#general・🧺
1 messages · Page 38 of 1
Design: Figma and affinity
Landing page design: Google Stitch
Development: export from stitch
When that come out?
guys I want to build a character for my app but with using google ai studio this is what i get
any tips
give it a reference img prolly, and try image -> svg instead of creating it from scratch.
bet
hi
this prolly sounds hella weird, ik this is stupid but plz tell me how to use the ingredients correctly bro i can not for the life of me figure it out n all the videos are on a older version
haiii
Hi guys
hello
this was kinda interesting https://telegra.ph/The-Apple-Doesnt-Fall-Far-From-the-Code--How-AI-Models-Inherit-Their-Founders-DNA-03-22
There is a persistent illusion in the tech world that Large Language Models are objective entities. We talk about them as if they are cold, calculating oracles birthed purely from mathematics, silicon, and terabytes of scraped data. But anyone who has ever built a platform from the ground up knows a fundamental truth about software: code is just...
Hello.
Hi
@molten dock
HELLO
Goodnight
Good
When Veo 4
should be around may 19 - may 21 2026
Hi
morning mate
Morning mate. It's 11 pm here. Where are you.
Hi 👋
Hi
jsut added search for all ai models on the internet zz time
what is this
can we talk about antigravity here?
Hello
was about to ask the same 🙂
hey
Custom program I'm making
Hello @winter sinew and @dusty sparrow while you can talk about antigravity here, there are no specific channels here for it. There is over in the Google Developer Community. There is the vibe-coders channel. You can find the server here. https://discord.gg/google-dev-community

Video Generation inside Gemini app Isn't working properly. I needed a video with 16:9 aspect ratio but it provided me a vertical 9:16 ratio video. I mentioned the aspect ratio in the prompt but it failed.
Use “landscape” keywords
Add multiple cues:
“landscape orientation”
“horizontal frame”
“wide shot composition”
Example - Create a horizontal cinematic video in 16:9 landscape orientation.
DO NOT use vertical format. Strictly 1920x1080 resolution.
Wide frame, horizontal composition.
Hello.
Does Flow usually not work at around this time?

has anyone created an animated character using google ai studio
Hi
Hope you are doing well today
Yoo what's good guyss
Fails spike during peak work hours.
Hey @grizzled lava - can we help you?
Hi all, sorry for the dumb question, but does anyone know where the Disco channel is? I can't seem to find it!
Hello, what is the solution to this problem? I registered with another account and it's still there.
hey! it might be that its not avalible in your country? can you try the flow website as whisk is merging with flow soon!
what is flo??
sorry, flow* its another labs product!
we have #flow-general・🌊 and many more flow channles to find out more about the product and the use cases!
okay thank you
haiii
Hi
What's up?
not much im just working on a video
Editing?
That's great, I guess.
You're from US. You know what I have met many people and about 60-70 percent were from US.
yeah lol probably because us makes up like 4% of the world, and you take out the people who dont have internet access, and then factor in the amount of people chronically online... yeah americans would make up a large portion of that probably
yeah a lot of people on the internet are from USA or India simply because of the population. and not many from China because of The Great Firewall
😭😭
Maybe it doesn't. It totally depends.
eh i never studied when i was in school (bad idea, you should probably study)
oh what tablet?
oh nice i have a pixel tablet
apple kinda sucks i hate apple 😂
ive heard good things about xiaomi though
and im also an android user
ive used apple mac computers in high school and i hated them so much
yeah the processors are good but i just hate the interface
You were a windows user before that?
linux
230k rupees is not very much 😭
yeah i want to move to a nordic country or portugal
yeah ive heard good things about them
oh i have 3 raspberry pis they are fun
1 of them im just using as a server and another is for testing stuff on and the 3rd is just for extra
good morning guys
Morning
Good
hii
Hii
yow
Anti Gravity has solved so many problems cursor chokes on.. cursor is great, but I’ve had it maxed for like 2 months now.. Anti Gravity😍
Good day everyone
Is it zealy offer still active or expire?
@silent dawnHi, just a friendly reminder to keep your messages relevant and avoid excessive emojis, spam, or unclear content. We’ve removed the recent messages that didn’t comply. Thanks for helping keep the chat pleasant for everyone.
You can ask in #flow-general・🌊 and if you suspect its a bug, you can make a thread in #1238621511401734254 !
Might wanna clear the chat spam in #flow-general・🌊 while you're at it
hello
Can anyone tell me which linux is best for me as I'm a begginer
Hi
There is no single best. Just choose the one that fits you the most
I'm begginer as I mentioned I have no knowledge
then it really doesn't matter. Just go with any, and if you don't like it move to the next Distro
This is not good
I'm asking for your advice
that is the best way to learn imho. Go with Ubuntu.
Whuch OS do you use?
Windows. Remember Linux does not use OS's
🥲 okay
there are tons of YT videos about this. Most of them are valid.
But honestly, just start with Ubuntu. You will learn how to use the terminal. You are going to need it.
Ohh okay
Hello! I signed up to the Disco private invitation more than a week ago, went though the whole process but did not get access yet!
What the best way for me to follow up? Anyone can point me to the right direction? Thank you 🙏🏽
Hey folks, just a heads up that non-Google Labs chatter is best for #off-topic・🐢
Thank you, good winds.
please be aware that veo 4 has not yet been announced and it is all speculation!
Go someone know how to enable the Chrome-Gemini update ?
You broke tos for discord
how?
AI in moderation not allowed
oops is it illegal?
no
oh then its fine
Just break there tos, they can remove it
oh
check discord support for more infomation, not me
thx for informing
or else me cooked
but why isn't it allowed though AI is soo acurate these days
Hi
Hi
Ah, thank you for that!
oof rip doppl
hiii
hello
Gee I wake up and already find everything around's here burning. Nothing new then.
Helloo
hi friends
Sıfırdan, kod yazmadan ve tamamen ücretsiz bir şekilde kendi Android uygulamanızı nasıl yapabileceğinizi merak ediyor musunuz? Bu videoda, Google'ın yeni nesil no-code (kodsuz) araçları olan Google Antigravity ve Google Stitch'i kullanarak baştan sona mobil uygulama geliştirme rehberi hazırladım. /reklam zorunlu
Google Stitch: ht...
I explained step by step how you can build a mobile app from scratch with Google Antigravity. It turned into a practical guide that will be genuinely useful for anyone getting started with app development. It will definitely help you.
Anyone here experimented with Stitch from Google already?
I tried it in the video I shared above 🙂
hey! try #stitch・🪡 🙂
👋
Does anyone has a petition to change the UI ? dear god it sucks, haha
Thanks m8 wasn’t able to find the channel
feel free to add tool/experiment feedback in #1238621511401734254! and please use the tags so we know which tool you're referencing 🙂
Hi
Has anyone gotten their Disco invite? I received an email March 5 to be a part of the first wave andsubmitted the form. I haven't heard anything since.
if i paid monthly for flow why im getting just 1000 credits
Pro subcriptions include 1,000 credits per month - here's more info
Members of the Google AI Plus, Google AI Pro, and Google AI Ultra plans now have AI credits that can be used across Flow, Whisk, and Google Antigravity. Learn more about Flow.
yes but its like nothing
for 1 minute video 150 credits its gone
if i want to create 2 videos or 3 per day
1 week and they are all gone
Do you believe that the amount you have to pay in exchange for what you receive is not worth it?
what i recive bro i just started
It's really expensive, bruh
Hello! I got the Disco invite 2 days ago, and I've filled out the form as well. Is there anything else I need to do to get access?
Hello
Hello
When I'm using NotebookLM to create deep dives for my Physics papers, I have to always add:
No video game, TV, DVD, or other consumer electronics analogies or metaphors.
It's the #1 problem with NotebookLM!
