flow-creative-challenges・🎨

[golden bronze]

Welcome to #flow-creative-challenges・🎨

Each week, we'll post a prompt here to inspire your next Flow generation. share your gen here in this channel where you may be featured on Google Labs socials, and stay tuned for other surprises too 🙂

📣 **Please note: **we will share your content across Google Labs socials only if you've opted-in to content sharing by reacting to our message in #start-here・📍.

[golden bronze]

This week's prompt is: Consistency

We're providing two ingredients for our very first Flow prompt: featuring our iconic Otter 🦦 and Snail 🐌 duo.

In filmmaking, consistency is key to a strong narrative. Using Flow, you can build stories with consistent characters, scenes and objects across multiple clips, giving you more creative control than ever. When a film lacks consistency, the audience notices and the story falls apart. In AI filmmaking, consistent characters are a huge challenge. Flow makes this simple with Ingredients to Video, which is now available for both Ultra AND Pro users!

The Prompt:

using Flow, select "Ingredients to Video" and upload both the provided Otter & Snail images. from there, create a video focusing on consistency, whether that be the scene or situation - the rest is up to you.

Share outputs directly in this channel (similar to our today's special channel ).

Questions? Ask in the below thread!

[gleaming lance]

Veo3 not suported two ingredients and voice

[dapper idol]

Yes only supported in Veo2

[gleaming lance]

Flow is not a consistint tool i am using it since 3 months ultra version but the output is not as good compare to runway leonardo pixer or other ai video creation tools and the app ui is also not intractiv

[grand shard]

"Disco Night! 🪩"

[zenith marsh]

I didnt expect the ingredients to video feature can blend so seamlessly with another ingredient that uses a completely different art style

[golden bronze]

thanks for sharing this 🙂 and for giving it a try! 🦦 🐌

[storm granite]

Late entry. 👁️ 🐌 :otter

I love the screaming faces on the shell… 🐚

[dire steeple]

Hi guys, can I change the voice for Google Flow? If not, do I need to tweak it into the prompt on what sort of voice I want?

[grand shard]

You can prompt for the style of voice and dialogue, but you don’t have total control yet. Audio in Veo 3 is in beta, so expect to have more control soon.

[zenith marsh]

Wait I just noticed that now that you mention that ahhaha

[hearty echo]

Before you start with the challenges, I suggest you first restore the ability to generate people based on realistic photos, because few people are interested in generating animals! You're mocking users who spend money on yours AI!

[zenith marsh]

🏄‍♂️

with sound!

[zenith marsh]

The workflow was basically: generate a video with Veo 2 Ingredients to Video → drop that video into Scenebuilder → hit “Save Frame as Asset” → then throw that frame into Veo 3 Frame to Video… boom, Veo 3 Ingredients to Video 🙂

[hearty echo]

It's true!

[vernal wharf]

Hey guys 👋🏼 . I'm exploring the server. what is this channel about?

[hearty echo]

lol

Can you add photos of realistic people?

Can you add photos of realistic people?

[zenith marsh]

Anyway move to #flow-general・🌊

[hearty echo]

This isn't about celebrity photos; it's about photos of ordinary people, generated by, for example, AI, who don't exist. People have huge problems with this. Check the section where flow issues are reported; there are already dozens of comments there. Can you give me examples of countries where this is possible?

[fiery prawn]

I spent 110 credits and 20 minutes on this and it shows! Lol fun though... back to work

[dusk rain]

Guys do u know why It stop at 40-50% of creating and It does a horrible work?? I mean some time ago It was better and It was "creating until 99%

[dapper idol]

How did you get the voices on there using V2 ingredients?

[neat moon]

what's up?

