell in austria even the energy supplier had to pay back
not to mention they call it degraded service. that usually means that the quality of a service is declining, which usually means that you would get a decompensation, instead of an extra fee
The announcement still is unclear. Is that on purpose? It still doesn't tell if 100% of the coins are refunded or almost none of them.
It also doesn't tell if it's going to cost 10 on the long run or only for a day.
Also, people paid for their subscriptions and you changed the TOS/ prices during the subscription period?
Seems weird. Hopefully there will be a more clear announcement with hin the next 12 hours. Else I don't know how to keep going with my subscription/ coins.
i will give them time til the weekend at least, then i check up with the customer protection agency thingy we have
its just short term pain, you guys will live
Is there an official prompt guide?
I was generating about 300 hundred to 400 hundred daily... not now.....
since you like to be compliant to sudden and unannounced charges
poor google, they sure seem to need a handout
They have post you can search it
https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1
also you can see the examples on genai
Who will think about the trillion dollar corporations! We all need to defend their honor as much as we can
"You'll live." is true, but it's worth adding "But you might lose your job"
I’m not defending charges. I’m just not crying about adapting. Tools change. Winners adjust. Losers complain.
"Oh hey, our ads dont pay as much anymore, youtube videos now cost 5$ to watch"
For the moment, I've canceled my subscription so it won't renew automatically. If they don't put the 0 creit back, I'll complain to the consumer protection office
“Hey I’m going to amputate your arm just because” “winners adjust. Losers complain”
At least i get my money back
They don't even bother to be clear in the announcements
I am with you. 👍
yeah, i "adapt" by canceling for now
I understand the change and hope the increased load is alleviated in the near future, but jumping from being able to generate 20x 8 second clips for free at a time to having every clip be 10 credits seems like a nuclear option, opposed to applying heavy limits such as only being able to generate 1x clip at a time at the cost of no credits - just to have that 0-credit option still available at a BARE minimum
i mean on top of an annoying "You are generating too fast"
I wouldn't mind if they tell me clearly I'll have X a hours a day of free generations, or that it will go back to 0 in X days
Imagine your monthly coins would go down to 20. Would you still be able to adjust?
I guess not.
Anyways. It's a change of the TOS after payment. That's not even legal, if it stays like that.
I never saw that and I generated probably 5k+ clips over the past week
good for you, its in the bugs and feedback, so its not just me
just reached the maximum generations at a time trying to get the right dialouge from scenes with 2 characters and complex details in the background deciding to bug out
If my coins dropped to 20, I’d just switch tools and keep working. Adapting doesn’t mean pretending everything is fine. It means not melting down over something I can replace.
That’s because mitigating heavy load was never the goal. The goal was to sunset a popular plan that worked because it was good and delivered what it promised. It was the best option on the market and once they got enough subscribers they said ok and pulled the rug on everyone who already paid for it.
Then imagine next tool does the same to you, after payment. And there's not that many replacements yet. Not for quality content
see grok and suddenly moderating their self praised uncensored model
And who said we are not going to replace? We are just giving them a warning.
38 generation to get only one with my character moving his lips when talking. Would be ok with limitation if I could get him talking with only one
Then they should have waited until next subscription period to change credit cost
It would have even been good if they had consulted for recommendations due to the high demand, as they ask, or if we could give feedback so they can improve their product.
Because to deal with the high demand, they could have implemented:
1. Priority by plan
Ultra > Pro > others (in case there are other plans)
• For users on the Ultra plan, for the 3.1 Fast model you pay with your current credits, but whatever you consumed throughout the day would be available again the next day. If you use credits on other models/functions that already had a cost, those would be deducted and not reloaded.
This way, credits are still consumed, but Ultra users are not limited as much during the high-demand period they are facing.
The way they did it, they'll have to refund if they don't put it back
At least in Europe
If being upset is the hobby here, I’ll let you all enjoy it 😭 adapt or cry
General chatter space for ALL THINGS Flow
+100 social credit score
but every dumb message costs 10 social credits
-10 social score and loving it... I dont follow like most
I dont care what others think
Nanobananа only works from chat. It has completely disappeared from FLOW. ((((
i have it in flow
i have it as well
No. It just takes photos. And if it's based on a reference, it's a little more than NOTHING.