I really dislike these metaphors. We're exploring an advanced concept, and suddenly we have to listen to some really cringy analogy comparing the math to a DVD, etc
I don't mind the technology, it's just that the metaphors NotebookLM makes are always really obsolete/outdated, like it's trying to explain the concept so your grandpa will understand
It's brilliant, as long as I always add the "No Consumer Electronics Metaphors" disclaimer. Seems NotebookLM is engineered to specifically attempt to make Consumer Electronics analogies.
hey! this might be better said in discord.gg/notebooklm as that is a notebooklm specific server and you might be able to get more help there!
cool
Galaxy M12 - to - Docker Server (Custom Kernel)
Hey everyone! Just wanted to share a quick win: I’ve successfully modified the Samsung Galaxy M12 stock kernel to support Docker and native containerization. It’s now a fully functional, pocket-sized Linux server! 🐧
The Mod
To get Docker running, I recompiled the kernel to enable:
cgroups & Namespaces (for process isolation)
Veth & IPtables (for container networking)
Optimization for 24/7 background stability.
Why do this?
Built-in UPS: The 5000mAh battery acts as a natural backup—no downtime during power blips.
Ultra-Efficient: ARM architecture uses way less power than a traditional PC.
Sustainability: Perfect way to repurpose hardware into a home lab node.
Use Cases
It’s now running Pi-hole, Home Assistant, and some Discord bots flawlessly. It’s a great way to save money on cloud hosting for small projects!
Anyone else here experimenting with "Mobile-to-Server" projects? Would love to compare notes on I/O performance!
@google Labs
@outer wagon i cant move the frames in stitch . And is missing a align grid button. "SEEMS AUTOLAYOUT " NOT ONLY TOP. RIGHT. LEFT.
please use #1238621511401734254 to submit any bugs you find also we have the #stitch・🪡 channel for focused stitch conversion!
Anyone know the best way to get videos like this? https://www.instagram.com/reel/DWKwV5KB5Mj/?utm_source=ig_web_copy_link
How to Target Every Part of Your Back
Angle changes muscle focus. Row at 30° to hit upper back, 45° with elbows out for rear delts, and keep elbows close pulling to hips to bias the lats. Small adjustments, different results.
Train your back with precision — train with @fitonomyapp 💪
#upperback #reardelts #lats #rowingform #backtraining
6610
god day guys
why that video look so bad 😭
What do you guys think,
Is it actually effective to start prompts by assigning the AI a role (e.g., “You are an expert teacher” or “You are a skilled product manager”), or does this have little to no real impact on the quality of the response compared to just giving clear instructions?
At least in image generation i've seen using the word "Professional" as prompt more often has great results. So "expert" or "Skilled" will most likely make AI try to use the best rated material possible.
However that might not always be exactly what you could be looking for... I imagine as teacher or product manager one would have to pair it with traits like "Empathy"
AI models are now more often designed to care for the safety of humans, but they also are much better than most of them at anything, and being the best or "Skilled" at something won't always mean you'd care for the means or not.
Empathy for example is not only used to care for people, it can also be used to understand what they really want or need, which is good for sales.
Why is this feature so expensive, and how will future competition prevent Google from raising its prices to the point where young and innovative leaders will not have to fork over their entire paycheck to utilize the AI interface?
Google's services aren't bad, but If you get good at prompting and keep searching alternatives I'll assure you that you'll find something better and cheaper. I know well some other video generators got so good that they could even go out of beta phase by now and remove the free tiers.
hloo
And I've seen exactly what things is VEO missing that those other video generators can do and yeah, It's kind of a big deal.
As for the image generators, yyyeah right now Google pretty much is competing against itself with the Nano Bananas, who knows why. It'd be so simple to keep original Nano Banana enabled at their sites like Google AI studio does and keep all the market for themselves.
I feel like we could do more to help capable business leaders exponentially increase their output through AI, but these companies are just not letting them do that, you know what I'm saying
Google's doing that monopoly thing again
When Veo 4
I don't think I can justify having Veo 3 anymore. Had it for 9 months straight now
Ultra edition
This is like Gemini with Imagen 3 when it was first released because they kept trying to use ImageFX from the AI Kitchen instead of giving people the full image generation they deserved
And it just kind of hurts to see as someone who's involved in business myself
tbh pretty much any other AI video generator is working almost the same with the exact same type of prompts. So best you can do is find the one that gives the most use per money or even free and practice there. You'd be surprised.
As someone who's had to work with people, meet with people for business meetings during dinner, it just hurts to see a company do so little to help cultivate business leaders
Especially when these young and talented entrepeneurs are the future
Also for the videos that I used to make the flow update completely ruined the tool. Scene builder is now trash and all you can do is extend. You used to be able to do everything in one place. Not have to leave the scene builder to make another scene
Yeah, there are so many issues going on, and it's insane
It is insane
New flow runs exponentially better and looks cool but I prefer the bugs and glitches of the last version
Apparently the excuse is that we don't leave that type of feedback at the "labs bugs and feedback" because messages get lost in the chat, even tho those chats are designed to talk about the tools too.
So yeah you might wanna leave that complaint at the labs bugs and feedback area so there's no excuse.
I'll write out something in a second. If they make scene builder able to do everything in one place again I'm sold but otherwise I'm probably gonna go with _________
Idk if I'm allowed to say so I won't
Then again we are indeed at general and it might pass over their heads here. But there's all sorts of complaints at the tool's channels too.
Oops sorry for talking in general instead of flow general
Nah here we can talk about most things anyway, plus it was just casual anyway.
Does anybody actually use the other Google ai tools ? I feel like (and maybe I'm wrong) that everything Google had made with ai tools is pretty...mid
Flow is pretty good up until recently
Well the other that's connected to the same billing account besides the Labs ones is Gemini.
The other labs tools aren't included in ultra anymore?
It does allow some other options, but it seems it's mostly casually
No they are, at least what you see at the billing page
The Gemini one should also be there, they call it app but it's mostly a separate webpage
But if you payed a plan, you also have that
I've used it every now and then. Still not sure what the best llm to use is
I paid $200 for the openai plan at one point but there's no way I could justify it and I didn't like sora that much
If you want to pay to talk with the AI chat models, probably even those chatbot sites would be better anyway, they have lots of models.
Also Deviant art just added video creation tools, and Veo is there, but creating them there costs the equivalent of 90 cents to $1 for 1 eight seconds video, they're drunk
And they have a minuscule token limit too, can't even do a decent prompt there. What a robbery.
is anyone willing to take a look at my app and give me harsh feedback
hey guys can anyone guide me that if im using google flow for video generation, i am only able to give 3 ingredients and none after that, should i use the frame method after i generated the ingredient video? how do i stitch them? i was using elevenlabs flows but it has a lot of limitations too. i want to create an ad/tutorial video for my app
Hi welcome
whatd vertex ai?
anyone know if agent parallelization is possible in antigravity? / spawning agents, delegation etc.
Um... hi. Is anyone having issues right now with generating? I keep getting a error saying it failed to generate
yeah me too, one after another
It’s really frustrating
its working like 1 out of 4
Yeah I get the same thing
Any campaign you have in mind
1 of 12 generations worked 🙁
hello
I have credit pro but i want now download 2k images and didn’t work why
Hello
yup same with me
Hi 👋
It happens to me sometimes
Hello
hey! i saw you sent me a dm! if you wish to speak to mods please open a @fringe nova ticket!!
Why
hey! welcome!
Hi
no more music sandbox nuuuuuuuuuuuuuuu 🙁
Hi everyone—could someone tell me why this issue is occurring, or who I can reach out to here (or where I can submit a support ticket)?
Error:
You are sending generation requests too quickly. Please try again in a few seconds.
best place for bugs is #1238621511401734254 !
I’m facing a ‘resources exhausted’ issue. The website shows ‘too many requests’ when I try to upload images.
I’ve stopped generating videos on this account for the past 2–3 days, but I still can’t upload images from the website. Could you please help me fix this issue?
I have got the same issue
It’s possible you all ran out of image generations for the day. Previously, when you hit the limit, Flow displayed a clear message like, “You ran out of Nano Banana 2/Pro image generations for today 🍌.” Now it instead says that you’re generating them too quickly.
But some accounts do seem to get limited for some hours even if they haven't been used for the day, eitherway the problem fixes itself 12 to 24 hours after it starts happening
So yeah if you still get that message after 5 minutes don't bother camping there and just come back way later.
That doesn't really make sense—especially when you're paying for it.
it's just some sort of bug that started some days ago when apparently flow overloaded with users. You get used to it, but yeah if you were already creating images today the account is probably fine don't worry.