[hallow wadi]

https://cdn.discordapp.com/attachments/1331469853646651403/1399632940790579231/Whisk_gif_gzmymq2mdy.gif

[jagged sapphire]

https://vm.tiktok.com/ZNdXxNkjh/

[main raft]

the hardest thing with Flow is to generate a fighting scene. and that takes a lot of credits(points) to get at least some decent result for each scene. who agrees? https://www.youtube.com/watch?v=xEaPcX4BYwQ

[fathom estuary]

How did you get that cool flying effect at 00:48 ? ive been making superhero ones too with fights , its always messy , and most lack the power of the punches , the glitching

[violet zinc]

waoo pretty cool.. will be very curious about the prompt 😉

My first try 😉

[main raft]

exactly, the worst think about flow is the fighting and flying scenes, i didn't get that flying scene in my first try, i had to try 4 times with the same prompt, 1 and 2 and 3 were messy but final one was decent but not like how i imagned.

[main raft]

nice scenes, 👍

[main raft]

nice start, not bad as a first try

[fathom estuary]

Trying to get him to use the speedforce to vanish from the camera , but instead it made it look like the reverse flash has beef with him:P

[granite dove]

Sorry for the spam, my apologies for sending the same videos across multiple channels.

[hard canopy]

https://www.tiktok.com/@glopezs25/video/7541978103988456710?is_from_webapp=1&sender_device=pc&web_id=7541977299814778373

[toxic crystal]

What is the challenge?

[jagged sapphire]

https://vm.tiktok.com/ZNdXsb1wY/

[static crescent]

https://www.youtube.com/watch?v=IyTs-pF9qdQ

[balmy holly]

https://youtu.be/KueOdjML3PQ - It was so hard to get characters to turn around and also maintain consistency in new scenes that they haven't been in before. Trying to get it to happen was making dollars fly our the window.

[thick ivy]

This is one of my favorite Veo3 prompts from FlowTv. Input otter instead of DJ at the start. The song was created in Udio. I created custom lyrics to match the scene.

[main raft]

u have to chill mate, and have something to do for yourself 😉 all ur comments are 🥱

[golden bronze]

last week's challenge is pinned 📌 you can also find it here: #flow-creative-challenges・🎨 message

we'll be posting a new challenge shortly.

[toxic crystal]

Cool, I did that with pip...should have posted it here I guess....

[golden bronze]

if you have gens with Flow that aren't related to the challenge set in this channel, please share them in #flow-videos・📹

[golden bronze]

you still can!!

[golden bronze]

This week's prompt is: Defying Gravity 🕴️

welcome to the 2nd Flow Creative Challenge 🙂 we're not including any ingredients this week, so take this prompt in whatever way you see fit - just don't forget the server rules & guidelines.

The Prompt:

using Flow, create a sequence where objects defy gravity.🕴️

📣 Interested in being featured on Labs socials? Be sure to opt-in to our content sharing message here: #start-here・📍 message. We will only share content from members who have opted-in. Not all submissions will be featured.

[grand shard]

"Flying Dwarfs"

[toxic crystal]

Hell yeah! That is really dope.

[rich elm]

Garden

[toxic crystal]

@golden bronze (sorry just tagging cause of convo earlier) This is Pip.

[graceful karma]

He defies anatomy 🤣

[ember moth]

One with 3 people = GPT

[granite dove]

What time is it where you live?

[ember moth]

2pm

[granite dove]

the result will come out as soon as at least one person gets it right

we already have 2 votes for ChatGPT with 3 people

[fallen vortex]

Fighting gravity and defeating gravity.

[granite dove]

thank you very much

be happy ☺️

[fathom estuary]

A little animated Superman clip

[jolly shuttle]

Hi

[faint fossil]

How do you make the video last twice as long? Thanks.

[graceful karma]

Multiple videos concatenated in Flow's scene builder.

[warm depot]

Made multiple commercials for a game I play. I originally posted this on our discord.

[sweet elm]

oi

hello

How do I generate a video? Can anyone help me?

[stable flicker]

You go to labs.google then click flow and put your prompt

[obsidian warren]

/Only a man is singing, a woman holding the microphone and moving in rhythm with the music, on the street, people are listening to music, and a camera from above

[civic shoal]

The Rising Arch District in Phoenix City (Promo 1 - Concept) - The Multiverse VR
Google Flow (5 generations) simple 1 line prompts, not in JSON format.