Hi guys, first post here for me ! Nice to see i'm not the only one chocked by paying fast generation now !
Just had a chat with support who at first told me it was written in payment documentation.... until i prove them not, and they told me that 10 token / gen would be permanent !
Keep that conversation saved for the incoming class action
oh god watch out it's mr "my lawyer already has the documents"
They also tried to change the subject ....
Google just hit up your dms. They want to offer you a free lifetime subscription for defending their honor
“Maybe if I defend the billionaires they’ll give me money” head ahh
Well that’s false advertising then
I'm not defending them, I'm laughing at your petty posturing
Chocked? Man, this is straight-up as1s of a situation. I swear, I hope aliens show up any day now and hook us up with their unlimited servers
Yeah they definitely couldve done a better job in communicating. Just dropping a "Hey this is how it is now" is just really rushed.
And weird way to touch base with your highest paying subscribers/power users
this is what it looks like when the flow devs wake up to find all of their tpus reallocated to gemini 3
The communication says "it will only be temporary" ... the customer service "it is permanent" !
I'm not surprised, its probably a rushed decision from the top. The announcement is very ambiguous about restoring some level of free gen most likely because they don't know and are seeing the # of unsubs first lol. But I have doubts it will return for 3.1
lets hope so. Its such a scummy move. Has crypto rug pull vibes.
each 8 second video we get on flow would cost $1.20 by api that makes 10 credits = $1.20.
you get 25,000.
that's $3,000 worth of api equivalent usage.
it's still a good deal. people who are that invested in unlimited video should spend like 4 months worth of ultra and just buy the gpu hardware to do it locally
Yeah
like my 5060 ti with only 16gb of vram can easily pump out 200-300 wan2.2 videos per day
They shouldve done the math before announcing ultra would get unlimited back in september. The issue is the lack of communication, as we can see here that nothing was announced until a bunch of ultra members talked about it. Thre's no emails or twitter announcements either
they probably did do the math, but nobody expects the people who are able to reach near superhuman levels of continuous video generation lol
It doesn’t matter if it’s a good deal or a bad deal. People bought ultra because it said it had unlimited generations.
nothing in their marketing ever labelled it as unlimited, just 0 credits for ultra. that's a "technical unlimited" in everything but wording. because they reserve the right to do exactly this
Thats deceiving
The issue is not the consumer, but the platform promising something then taking it away without any awareness.
They could have said Hey free generations for ultra members for 1 month and that wouldve given them plenty of new data to figure out a pricing model, but they didnt probably because they already had projections
Ok.
It doesn’t matter if it’s a good deal or a bad deal. People bought ultra because it said it had 0-credit generations.
If they don't say (temporarily) like they say in kling 2.5 turbo
Like why are we arguing semantics?
They say it’s included on this page ( screenshot taken this morning)
where they say until december 5th
We paid for an expensive deal .... contract is broken
"included" isn't "unlimited" though. companies have been sued for lying about it before, at&t vs ftc, I'm not on their side, but google has been careful not to use the specific "unlimited" in marketing because that's where things get dicey
Other generations have prices, we did not ask for unlimited but for free
doesnt matter, 0 credits is 0 credits
I hope SORA2 will drop access to the entire globe and wring the necks of these lovers of “unlimited access for 10 credits.”
Is there a tweet were we can complain ?
@flowbygoogle on twitter
alright gang, which company do we move to if the credit system stays in place?
In europe, they'd be force to give money back
Probably something like freepik
Each video generator has its own strengths, so you can leverage that
aside from third party providers reselling api and offering unlimited, there's always dream machine from lumalabs or runway's unlimited plans, but those can get throttle heavy
Yeah but no one has already reacted because we do not have a tweet from them ...
their communication has been pretty poor yeah
its probably a really small team
I hope it is simply they are going to bring it back soon, so they are not bothering
Once the new model's hype passes in a couple days
We are not against a limitation of one or 2 concurrents jobs in fast ... but paying one.... IF the veo3.1 was good enough to respect the prompt ... yes ... but it is not accurate !