Been dealing with that all week and my account been fine later too.
idk at this point a 12 hour wait seems to be good time to just "pick a pencil" like social media often suggests to us AI image generators
At first, I checked to see if there was something wrong with the account; I tested it on a different PC and IP address, but the error appeared almost immediately.
It's just tied to the account not the ip, it'll be fine
OH ALSO, it affects Nano Banana Pro WAY more than Nano Banana 2, so you might wanna try swapping to that one
I keep forgetting about that
Ok I’ll try it tomorrow ty
Seems way more people are using NB Pro than NB2.
Only youtubers that click and run on things keep saying NB2 is awesome, actual users don't seem to be liking it much, much worse over Whisk's original Nano Banana... or "Precise mode" as most think it's called for obvious reasons
Is google ai studio good for creating games
if not pls suggest me the best software for vibe coding games
<@&1173646484311384114>
good morning guys
I think Google Labs should do more to inspire young and innovative business leaders to pursue their passions by limiting the amount they charge future business leaders
Who in this server agrees with me that the corporation should do more to help partner with the youth in developing sustainable measures and metrics to bolster profits and foster innovation and collaboration through the ease and compatibility of AI integration?
Hello
AntiGravity Though 🙌 💯
Recently learned about the youtube, products, research, and history of Google - they are insane!❤️🔥
Good morning everyone

Mornin
hello good afternoon everyone
hello guy
Guys how much point does flow gives per day i am new ✌😅😅
Hello, can I ask something? Because I don´t understand one function.
I want to ask about credits, when I set monthly plan, can it be cancelled at any time, or how it works? Thank you for help. I never try it.
@dusk tideHey, please don't mass ping people without reason or context. We've deleted the recent message of yours that didn't comply. Thanks for helping us keep the chat clean!
Hey, you can learn more about credits and plans here: https://support.google.com/googleone/answer/16287445?hl=en
You can cancel any plan anytime though!
Members of the Google AI Plus, Google AI Pro, and Google AI Ultra plans now have AI credits that can be used across Flow, Whisk, and Google Antigravity. Learn more about Flow.
Notebook LM is such a powerhouse for roleplay compared to other AI. I just provide it all details about my genre of roleplay and then upload files on how to Write novels and its a master piece
Yes, that worked.
Hey
still so many errors, whats going on? 2 days like this?????
hii
Gm gm
hello guys i have an image to enhance the quality do you guys have any prompts i can use
hi
good morning
Which one was the IA /model which could make videos?
Underprovisioning. Look at the hours you get fails, and try at different times of the day.
Also if you're using Nano Banana Pro try changing to Nano Banana 2 cause better than nothing, that one's failing somewhat less, but when it starts failing non-stop too yeah you better just come back in like 12-24 hours.
this was using NB2
Flow isn't displaying the correct messages for some things too, it's also possible you ran out of image generations for the day and it isn't telling you. So yeah like Modal says would have to try back in some hours.
Tho with that error message it could also still be working sometimes
Im on Ultra, its not running out as it creates more as the day goes by, just seems like an error, just like yesterday
Start taking notes on when you get fails. Typically they spike during a few hours, when people are at work in the Americas and EU/Africa. You'll find a pattern.
Or just wait and try later.
Help! I couldn't finish creating the video, but I got deducted credit.
any help im new here where i should start
worth it ?
hey Zaak! welcome. you can learn more about Google Labs in #start-here・📍 - we're all about experimenting! we have several tools and experiments on our website, https://labs.google/ 🙂 please note that some are behind paid models and not all tools are available everywhere.
if you have questions about any of the tools/experiments, or want to showcase what you've made, check out the #server-directory・☎️ to find the best channel!
if you have moderation concerns, feel free to ping the Mods or send a ticket to us by DMing @fringe nova!

for those who are making app do you use stripe for your apps
to make money, or is their anoter thing that doesn't take as much percentage of your money
does the tiny lesson website have czech?
not at this time, no
hello i want to know the credits of free video maker on flow restarts everyday?
Now that I’m back to using labs.Google, and flow, are there rules about posting on places like Instagram? Or other social media outlets? I have a Google subscription.
How do I always get aired in these chats bro
not sure why pomeli can't read my site
How to use notebook lm to create website
Notebooklm is so underrated.
good day guys , Happy Ram Navami

happy Ram navami everyone
hi. if i'm someone who hasn't ever done or learned any coding ever... and i wanted to make a kings on the corner game app that i can play with my Gemini either in sidecar or on my mobile; would i use opal or stitch for that? or if i need multiple; which ones and in which order?
Hêllo
Try out Google AI Studio first, work on the mechanics, then for the designing portion, I'd go with Stitch. I haven't worked with opal too much, so i cant say much about it atm.
Hope that helps.
GS - https://aistudio.google.com/apps
Thx!
Hi guys, has your content generations been failing in the last 30 minutes?
yes
very difficult, there is a problem
On Whisk???
can I use Stitch in VS code?
still having the generation error
hello everyone
yo guys how to contact support or to get help ? about google flow
Flow still down?
In mine, I can generate images but I cant seem to download any of them
flow captcha error ?
{
"error": {
"code": 403,
"message": "reCAPTCHA evaluation failed",
"status": "PERMISSION_DENIED",
"details": [
{
"@type": "type.googleapis.com/google.rpc.ErrorInfo",
"reason": "PUBLIC_ERROR_SOMETHING_WENT_WRONG"
}
]
}
}
<@&1366831844502274180>
Google Flow dont work today Scared!!
What's wrong with whisk today that isn't producing anything?
Bug general
Just today I was planning to record something about generative AI
Murphy's law always works this way
It will be tomorrow, today I'll finish the script about generative AI without spending money
its been like this with flow for days, nano banana 2/pro nothing works, people on here telling you to wait while google NEVER waits to charge you 250$ a month
It works on Gemini at least, the bug must be on Google's end.
Bug solved!
it was time
Whats up with this "Oops! Something went wrong!" crap this morning? Can't create anything this morning
Same here .. it sucks 😭😭😭
pomelli is not working either
Hey, I’m experimenting with AI cinematic storytelling.
Trying to create a concept around identity control.
What do you think works better:
– subtle transformation
– or aggressive visual change?
Both have their merits, You could even take a little from both piles and have a short aggressive change that only does a portion of the whole transition.
That actually makes sense — mix of both could hit harder.
I tried mostly subtle, but maybe adding a short aggressive moment will make it stronger.
Appreciate it 🙌
I've been interested in AI art for a little while and I was trying my hand at it.
what have you tried so far? 
I've been doing some tokusatsu-based artwork so far
Hey hey
Hello?
Is it normal that the image generation of Flow is currently not working?
This will be resolved before tomorrow?
welcome to the community. I've posted an update [here](#flow-general・🌊 message) that we're currently investigating. thanks for your patience.
I got access to disco, anyone know how long that takes?
can take a few days since I need to manually verify everyone, feel free to dm me your email address
Hi
By any chance is it possible for me to be accepted for google cc, I am on the waitlist?
Will producer.ai get integrated into google subscriptions? When will it happen
hi
hi
flow nanobanana is down? because i cant generate anithing
Hello everyone, I have a problem, everytime i try using nano banana, this message appear : Failure
You are requesting generations too quickly. Please wait, then try again.
I am using google ai pro, does anyone know when i can generate image again? it's been a day
its happened me too
Interesting
my account has faced this problem last 5 days (geneation too quickly)
I have a ultra plan
please help me. Please resolve the problem
or suggest me, if I purchase a new ultra plan or not.
hi
hi Everyone! I just jump to this server!
I have accepted DISCO invitation, signed NDA and all, yet I don't see 🔒 #disco-🪩
Am I missing something?
Flow is working much better right now, but if that keeps happening to you check if you're using the Nano Banana Pro model, and if you are try swapping back to Nano Banana 2. Pro still has a very little daily limit
And if both models still give you the "You're generating too quickly" message more likely you just ran out of image generations for the day and gotta come back in 12-24 hours. Flow is just not displaying the correct message for when you run out of image generations for the day atm.
hallooooo
Henlo
hnleoo
Because pretty much any other of those problems Flow had seem to have been fixed atm.