[granite wedge]

Hope this meets the criteria I'm new to all this

[tiny dawn]

Or gif?

[sudden ember]

I'm new to this creative challenges!

I make a plate of cake jump. Does it has to be gravity for a plate of cake?

[acoustic blade]

I am doing animation videos and use veo 2. I use veo 2 b/c I don't want any mouth movement. No matter what I prompt, 8 times out of 10 I get mouth movement and am losing precious (expensive) credits. Does anyone know how to resolve this? Or at least get credits back for poor outcomes? Help.

[gentle sail]

Defying Gravity

[uneven crest]

The Rising Arch District in Phoenix City

[granite dove]

👀

[surreal hearth]

I received an error mssg on MY Creation: "Failed Generation :against company policy to recreate prominent people." This is MY own creation. I sent the error message to the "send feedback link." Is there anyone who's able to straighten this out?? Why would it give me this message??
PLEASE HELP!

[granite dove]

No, yes, all the creations are ours, but the creations were made using other policies, you know? You need to look at your creation and see which policy that was.😅

[granite dove]

For example, I can create a completely original story with girls using Spider-Man.😅

[granite dove]

can I see the prompt?

[surreal hearth]

There's no prominent person in it whatsoever. This person I created has never existed. I took a photo of myself and used it to create the person, so there is no prominent person at all.

[granite dove]

Okay, but you just said you used a photo of yourself as a reference, right? And when you use a photo of yourself as a reference, that adds a human aspect to it, you see? That’s where this is coming from.

[surreal hearth]

{ "scene": { "character": { "description": "A beautiful woman in her mid-30s with a sparkling caramel complexion, almond-shaped hazel eyes, and high cheekbones. She radiates warmth and confidence. Her long, smooth brown hair flows in soft waves with subtle blonde highlights. She wears a bold red blouse symbolizing passion, strength, and empowerment. Gold hoop earrings frame her face with sophistication. She sits on an elegant cream-colored chair with gold trim, projecting a regal yet approachable vibe.", "props": { "microphone": "A microphone with a striking red foam cover, complementing her outfit and emphasizing her role as a powerful speaker.", "desk": "Desk with neatly stacked books signaling wisdom and growth." } }, "background": { "description": "A modern, chic, and feminine studio with warm lighting, white walls, and shelves decorated with touches of gold, red, and neutral tones. Potted plants add freshness and life, balancing professionalism with comfort. The atmosphere feels inviting yet empowering—ideal for conversations about self-worth, healing, and personal empowerment." }, "audio": { "voiceover": { "tone": "Warm, confident, and uplifting", "script": " This is for anyone who has ever felt overlooked, undervalued, or questioned their place in the world." } }, "mood": "Empowering, warm, and inspiring", "camera": { "framing": "Close-up shot focused on the woman's face and upper shoulders, keeping attention on her expression only.", "movement": "Static (no panning, zooming, or shifting)." }, "visual_text": { "captions": false, "subtitles": false, "on_screen_text": null } } }

How do we fix it? I've put so much work into this.

[granite dove]

{
"scene": {
"character": {
"description": "A beautiful woman in her mid-30s with a caramel complexion, hazel eyes, high cheekbones, long brown hair with subtle blonde highlights. Bold red blouse, gold hoop earrings. Sitting on a cream chair with gold trim.",
"props": {
"microphone": "Red foam cover microphone matching outfit.",
"desk": "Desk with neatly stacked books."
}
},
"background": {
"description": "Modern feminine studio with warm lighting, white walls, gold/red/neutral accents, and potted plants; inviting yet empowering."
},
"audio": {
"voiceover": {
"tone": "warm_confident_uplifting",
"script": "This is for anyone who has ever felt overlooked, undervalued, or questioned their place in the world."
}
},
"mood": ["empowering", "warm", "inspiring"],
"camera": {
"framing": "close_up",
"movement": "static"
},
"visual_text": {
"captions": false,
"subtitles": false
}
}
}

[granite dove]

@surreal hearth test now

The error occurred because the JSON had text outside the structure and an unclosed quote, making it invalid and unreadable by the API.