If its 5 december, i will have already cancel my plan by then and look for another way !
sooo... they just unironically lied in this post?
why 10? why not 5?
TO you have the link ?
"some level" is so vague
dawg look in the flow announcements
my sub will end in 5 december anyway..but why 10 credit? not fair
I cant link cause this server wont let me for some reason
I’ve been an ultra user for awhile and I’m personally going to just hold my tounge because there’s conflicting information on the 10 credits issue right on whether it’s permanent and not
I get the traffic overflow and spam issues right now but why are we getting multiple answers from different sources
you think the 10 credits will stay?
I have no idea
Yes that's how income tax started and is still here
If it stays consider my ultra subscription over since there’s no point of keeping it just for some more points you can just buy anyways
right! me too!
But like I said there’s conflicting information on the issue right now. So I would wait to see what any of googles guys in the server responses are to @sullen mountain’s support response if any.
i hope they will notice on it..
They did not say anything yet ?
Not to your post no.
they put a message in flow saying will restore some amount of 0 credits
I would honest throw that screenshot you have in the support form in the server an see what the response is to that
How many are some, who knows?10 videos, 100 videos, 1000 videos?
That’s my pov, I’ll just downgrade and buy more credits as needed, likely not use it as much. HF sometimes offer free generation deals for a week
I have to hope their team is reading this chat and the reddit, whether or not they're covered legally by careful wording, it is a blatant bait-and-switch of what was promised that will lead to a lot of unsubscriptions
Since there’s now genuinely conflicting information floating around
guys lets wait for it
Pretty much what I plan on doing if this garbage sticks and they really are trying to be slick about the 10 credits being the new norm
I appreciate your answer, but dont you think we customers bought the Ultra Plan because of its benefits and suddenly and abruptly it gets changed without even noticing us? Dont you think its unfair for us that it will be restored to some extent but not the same plan we bought
But like i said, i would hold my tounge intill that conflicting information is clarified
i hope they will lower the price of ultra too
With unlimited videos, 125 bucks is really good
for 3 months
The 0 credit gens is only reason I scraped the money together for it. It’s $220 a month at half price here
then create a new account and restart
I doubt that but even if they don’t lower the price. If they really are trying to sit here and bait and switch up on the top tier. Consider me out man
If sora 2 gets the ingredients feature of using 3 images instead of their 1 right now, its over for veo. Their few free generation a day already do more for me than my ultra subscription. 15 seconds of legit video compare to 8 sec of random veo one who mixes up dialogue all the time. Takes me like 30-40 veo renders to get the right ppl to say their line.
with another credit card?
Literally the only reason I have the ultra tier was for that benefit
I believe they don’t really do real people though last I heard? (Could be wrong, haven’t used it)
I mean realistic people
Same with me
I use both for animation and veo has crap clunky animation compare to sora
Yeah I’m personally considering other options now with everything going on with flow at this point
THe more competitors the cheaper it will get
I’m glad we have the extension feature now but why not an actual increase in the time options for the generated videos. I wouldn’t even mind if they had so that it was only available for higher tiers whatever at that point but the fact Sora is kinda walking circles around Veo in many attributes kinda is making me consider my options
And with the all of the problems Veo seems to be facing every week at this rate idk if I can justify the money they want every month for it
I only use veo because ingredient makes my character consistent in all my scenes. Otherwise when I use sora 2 to make my animation, the flow of the animation is 10x better, less effort since videos are 15 seconds long
High action scene sora > veo
also good to keep around with a timestamp
Hello! No, there were no lies in that post. We will indeed restore some level of 0-credit generations as soon as possible, but I'm not sure yet when we can safely handle the increased load that brings.
And the one saving grace Veo had apparently (intill properly clarified) being bated and switched definitely has me considering
Ah well speak of the devil lol 😂
I feel they should too to like $99.99
yep.
You say some level, that's too relative
Some level. Some time. That could be one gen a day a year from now.