Nano Banana Pro is just about 10% better than Nano Banana 2 anyway, but Nano Banana 2 is faster and more cost effective so yeah, your account gets more images for that one.
Hey!
happy weekend mate
yess happy weekend!
Hey
Hi, I haven’t generated anything for the past 4–5 days, but I’m still getting the “you’re generating too quickly” message.
Also, I’m unable to upload any images at all. It seems like the issue is not related to daily limits.
Could you please check if there’s any restriction or problem with my account?
Sorry I'm not staff bro, just a little chameleon. But yeah that issue is unusual atm, you will have to open a thread at the "Labs-bugs-feedback" area for staff to check what happens to you
But just in case do check you're not using "Nano Banana Pro" atm and swap to Nano Banana 2, that could also be it
thanks brother for your helping hand
Also in what country/region are you? That could help them
Is there a channel for anti-gravity app talks?
I basically need help with the always proceed, which doesn't work in the anti-gravity app
hi
Anyone has any opinion on Jules?
im not sure if it's right place to talk about it.
ah, found #jules-general・🦑
is there way to create a longer video on flow ?
hello everyone.
I come from a c background and this semester of my cllg we r tasked making a mini project on the topic resume screening. we got very little time. can anyone help me through this. I have started learning python basics but tell me how much python to do and how to proceed.
we basically need to use some kind of ai in the project
not sure if this the right place for this but I did a livestream with Ivan Leo (DeepMind) today about how build a deep research agent with Gemini! https://www.youtube.com/live/LUfqQgz1-Os?si=xeAI-iOwi41UlX5-
repo here: https://github.com/hugobowne/build-your-own-deep-research-agent
all feedback and thoughts welcome and encouraged!
Ivan Leo built production agents at Manus before joining DeepMind. In this workshop he shows how to build a deep research agent from scratch with Gemini, sta...
Ehhh maybe the #off-topic・🐢 I'd say this is alright though we should have a dev showcase channel though.
I'm sure @outer wagon will let you know fs when shes back.
cool lmk 🙂 thx!
hello
i have a doubt
i have the student pro plan so i should be able to gen a 3 min song using lyria right? but why does it gen 30 sec?
Hi
i tried all the model like pro fast thinking it gen 30 sec
Does it have a name the thing which Google Ai Studio use to make scripts for the apps, etc?
Lyria 3
its in gemini and google ai studio i think
Is there any server or channel to talk about that thing or suggest?
sadly no ;-;
i dont see one
and am unsure where to reach out
ok
Is anyone up?
yes
thanks for reaching out! the Disco team adds new members each day. if approved, you should receive access to the new channel within 24-48 hours, but please note the team is away for the weekend. appreciate your patience!
Hi there, there is not a dedicated antigravity channel here. We recommend you head on over to the Google Developer Community. They have the vibe-coders channel for antigravity. Here is the join link. https://discord.gg/google-dev-community
Hi there, while off topic would be a better channel here. Things get lost in general. You would really be better off in the Gemini server. There is a dedicated deep research channel. Here is the link to that server. https://discord.gg/gemini
Hi and welcome. For lyria you would want to head over to the Gemini server. There is a lyria channel to discuss all things lyria and a couple event recaps. Follow the link in this message to join the server. #general・🧺 message
U
hello guys
@errant pelicanUnfortunately this isn't Google Support so no one here is able to assist with this issue. https://support.google.com/ will be able to assist with account issues. Sorry you're having trouble and best of luck
Hi Google, when will Gemini have a Connector app?
Good afternoon. Does anyone know any Google apps or websites that can be used to storyboard or any other apps in general
Hello. I'm new
hi
are you currently working or still studying?
dont they already have that
I have a question about the flow. If the image I generated involves a child, can I still use that generated image from the flow and use it as a reference in my second prompt so that the character stays consistent? Does that count as involving minors?
I am working on a short film using flow/veo3. There is a scene with two guys where only one guy is speaking. He has two sentence dialogue in the scene. Veo keeps making the first guy speak the first sentence and the second guy speak the second sentence. How do I prevent this? Even after mentioning in the prompt that only one person speaks, it still doesn't follow it.
Obviously, since it's a child, doesn't matter if it's uploaded or you made it yourself. But Flow, even with adults seem to treat images made there the same as uploaded images, filter strictness wise. This is not the case with Gemini
but is it breaking the rules if i made it with flow?
Well, I know Flow and WhiskFX don't allow the upload of images with minors, but that refers just to uploading. I mean, unless you do sussy stuff, you're fine.
ok thanks
hey! is it possible to add nano banana 2 on opal [Experimental] ?
works the wqen can go on browser and see stuff 😄
Does anyone know where i can post a (relatively easy) feature-request for Google AI Studio?
Google has a site for that
The issue tracker can be used for feature requests, support requests, or any other type of development task that a project needs to track. You can file a feature request by clicking "New Issue" and en
oof
i think i fall into the "general public" of their terms of service(attached) but that led me in the right direction of this embedded modal on their FAQ
https://aistudio.google.com/apps?source=faq
thanks, @bronze marsh ! i posted there but got autoblocked. any ideas?
I do not want to speculate on what happens in a community I am not involved in. You would be best to reach out to a moderator in that server. If it was an autoblock from the bot though, it usually will tell you why it was blocked.
Has anyone else experienced this issue when trying to generate images in Nano banana Pro 2?
Error: “You are requesting generations too quickly. Please wait a moment and try again.”
@mental sluice so far, so good! Looks like a cool race car.
Anyone ever looked into complex issues
good day guys
Good
Flow, why doesn't it have a choice of images on the phone like Gemini does?
Nice.
hiiii
hello ^^
Hello everyone!
Hello Everyone 🙌
is it just me or has google once again changed the quotas on antigravity? using the AI pro plan
wassup
forza horizan 6 is comeng
Are the 50% promos done for ultra or still a way to get it?
hello everyone

(This isn't a promotion it's built for stitch)
https://blog.google/innovation-and-ai/models-and-research/google-deepmind/dear-upstairs-neighbors/ any more info about his video or when the customize gonna come in flow ?
When exactly will code wiki for personal repos be launched
longest 10 minutes of my life. ANYWAY
TurboQuant really boggles my mind.
Oof. I hate the cooldowns sometimes because some platforms are a little too liberal with them.
Bro the bot spam is crazy
Discord is like email with spam.
Hopefully that gets adopted widely, and quickly.
Do you not see this on the pricing page?
😱
Hi 👋
Hello space x
Hi there,
I have a question regarding the Antigravity software and the Google AI Ultra subscription...
When interacting with the agent, after a long exchange, we eventually start consuming AI Credits, which is normal. However, once a certain number of credits is consumed, the interface freezes and prevents us from continuing the conversation... even after several hours of chatting...
And when we check our activity, the AI Credit balance is far from being depleted...
You're hitting a bottle neck in local memory
Its basically going to be like that regardless due to memory issues
hi my name is brad, i am a ping pong ball, i don know how i bounced in here .
Why is the lame chatbot not banned yet? Are there any mods on this discord?
video generation is expensive and hence nobody will give it totally free !
in the past it has restricted me using minors.so i use different tool which involves generation of videos in which a minor is involved
Is there an issue with video and audio generation in particular? Almost all of my prompts are failing with the message: Failed Audio generation failed. I don't even need audio and have added this into prompt and it still does it. No problems on my side with bandwidth or space.
Do you have option to generate using veo 2.0 (no audio) ?
its cheaper as well as you dont need audio
Used to but now only 3.1
ah..so it does not show an option to select 2.0 anymore ?
Is there a way to revert to previous version of Flow?
I am no more on pro plan hence dont know abt the latest..in the past i used to select 2.0 when i did not want audio as it was cheaper
yup. I might be joining you in leaving the Pro Plan. Might look for a better video generation AI with more reliability. Any suggestions?
ah. 🙂
Hi, if there is something you need to report to a moderator you can DM @fringe nova . Thank you!
Google is good with Video Generation (You need to know the limitations and play around it).I also use MidJourney at times when i cannot get somethign out of Veo.