[surreal hearth]

I received the same error message. "Failed Generation :against company policy to recreate prominent people."

[granite dove]

{
"scene": {
"character": {
"description": "Original, fictional woman in her early 30s (not based on any real person or public figure). Medium warm complexion, soft almond eyes, gentle cheekbones, shoulder-length brown hair with subtle highlights. Casual elegant red blouse, simple gold earrings. Seated on a cream chair with minimal gold details. The design must avoid resemblance to any known personality.",
"props": {
"microphone": "Table microphone for a home recording hobbyist (not a broadcast set). No logos.",
"desk": "Small desk with a couple of generic books (no titles or recognizable covers)."
}
},
"background": {
"description": "Cozy home-studio corner (non-broadcast), warm practical lighting, white wall with neutral decor and a small plant. No brand marks, no network set, no audience cues."
},
"audio": {
"voiceover": {
"tone": "warm_confident_uplifting",
"script": "This is for anyone who has ever felt overlooked, undervalued, or questioned their place in the world."
}
},
"mood": ["encouraging", "warm", "inspiring"],
"camera": {
"framing": "close_up",
"movement": "static"
},
"visual_text": {
"captions": false,
"subtitles": false
}
}
}

@surreal hearth So, that’s exactly what I explained to you, you know? When you took a photo of yourself and used it as a reference, it’s picking up features of your face and identifying them as a human figure. And that kind of falls under the category of commercial use, you see? That’s why I said that since you used a photo of yourself, it’s a real person. You need to make it clear that the person doesn’t exist that it’s fictional.

[surreal hearth]

The photo of me was morphed into this other person. It is not the original me. It would have no reference point for this person. This person doesn't exist in the world anywhere... not on socials... no where.

I didn't use any of the AI tools to create this person, just my own graphic design.

I've seen so many people on YouTube with a model of someone. These people have MULTIPLE videos of one character. Why is this possible for them and not for me? It doesn't make sense. My character is more original than any of those.

[granite dove]

No, but it doesn’t work like that, you know? Human aspects, human characteristics… it doesn’t mean you have to take a photo and post it somewhere. That’s not how it works. Human features, human aspects… even if you’ve never taken a photo, a machine can still recognize that you’re human. So I think now you understand what I mean. Human aspect, human characteristics it’s not about whether you post a photo or not.

[surreal hearth]

So why do other people have multiple videos with actual prominent people that we all know??

Why wouldn't it give them the same error message?

[granite dove]

Yes, that’s already another matter now we’re talking about the model itself, you know? For example, OpenAI’s Sora does this. You can give it the name of a celebrity, and it will literally pick up the features of that celebrity and put them into the image, just like what was happening. This varies from model to model, you see? It’s something specific to what the model does.

[surreal hearth]

So I can't make up a character in Veo 3, is what you're saying? Veo 3 created 2 other videos for me prior to this with no problem.

[granite dove]

Okay, but if you still want to use a photo of your face as a reference, here’s a tip. Take a high-quality photo of your face, and then in a video editing app, replace your eyes with the eyes of another person. It doesn’t matter who, just make it random just modify that part only.

[granite dove]

So then, the most important thing is that you already have the reference. You need to analyze what you did in that video note that slipped through without being blocked, you know? You basically already have the answer to your question you just need to analyze it.

[surreal hearth]

It doesn't have my eyes or anything. It doesn't look like me... only has my skin tone and similar body. When I say I used my photo, I did not put my photo in somewhere and ask for it to reproduce me. I only said to myself this is how I want it to look. THERE IS NO REFERENCE ANYWHERE FOR THIS CHARACTER.

I think you're thinking that this character looks like me... it does not look like me.