I use Flow mainly to animate static images, I have found to date the renders veer to unrealistic or obviously wrong. My prompts try to ensure subtle animations so planes don 't suddenly appear or cars appear and drive on the wrong side of the road through another car 🤣
Am I in the right server for questions about my Gmail?
I add something like this at the bottom to avoid that, more often it understands.
[Fixes: No unnatural clipping/intersecting/overlapping. No occlussion errors.]
That also helps limbs or clothes not to ghost through each other. VEO has other bigger problems specially with voice, it doesn't know what "Vocal distress interjections" mean, other video generators like grok and wan have no prob with it.
Unless you specifically prompt those it just can't do it
And yeah also gonna cancel everything Flow related past the 31st, gonna need that money to sub to a place that still has original Nano Banana available and there's plenty. Nano Banana 2 can do similar things but it just takes way more time unless you need it for writing random text.
Been testing some other sites and NB1 still works fine there.
good morning guys
Nope
Hello, everyone
🥹
Trapped in the snow-capped mountain for three days, I discovered a container.
veo+kling
hello
Hi
Hello guys
Hello Guys,
Since i did the 1.21.6 update on windows with wsl 2, the agent is locked in a sandbox mode in my wsl without network, privileges etc. I've tried all the recommended settings to "release" the agents from the sandbox mode but without results.
Do you guys experience this too ? Has anyone tips or workaround ?
is there a gemini cli channel here?
I feel like even on ultra subs the model is taking an huge amount of time to respond " ⠏ This is taking a bit longer, we're still on it. (esc to cancel, 1h 49m 4"
I just paid in full, and now this happens. There are still 200 videos waiting—doing everything manually—and my kids are crying nonstop. When will I ever finish to deliver to the client? I think Google Labs really needs to reconsider how they operate and treat customers, or people will just switch over to Grok.
Content about vegetables deserves to be limited? Then what are you even building AI for, Google Labs?
Now that Flow does not have Nano Banana, what is the replacement?(That has cheaper credits)
Can I take inspiration from copyrighted images and use that to create my own image with AI
@onyx shardHello as 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. If you are asking for prompt advice please do so in the appropriate channel. Thank you!
reminder, general is not the place to ask questions about products as we have #flow-general・🌊 , #whisk-general・💭 , #jules-general・🦑 , #opal-general・✨ and many more channels where your product specific questions can be asked. Thank you
Failed
Something went wrong loading your media.
** anyone getting this error today in the past 2 hours on all videos in Flow?
ok
FLOW has been working horribly for a week now and they're not giving any explanations
Is nano banana down?
on Flow?
Ya, every generate is failing
yeah, same for me. most of them atleast
Just noticed Google docs now have "premium" features that require a Google One subscription, or a Google Workspace..
reminder, general is not the place to ask questions about products as we have #flow-general・🌊 , #whisk-general・💭 , #jules-general・🦑 , #opal-general・✨ and many more channels where your product specific questions can be asked. Thank you
Did you have a paid account before?
You're not supposed to, but the filters are mostly about stopping people from using images with other real people, so if it's stock images or something it might get through.
Had something almost crash on me yesterday
hi
Hi. I am looking to get access to the disco channel. I have done all the steps in the emails and am currently using the product, so something is right!
Anyone has tips when using reference images so the models don't lose their creativity? It seems when I put a reference image it's hard for the model to keep the style but still be creative in the output's structure and elements. Even if I say things like "only use the reference image as a color palette inspiration" it will introduce elements from the reference image into the outputs
Do literally try that then. Be specific on what kind of creativity you want and put in the prompt tho, for example
"Freely creatively with all creativity freedom apply any professional coherent art style, composition and shading"
Shading does wonders with whatever.
I like shading like this
[Shading: Best professional shading style available with depth cues correct physics realistic perspective scale proportions, and spatial layout.]
Using terms alike "professional" "Expert" "best rated" will cause the model to use what knows it's best.
And yeah also ending the prompt with stuff alike
[Image quality: Remastered 4K HD ultra high resolution, sharpest definition, denoised, unblurry, no jpg artifacts, clean lines, High contrast.]
Hallo zusammen
Is good idea, you might do best in finding your own pace and style tho.
Sometimes that can be a litle dark in Nano Banana 2 compared to original Nano Banana, so you can also add an illumination slot if necessary, "Clear natural light" "Well illuminated" "No dark areas" etc.
A specific hour works too for natural lighting
Yeah everyone's got their favorites, you'll know when you do
Also the opposite works here, if you don't like how creative the model is being with the characters or whatever add "exactly without any creativity freedom" and it will obey the reference image more unless it's conflicting with other prompt.
The prompts set at the top will always take more priority.
Kamalu: Legend of Amadioha | Episode 1 ⚡
The God of Thunder, Amadioha, has been reborn, and not everyone is happy about it. As Kalu grapples with his new identity, a dark threat from the past looms over his future.
Kamalu: Legend of Amadioha | ⚡️
What happens when an ancient legend becomes your reality?
Meet Kalu, a young man discovering extraordinary powers hidden within him. But his journey didn’t start here. We take you 15 years back to Ozuzu Village, where a powerful prophecy was born.
The God of Thunder, Amadioha, has been reborn, and not ever...
Is there a separate discord channel for Gemini CLI and Gemini Code Assist?
Hi Micheal, we don't have a channel here in Labs for Gemini CLI and Code Assist. Gemini in fact has their own dedicated server here's the link to that server. https://discord.gg/gemini
That's awesome... but looks more like Seedance
is there a support channel for gemini/veo related bug reporting?
Hello, I'm looking for help with getting access to Disco. I've signed the agreement and gotten a (very shady looking) email for installing Disco, but the installer is not working on my Mac. 🧵
morning mate
Good morning
Morning at 10:50?
I am in the US, it’s late night to early morning for most of the country
West Coast it’s 11 PM, Monday. East Coast it’s 2 AM, Tuesday.
That’s obviously not including Alaska, Hawaii, or any far off US territories like Guam
I think there’s a moth in here damnit
Well, I live in the South Asia
Well also my sleep schedule is screwed up since college for a long time especially given the type of work I’ve done or do
I see...
HIII
Hey guys, is anyone interested in FEP and Active Inference? Why wouldn't it work, even though the theory is so elegant?
Where is help channel
Does Antigravity not have seperate channel?
They should really have a Antigravity channel.
When IDE starts writing code directly even tho it's in planning mode. Is it model fault or IDE fault?
hello
Hello
do what?
work for 3 hour 20 mins without any input
also this isnt even gemini 3.1 pro
this is 3 flash
Hello everyone
Hello everyone, is everyone well?
I have here a video and I need your help on how I can improve the level of quality in this video. Do you think that it gives signals of AI? Do you think that it's a problem with my image, my first image frame, or is it based on the video generation variables or issues that maybe I'm not seeing but hopefully that you'll be able to see things and identify for me to really make it way more realistic?
Actually it was all Kling
antigravity is very bad
But it's just inspiration right? Because I want to generate a shoe based on other shoe brands like Nike and Adidas. It's just a custom brand. It's not going to be a real brand. Just a programming project
I know there’s heavy traffic, but what doesn’t make sense is that I had Claude working for 14 minutes, then got that error which messed up everything it coded, forcing me to roll back to my save and use up my credits.
It would be easier to use Antigravity if you know some coding basics, well, based on the coding languages you're using
I've made pretty much everything, i just hate waiting 15 minutes to get a high traffic error in mid generation
Hi everyone
The first emotion looks unusual
You ca tell it's AI
Thanks I will take this into account
Hey everyone, I'm trying to create some cool sci-fi type images that might've looked like something from a movie scene and my generations keep looking like comic book images. Anyone have any tips or resource guides on how to make futuristic images look real?
For example I'm trying to do a human versus aliens battle scene
get references
get reference on how you want it to look, and tell the AI to take inspiration from that reference
Thanks, I am using references but it may be how I am prompting it to understand to provide photo quality. But I have told it specific things like shoot it as if it's coming from the film camera. But it's still all looks like cartoon.
Is there something specific you say with regards to the reference that gets it to look like a real photo?
I don't know anything else other than a reference. If it's not the reference, then you should take a look at your prompt
hi
Yo https://labs.google/fx/tools/flow not working right now right?
Flow is an AI filmmaking tool that lets you seamlessly create cinematic clips and scenes using Google DeepMind's most capable generative video model, Veo
Probably Google Developer server
How'd you trigger that?
I just told you what the rules are. I don't make them.
Branded products are also against the rules... I think. Anything that's intellectual property.
Lower the quality. Give it a film grain. Don't make the face so close to the camera.
Share your prompts.
I gave it a huge plan really big one
And have all permission
Might be sitting on something pretty big
Potentially found a way to solve a massive thermal wall
Let’s steer clear of sensitive or potentially divisive topics, this includes all mentions of politics and religion. We aim to keep this space welcoming and comfortable for all ages and backgrounds
hi
Thermal wall im looking into has millimeter Kelvin environment while a pulsed laser hits a substrate with 1350c in peak power
Math so far works
Your face looks familiar
Dudee what is veo 3.1 lite
give me seedance 2 destroyer
is the winner gng working hard on veo 4 in google?
I have a lot of inputs, i use veo and NB every day for hours
We need Veo 4 🔥🔥🔥
Google is due for an update 👀👀👀
Perfect opportunity since SORA is going down
Yeah. There hasn't been an update in a long time 🕒
I feel like i don't get very much quota on antigravity. Can easily blow the weekly quota doing a couple tasks. The flash model seems to provide more usaged, but i barely use it due to the fact it tends to loop and get stuck on itself constantly. Anyone else find this?
They might have lowered the rate limit because their servers can't keep up.
It's possible. I believe this is my second week now trying things out, last week felt the same as this week. Comparatively to something like codex i feel like i get less value. I do jump around a lot between different providers to not be reliant on a single one, but it would be nice to kind of stick with one eventually. 😅 ( should clarify, less value for the coding specifically, googles plan has more value overall as it includes other things, but i dont necessarily use those, but i do appreciate the ones i do use! )
nice to see ya RoRo!
since sora is gone, i am going back to the other models. I was a top creator on the platform. I still have my old space video from veo
planning to upload around the day NASA is flying around the moon
plus the main tool i was using to use nano banana and other tools were removed, i forgot it was on flow
ah.. imagen 4, good times
Flow app for iPhone would be peak
Gemini app only allows 3 vid generations a day even with google 1 pro
I thought Antigravity was free.
PS - Oh, I see. It's just rolled into AI Pro. They should roll it into G Cloud instead.
"What if we had an even worse model??"
Just got the email from Google about the tiers and I have no idea what it means. I pay for the service. Does this mean I have to pay more?
Lol noone knows.
Haalu
I too got it! They are just putting some limits on the spending so you don't empty your coffers
thats only if you use gemini models via the API or through ai studio. if you use it through https://gemini.google.com/ then it wont affect you
is seedance 2 worse than what we have now?
good day mate
Yo
Yo
hello
Hey Everyone,
I’m a college student and I run a small tech community where we’re learning development and building beginner projects together.
We’re looking for someone experienced who can occasionally guide us—like reviewing projects or giving advice when we’re stuck. It wouldn’t take much of your time, even a short session or occasional feedback would mean a lot.
Would you be open to being a mentor or guiding us in some way?
Thanks for your time 🙌
yo
hi
first and last frame, what a wonderful addition.
Hi @calm hatch I am a Senior AI Business Analyst with experience in application development and agentic AI solutions. Please let me know if I can contribute or support any of your projects. I would be happy to help.
@twin raft haha, where are you from?
from china
Nice i love China. i want to travel. which city are you from?
Yo
beijing, welcome
Just need some guidance and mentorship for our beginners members and us too we believe in instead of just learning. We'll show you our project whenever you have time can interact with us and guide fixes or any other improvements in our projects and learnings
I'm talking about Veo
💯 🥺 😭 😔
@eager talonHi, just a friendly reminder to keep your messages relevant and avoid spam, or unclear content. We’ve removed the recent messages that didn’t comply. Thanks for helping keep the chat pleasant for everyone.



Hm
What's the trouble
You can try use seedance... they are awesome
what the hell is going on with antigravity
dude it used up my quota and i didnt do anything, chats dissapear and its a total total mess?
where can i report this?
Use ai lol who codes 🤔 now in this age
I have been :) been following it since day one
Hi, quick check before I share anything publicly here.
I recently open-sourced a project around reducing the cost of high-volume LLM classification workflows by offloading simple cases to a lightweight model under a quality guarantee.
I saw the rule about no advertising / no call-to-actions, so I wanted to ask first whether it would be okay to share it somewhere appropriate for feedback and discussion, or alternatively only by DM with people who are explicitly interested.
I wanted to make sure I’m respecting the server rules before posting anything, happy to delete this post if needed.
Thanks a lot.
How can you even get access to the site?
wdym
nanobananan 2 is down again? many fails generations
every day now nanobanana 2 is down
Expensive, but for realism defintely the best right now.
All the big AI companies aren't able to keep up with demand, so for a few hours most work days they either have to limit use, or fail out generations.
Still many failed generation, do you agree?
It's worst during the work day M-F in the EU and US, when the most people using compute-heavy models are trying to use them.
hola, alguien algun latino que sepa utilizar flow? veo 3
uhh its not working the ai video generation veo 3.1 fast
when i try generating with a reference image
it works without a reference image weirdly
guys veo 3.1lite is at 720p but the upscale version look good and we save half of our credits
hi, this is the #general・🧺 channel and is not a place for formal support. please submit a ticket in #1238621511401734254 or join Flow conversations in #flow-general・🌊.
Hello
please use #1238621511401734254 to submit any bugs you find!
Is it here that we'll get the reording?
Still waiting on access to the Disco channel. Admin please help?
Nice bugs 😁
Masterclass perfected ☕️✌️
is today's masterclass available for download somewhere?
ill take just a stream cause I had other stuff going on and couldn't make it.
thanks for ai God bless Ai
Anyone have recommendations using Google AI as a graduate student?
Like with making detailed study guides
So any word on the Antigravity channel yet?
Hey
Hello
What did you mean when you said "ai God bless Ai"?
Is that the name of a AI agent? Or you just said god bless that we have AI?
They have a new tool called 'Learn Your Way' where you can upload textbooks and basically chat with a bot that's built to help you understand the content in that textbook: https://learnyourway.withgoogle.com/
I was also wondering. I couldn't make it in time because of work 🥲
😤
I am making tinder for food, but I don't know how to upload a picture of the food for cheap any recommendation on api keys I should purchase or anything else I should do, the thing is the food is randomly generated so even I don't know whats' next so I need my ai to upload a corresponding image
what is an api key that is very cheap to generate images?
Hi
yo
hello
good morning guys
HII GUYZZZ I AM NEW TO THIS WORLD OF AI thingy and heard a lot about API KEYS. CAN SOMEONE TELL ME WHAT THESE ARE PZZ
I am shocked by how low the quality and intelligence of Gemini 3.1 Pro and Flash are. It cannot handle a single task of my physics work. Every single task is a failure.
For example, I'm using Antigravity. I provide my 2000 line academic paper to Gemini, asking for a rewrite with corrections and reordered sections.
- It outputs a 300 line paper.
I ask again, to create 20 different files, for each Section or sub-section of the paper.
- It creates a bunch of 30 line files, totalling to 600 lines. 75% of equations and materials are deleted.
This is the worst AI model that I ever used. It's far below any Chinese open weight model
It's total complete failure at every turn, at every request
This is a literal joke. 16 lines? 22 lines?
I provided explicit instructions strongly emphasizing not to delete or remove materials
... an Appendix with 1 entry.
I've never seen any AI model that is this pathetic except for some 8 billion parameter stuff. Not exaggerating.
The poor AI model felt bad about it, and created me a 5 line Discussion and a 3 line Conclusion section. Wow
How is this even real
It can't even create a 200 line markdown file. How is this real?
Google Gemini Pro struggles to write 40 lines. How is this possible?
I can write a 3000 line document with Claude
i use gemini to code in unity and it work fine. we just have to goto instruction gemini and fill what we want gemi to do
Gemini = TEMU LLM
have you tried breaking down the tasks so its more manageable in one session?