Pretend I did not mention that I used myself as a model, what would be your reasoning as to why I'm getting this message?

[fast ridge]

hey @marble ridge as stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now."

[maiden raptor]

I know you said forget that you used your own photo, but in order to help troubleshoot the error, can you clarify if your prompt is combined with uploading your personal photo as a reference? If the answer is yes. Then I would ask have you tried starting a fresh session and using the prompt without adding an image to reference? If you tried that and no longer received the error, then it would indicate the system is incorrectly identifying you as a prominent figure that is attempting to recreate. In fact I just generated a video with your exact prompt and no photo reference and it worked.

@surreal hearth You could take the same photo of yourself and ask Gemini to describe it to get an indication of what prominent figure you are being mistaken as

@surreal hearth this what I got by going to Gemini and clicking Video and then copy and pasting your prompt as you shared.

[golden bronze]

we featured some of your creations on the newest FlowbyGoogle Twitter account 🙂 check it out: https://x.com/FlowbyGoogle/status/1961153180559184334

FlowbyGoogle (@FlowbyGoogle)

We posed a creative challenge to our Discord community: start with the exact same ingredients and a prompt to create something in Flow.

The results... delightful! ✨

X · Aug 28, 2025, 7:45 PM

[cold sand]

Lol

It is ok we forgive you

[golden bronze]

putting myself in timeout

[twin flower]

ghost ping?

[modern walrus]

mulher com magia e maça pq toda vez que tem as 3 coisas juntas da uma merda terrivel?

[trim grail]

Send her to the dungeon!

[golden bronze]

sorry everyone for the accidental ping, i am a dingbat (known). hope everyone's doing well and you can forgive me.

[trim grail]

[twin flower]

hey @golden bronze why am i having camera role? what does it mean

[dull ice]

[zenith marsh]

WELL EITHER WAY i am thankful for that phantom ping ❤️

[last mortar]

No I can't forgive you georgi

[golden bronze]

it means you opted into content sharing in #start-here・📍! you can remove anytime

[twin flower]

it is fine, thank you :DDD

[fathom estuary]

that is coooool !!!

[hard knot]

Awesome!!

[golden bronze]

valid

[midnight quarry]

Hello?

[midnight quarry]

ahh np

[golden bronze]

i am sorryyy 😭

[midnight quarry]

You're good!

[trim grail]

it's Friday!!

time to push to prod

[ionic shuttle]

Final output after edits and music both applied 💪 😈.

[cyan estuary]

So you didn't share my awesome 8 seconds? Awwww

[ionic shuttle]

It’s okay 😆 I also have dingbat like properties so this one can slide 🛝

[golden bronze]

I don't see your submission in this channel! 😭 let me know if I missed it. I forward all opted-in creations to our social crews who take it from there!

[honest spruce]

I'm from India, flow is nice tool for veo 3👍🏻

[rotund trench]

I like that you still call it Twitter too 🫶🏾

[dense mason]

no way that is crazy what the hell oh my god

[surreal hearth]

Hi, Thx for the reply! I did the same thing that you did and received something similar. The thing is I'm wanting to create an AI influencer who has a YouTube channel, which means I want to use the same persona all the time/consistently. With the previous one, I generated the persona and made a few videos with her on Veo 3, but then suddenly I started getting the error message. So I don't want to start using a persona and then after 2- 3 videos have it give the same error because Veo 3 believes it needs to prevent generating real people without consent. The point is to stick with ONE persona. Do you know what I mean? How do I know that won't happen again?

[surreal hearth]

Hey Frontier Modal! Appreciate your comment. Keep talking to me, I'm interested in your perspective. So how do other AI influencers keep generating videos without this problem - Aitana López, Imma, Lil Miquela, etc.? Many are doing it. Do you think it's the use of multiple AI tools instead of trying to stick with one (Veo 3 in my case)? Definitely want to hear your perspective on this.