I literally did that. I tried making it split my academic paper up into 35 separate files, just so it would write more text. It still failed.
It can't handle it. Every generation it attempts to write the paper, 50% of the materials are missing, randomly
I give it to a Chinese AI, I get 1:1 perfect reproduction with every line in tact, everything preserved, and all corrections made
Gemini Pro intelligence is too low, so I gave it to Z.ai GLM -- it finished the task in a single request. Perfect. No wasted time
I wasted 4 hours with Gemini, battling to even get a single Section of my paper correct
Hi guys!
Guys, when I try to use the Nano Banana Pro, this message appears: Failure! You are requesting generations too quickly. Please wait a moment and try again.
However, when I use the Nano Banana 2, it works normally!
Can someone explain this to me?
Hello~ How to create an image with Nano banana in discord?
Hey everyone! Just a friendly reminder to keep all conversations in the supported language for this server. This helps our entire community engage effectively, and we've gone ahead and deleted the recent posts that didn't comply. Thanks for your cooperation!
Hello, I understand. Sorry
How to install Google
Hey, quick question.
Has anyone here used antigravity.google?
Is it fully free right now or are there limits?
How fast do you hit the free quota?
Can it build full apps from one prompt or do you need to guide it step by step?
What kind of projects work best without hitting limits?
I’m new to coding and want to build simple apps or games, so I’m trying to understand how far the free version goes.
I use it most of the time. I pair it with a Google AI Ultra subscription, so I've not hit any limits. Not sure about the free version. Biggest benefit is access to Frontier models like Opus 4.6 and Gemini 3.1.
I use Microsoft edge but not Google because i dont know How to install it
But there is a free plan
You can use one on Edge?
Is anyone on the Google AI Ultra plan experiencing errors when generating images and even videos?
same, throughout this week!
It has been pretty reliable for me this week. I run a lot of stuff straight from Antigravity though.
has anyone seen those wizardcore videos or the dripwarts videos?
how are their movements so smooth, any idea what video model they are using?
You cant.
Google is a website.
Hey guys, Google just released their new Gemma 4 models and you can run them locally now! https://x.com/UnslothAI/status/2039739190536286313
Google releases Gemma 4. ✨
Gemma 4 introduces 4 models: E2B, E4B, 26B-A4B, 31B.
The multimodal reasoning models are under Apache 2.0.
Run E2B and E4B on ~6GB RAM, and on phones.
Run 26B-A4B and 31B on ~18GB.
GGUFs: https://t.co/fpX21yWbge
Guide: https://t.co/odRo9WjRpA

hello
Hi fam, I would like to create an animal avatar and do different clips of it, using always the same avatar. Is this posible with Google?
Yes.
ok
anyone else experiencing internal errors on gemini flash in aistudio?
it works fine with just regular generations but with image uploads I keep getting this :/
what would be the best place to report concerns regarding gemma 4 tokenizer integrity? i have reason to assume its "broken"
Is this an appropriate discord to ask questions about Antigravity? I've been getting the worst infinite loops with development lately 🙁
hey! we don’t have an official Antigravity space atm. there’s a vibe coding channel in the Devs community where some folks are discussing it! discord.gg/google-dev-community
The answer to this yesterday was no.
It changed to yes just today.
Depends on how many videos you want to make, where you live and how long those videos are.
Using google gemini app is so bad, why don't they have option to disable search when answering query? the internet already being dominated by AI and a lot of creator even the big one use AI to write their report or blog, it make it worse for gemini than anything
How to get google ai ultra users role🙂
Im suing a drunk driver
can anyone over here help me build a roblox game?
Can anyone help me get access to Google Disco?
you should be able to purchase yearly google ai ultra
which tool shall i use which use my voice real time making generative video
please don't turn off the google whisk
it is much faster than FLOW in terms of image generations
very good morning guys
Hello friends this selebrate
Are they shutting down whisk??
Whisk is generating images in seconds, when flow takes much longer and 90% generations is failing.
Hello everyone
hey guys, not sure if thats the correct channel, but has anyone else problems with antigravity currently? endless loading circle, no answers
restart pc, log out and log in again
For me, disabling my AV helped for some reason
your AV?
hi guys👋
is there a support chanel for antigravity..
Hey, we don’t have an official Antigravity server, but there’s a vibe coding channel in the Devs community where some folks are discussing it! Keep in mind that it’s only community support from other users! discord.gg/google-dev-community
Hello guys
Sure, but take into account that Flow uses much better and heavier models for generation
hi
the computing will probably go to Whisk
i mean flow
Hi
did anyone release the uncensored version of gemma 4 on huggingface or something yet?
Can someone help me find the recording of the webinar about flow please?
Hi, quick question. Has anyone here tried Google Antigravity?
Is it free to use?
How do you actually use it properly?
Is it something you can build with or is it more of a demo?
Are there any limits I should know about, like usage caps or restrictions?
If you’ve used it, a simple explanation would help a lot.
its actually pretty good
its like vs code insiders + agents
currently useing antigravity it deffenatly fills in the gaps. one thing i notice is too much automony causes it to drift. what i find is a pre thought out instruction specifing the exact paramiters before you allow it to input work flows, is the best way to utilize this capability
There was a problem sending a verification code to [Phone number]. Please enter a different phone number.
anyone know how to fix this?
yeahhh if you completely let it go without a proper implementation plan it drifts away
also i will recommend not to overuse it
instead use it like vs code insiders
right! and iof you want to rebuild a project just sit there keep hitting buttons and you will have no choice!🙃
So, it's NOT OK to ask Flow to create a couple in bed, kissing. But it' IS OK for them to shoot each other with guns and slash them with knives. Got it. What's wrong with this world? Here's a clue.
could you make two seperate images and edit parts of one image into the other possibly multiple edit like two or three would it catch on!?🙃
one is called "innapropriate" one is innaproprate
mb for spelling
i once recreated the rick and morty scene where they eat a person's intestines as spaggeti and it didnt restrict it
does anyone try google app search yet? This is what it look like on windows same amount as a normal browser damn...
its kinda good for analysis
yea kinda good, finally we have some good decent raycast like (50%) on mac for windows
Seedance 2.0 will change things.
The iphone realism + physics here...
One-shotted this:
+ Used omni-reference.
+ (2) images from an amazon listing.
+ (1) Prompt detailed scenes...cuts...and timing.
+ Very specific language on the hook visual copy + pacing.
(1) Prompt - (1)
this is crazy if actually 100% ai
i need veo 4 to be better
Only thing that gives it away is the hose hookup, but it's otherwise super convincing
I'm sure it will be much different in quality. Given that seedance was trained illegally on clips that they can't even release the model publicly for legal issues. While veo 3.1 has its own quality. I'm expecting some improvements on its own style
Especially if we get less hallucinations in Veo 4, would be great
dude there is a specific problem with veo that makes me wanna cry
idk how to explain but here in the starting scene can you see the carpet being generated again and again or something weird happening with it
i dont really know how to explain what that is
even on the tiles and wood, and in general there is this weird grain glitch or something that happens maybe some google employee can explain
throbbing, shifting, temporal inconsistency
texture crawling whatever its called
Waiting for disco channel access, admin #help

Its like weird lighting shifts/increases. Thats odd, looks like were getting static electricity in flow before GTA VI.
Hello everyone, this is the only server on the discover page on Discord that I could find that it a bit more relevant to my question.
I am trying to set up a Google Business Profile for a client, and for HVAC/Air Conditoning businesses you need to submit a Business video which contains stuff like branded vehicles being unlocked, the area etc. to verify that you are a real business.
Well, this is the second time that the account has been suspended.
It says the following:
Your business profile has been suspended due to a policy violation. Content that violates our policies on deceptive content and behavior isn't allowed. Deceptive content intentionally misleads or deceives others. Your content was removed by our automated moderation system. Routing ID: DPNB
The business name has no keywords
The business category is correct
Should I just create a new new google business profile.. or what should I do 😭
Oii
Flow needs a way to keep using the same voice so when we make a movie or scenes it keeps the dialogue easier to work with. If the voice is already in the system should be a way to keep using the same one no?
i have a question
is google staff in this server really hired people working at Google?
and if so, is there a dedicated sector of google employees that work in this server?
yes! everyone with the Googler role works at Google. though they may not all work on the labs projects!
wow really? its amazing Google invests in servers and communities
Flow needs a way to keep using the same voice so when we make a movie or scenes it keeps the dialogue easier to work with. If the voice is already in the system should be a way to keep using the same one no?
Someone from flow labs please pay attention to this you guys really need to fix this. We need consistency in scenes. And more than 8 seconds like 15 or 30 seconds
What’s the point of making a great scene and then the voice changes in the next scene?
That would be frustrated
Sometimes you get a really good voice but then keeping it in more scenes is the problem. I don’t know why they can’t nip that in the bud and we pick a voice and stay with it through scenes.
would anyone be down to call on discord i would really apperciate any advice on my app that I am trying to build
bro it's genunly so hard to get someone to tell me about my app

Gm
very good morning zoe
Hello guys
@jaunty arch
I got a notification about it, but when I went to Grok there was no switch to turn on quality mode. Either mid-rollout or only for the higher end account levels.
ah damn
i really really want veo4 to be as good of an upgrade like NBP
how is everybody?
Hello
Hello 🪺
Helu helu
hello
Hello
Wssp
Wssp
Hello
Hello Everyone 
hi
hi
Gm
hello how many are in hackathon of gemma in kaggle
Hello from Poland!
Hello 👋
Heu
stitch can be export to a proper app right?
hi there people , can I ask for some help on the scenebuilder tool?
hi
Hello
hi
hi why mi scenebuilder doesn't have got the info to expand , or the plus icon to add the picture
hi
@rustic shoal what happened?
Hello, I wanted to know which subscription you recommend, Pro or Plus, if my intention is to have more images and higher quality, especially from nano banana pro.
They moved the expand fuction. Click on the clip you want to expand in the grid, and it's there, under the prompt box.
Hi
hi
why is it whisk is going under?
Ya the new design is trash
nah thats not what i mean, i asked stitch to change the font of a button, 10 times, and it just wont do it, i then asked it to take a screenshot mockup of a page i made, lets say "support page" asked it to take the mockup and turn it into a real page, but instead it did a whole landing page with cars and galaxies 😂
elloo community
Google Labs projects usually inform future production apps, but not in the form they start. VideoFX was killed off, and we got Flow (much better). I don't think Whisk ever caught on enough, even though it's been out for a long time and was free, but Nano Banana exploded onto the scene, and is wildly popular. Hopefully elements of Whisk will get incorporated into future versions of Gemini, Flow or other image apps they might roll out, but we'll see.
Hello, everyone!
see for me, Whisk's main draw was how precise it's image editor could be with prompts
im scared to lose that
huh?
Whisk just uses the original Nano Banana (Gemini 2.5 Flash image) model for that precise mode, that they did have at Flow before and they removed for no explained reason. You can still use it at Google Ai studio and wherever they'd have it so you get used to what you need to have to continue your projects.
They really won't listen to all the negative user feedback about taking Whisk offline or at least keeping the model used for precise mode at Flow, so I at least already threw the towel.
And yes like I still drop by sometimes to check and Whisk's general area is still full of complaints and confusion about it. Listening to what users want is like a core thing for feedback to really work.
It’s a bit disingenuous and highly illogical to assume people don’t want to use the original Nano Banana just because it wasn’t being used much at Flow before, when it's actually the main reason Whisk was being used and people thought they were being brought over for— just to find out the same type of image generation wasn't available and to constantly complain about it, whether they realized what's going on or not.
It's not the original Nano Banana. It was out before the first Nano Banana came out, and ran on Imagen.
More importantly: That's not what anyone I've seen has said. If you want to see the popularity of the two, just use a social monitoring tool, or Google Trends. We have no idea what the next updates to Flow will look like, or if they're going to roll out an app designed more for image generation than video like Flow.
You're right about imagen being the base for Nano Models, which is why when you run out of images for Nano Banana 2 at Flow you also run out of imagen 4 images. In the same way, the specific imagen 3 variation Whisk uses is indeed called Nano Banana Gemini 2.5 flash image.
You can only know this if you get an exact same prompt that you ran at Whisk, then you take the prompt to Google AI studio and run it with it's original Nano Banana, you'd get the exact same image. We've already been through with this challenge with someone else here.
Have you ever actually done this?
And yes this is not what people have said, but Whisk users at least have indeed said they want Whisk's precise mode like user above just did, which HAS repeatedly been a feature request for Flow. And the way to get "precise mode" even anywhere else, is simply just to use original Nano because they're the exact same thing and there'd be no other way to do it. So they just don't know that's what they're asking for.
So I'm just gonna repost the same test, feel free to reproduce it.
huh?
Here's the prompt
Ok again here's the prompt at Whisk
Here's the Whisk result with the precise mode on.
And here's the same prompt at Google AI studio with Nano Banana (Gemini 2.5 flash image) selected at the right. Giving the same image quality result and even the same poses.
These have the Gemini watermark and the Synth ID so they surely come from Google AI studio.
And this is what Flow's Nano Banana 2 does with the same prompt. It's similar but in almost all the images it ignores the "vintage" prompt that should be causing the image to look like old anime made in acetate nitrate anime film.
Which Whisk's/Google AI studio's original Nano Banana did get correctly with the vintage filter in all the images, while Nano Banana 2 instead of vintage made them look like new digital art anyone coulda made yesterday.
Anyway if anyone has more doubts about Whisk's precise mode being just original Nano Banana (Gemini 2.5 flash image) be my guest, run a prompt you're familiar with at Whisk in precise mode that you know generates the same images you're familiar with, then run at it Google AI studio with that model selected. You will get the same image from the same model Whisk is using.
And if you like Whisk's precise mode images, that's an option without a doubt after Whisk is gone, whatever popularity amount each thing has.
Also as a bonus, I do have witnesses that while original Nano Banana/NB1 was enabled at Flow it was generating the exact same image quality as Whisk. It's just really few people had chance to try it before it was arbitrarily taken down.
And as seen there keeping the model would have been the ideal solution for people wanting to generate the same image quality from Whisk at Flow since it WAS already working.
Then again gotta be using Whisk and the original Nano Model (specially at other places besides Whisk) a lot to notice most of this cause it can't be seen just from reading docs... and the confusion is just rampant.
Visual cues is the only way to get it.
So once again, if people complain that Flow doesn’t have the same image quality as Whisk—or lacks features like precise mode and ‘seed,’ among other things, it's because most users didn’t even know what model they were using at Whisk, since Whisk never showed the model name, so obviously almost everyone doesn't know what they are asking for.... but it's still the exact same thing as asking for original Nano Banana.
Because tbh I don’t see in what other way Flow could have those features unless using that same model... so yeah that's also why it isn't happening.
good day guys
Same.
Let's do a vc
Guys tell me where are you from
Also I did just listen and went to do some metrics too at the sites people usually go to... and yeah if this is right then I'm Julio Iglesias. It's believable NB2 would be so high there, but that GPT is higher just sent all their credibility to the dust. I mean GPT is good for data but not for images. And the amount of votes NB1 still has is insane too, something here's definitely not right.
They're visibly pushing some models more than others, Google's metrics are similar, NB2 is above by half at least over NB1.
But then again Google has been removing NB1 from many places so how is that fair. Let's ask someone impartial.
Thiiiiss is the part.
Of course if people are being literally forced to get used to NB2, it will be used.
But enough people keep saying daily they don't like it
Sometimes it's just about how you ask things
And that's the actual truth tbh
Then again it's not really about competition, it's about understanding both models are tools used for different things, and that's really what AI is saying. NB1 absolutely is more artistical than technical as NB2 is.
Will I get banned if I ask how to get rid of the Synth ID?
ig dude rlly got banned from discord themselves
he didnt lol
is there a separate channel for antigravity
Why is Google VEO adding watermarks to my videos on an ultra subscription??
What's the point of ultra when I can't remove the veo watermark.....
I'm just asking for confirmation
I'm still learning
Is opal included in the freebies?
Thank you, I have tried many in the labs
Hii
noys