[tiny dawn]

We doing another one i wanna start getting in on these I have some really cool content due to my Google pro

[golden bronze]

the current prompt is [Defying Gravity](#flow-creative-challenges・🎨 message)! we'll have another next week too 👍

[surreal hearth]

It's not the difference between a good or bad generation. We know if it's a bad generation, just trash it and generate again. it's more so Veo 3 is saying we will never generate this particular persona again.

It's not the text because I've generated 3 videos perfectly... absolutely beautiful! After that I got the error message from Veo 3 "Failed Generation :against company policy to recreate prominent people." Although it's a generated image and not a real person at all.

Never mind, I just found someone on Upworks who says that they've run into this error many times on different generations and they know how to fix it.

I mentioned it from the very beginning. Anyway, thanks for your input. I've found someone on Upworks who says they've run into this same error code on Veo 3 and know how to fix it. Enjoy your day.

[mortal onyx]

Sorry for what might be a stupid question but for the creative challenges in here what exactly is :
…just don't forget the server rules & guidelines.

What rules and guidelines must we take into account when making a video creation for the creative challenge?
Thanks.

[mortal onyx]

Hope this is within the rules for what can be submitted and hope people like it. It's my first go so please be gentle.

[tiny dawn]

Oops I sent the gif version

[scenic pond]

Honestly I have no idea how to submit my video to the server i got frustrated with trying let me add my mind is full of thoughts and ideas hoping there will be a time when I can share with a community or world only thing I’m having trouble setting up

[late pollen]

Levitating + flying

[mortal onyx]

Honestly I have no idea how to submit my

[dreamy remnant]

Gravity Falls

[mortal onyx]

I really wanted that to continue. 🙂

[fast ridge]

hey @jagged sapphire As stated in #start-here・📍 , “Currently, our team can support engagement in English; the discussion in this server will largely remain in English for now.”

[jagged sapphire]

Ok, I just wanted to share my video made on veo3

[graceful karma]

I don't understand … most of the videos here have nothing to do with the topic.

[analog skiff]

reminder! the current challenge is defying Gravity!

[odd walrus]

where can i see the challenge ?

[graceful karma]

It's in the pinned messages for this channel.

[odd walrus]

got it

No one is actually creating anything related to the challenge, I think they are just sending their flow's render here

[graceful karma]

“No one” isn't quite right. There are actually a few on-topic videos here, fortunately.

[odd walrus]

I see

[graceful karma]

I like this one, for example.

[tiny dawn]

It said defying gravity

[tiny dawn]

Thanks for sharing this is forgot to click the emoji

All my content is created with Gemini and Google labs lol so only fair to share

[left umbra]

the future of seoul 🤣
this is my first work on this topic !

[hushed salmon]

Test run

[unreal night]

Faça um vídeo de uma mulher caindo

[odd walrus]

can u share the prompt ?

[karmic void]

Follow me on ig for prompts

[pastel breach]

ah I walked past this office a few times! I was interning at the meta NYC office in midtown good memories

[zenith marsh]

she's holding space

[golden bronze]

This week's Flow prompt is: Time Lapse 🕐

we're focusing on new perspectives this week with Flow. 🌊 using Flow, create a story told through a time lapse. this can be the entire story or an element of the story, the choice is yours.

Some examples:

• a time lapse of a single room could show the progression of a relationship as furniture shifts and dust collects
• a time-lapse of a desk could show the creative process of an artist, from a blank page to a finished masterpiece.

Share creations right here in this channel

[modern walrus]

modo foto na minha ia de relacionamento

[inner blaze]

Big City Nights

[wicked mason]

The many years of growth of stalagmites and stalactites

[left umbra]

🌌 Cosmic Dance of the Solar System ☀️🪐
The Sun and planets flow in a mesmerizing timelapse cosmic dance ✨
🔧 nano banana hidden feature + flow veo3

[queen sphinx]

https://tenor.com/r9HN.gif

[graceful mural]

OMG Alex! you just gave me the best idea ever. My grandfather was the first graphic reporter of my country, his pictures have the same tones-vibes of your video, I don't know how I missed to think on making sth like this ♥

[fallen vortex]

@graceful mural You are welcome ). I will be glad to see the result.

[ionic pulsar]

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

[willow gyro]

Time lapse 🇯🇲

[mystic maple]

Which website do you use to generate your videos?

[ionic pulsar]

Google Flow I have the ai ultra plan

[mystic maple]

Ok. The trick I use to keep my characters consistent or the same is to first generate the video script, then create the images, and finally animate the images using the "image to video" feature

[ionic pulsar]

What if I’m doing like the swim in this ocean one

And it’s like

jello ocean

Lego ocean

[mystic maple]

Ok. But do you write the prompt to generate the script? Send it to me and if there are any corrections, I can guide you to fix them.

[trail laurel]

(https://www.youtube.com/watch?v=NS9nImre3aA) This is a short film made entirely with flow. It was selected at several film festivals dedicated to AI.

[undone cape]

How did you do this 😭

[golden bronze]

This week's Flow prompt is: Reflection Remixes 🪞

we're celebrating the latest update to Flow: vertical aspect ratios are now available📱🔥 learn more from our [latest announcement](#flow-announcements・📣 message) with that, we're focusing on remixing Flow TV prompts for our "Reflection Remixes" theme 🔀

How to get Started 📽️

1. Head to the "Time to Reflect" channel on Flow TV
2. Select "Reuse Prompt" on a prompt of interest
3. When prompted, create a new project with said prompt
4. Select Portrait (9:16) format within the settings
5. Be sure to replace the "wide shot" language within the existing prompt.
6. Remix the prompt to preserve the reflection component while making it your own.

[golden bronze]

🦢🦦🤸

[last cosmos]

Swan got Mike Wazkowsi'ed

[inner blaze]

This was not really what I had wanted it to do. The reflection was supposed to become its own entity of sorts. not even close

[undone cape]

I have issues I wanted to make him show the screen then back to pressing his phone screen

[undone cape]

How do you mean

[ionic pulsar]

Hey can you dm?

[undone cape]

How did you get this

[golden bronze]

try #flow-tips-n-tricks・✏️ for prompting collabs! might help with some refining 🙂

[jade summit]

You have to generate the images first which then you can use to create videos. I have some examples if you want to take a look...

[ionic pulsar]

Where can I generate images because I only see text to video

[jade summit]

how many images do you want to generate? Gemini is one good option if your usage is low. I am creating an app for similar purposes (VAKPix), mostly focused on content creators. check DM

[jade summit]

you can not create images from flow. You have to use different apps

[inner blaze]

try Whisk

[jade summit]

How? I never had any option to create an image...

@clever lantern

Also, I am developing an image gen/edit app based on nano banana. Please let me know if there is any valid way to promote it. Thanks!

Thanks! Got it. but it can not create an image from a reference image...!

Okay. it can...

my bad. Thanks for the info btw!

[jade summit]

Anything on this?

yes. Trying it.

works. Thanks!

[fleet dew]

@jade summit how to edit foto to video veo3

[jade summit]

Hi, you can choose "Frames to Video" option from chat box menu. there is also "Ingredient to Video" options for Pro users

[fast ridge]

@high niche we are unable to generate images in this server 🙂

[jade summit]

why did you tag me? I am a human 👀

[fleet dew]

Okay. Thank you 👍

@jade summit Do you have a YouTube channel?

[jade summit]

Yes. Why?

[frail bay]

how to generate prompt to video?

[zenith marsh]

sorry not sorry i am late 🙇‍♂️

[cunning quail]

Nano banana. Someone wanted a realistic Sea Otter. (Sorry can't find the correct place for, Nana banana images.)

[cunning quail]

My lovely lone gone Border collie Megan. Giving her a new live. Nano banana. Enjoy.

[forest palm]

Time to Reflect

[golden bronze]

This week's Flow prompt is: Extreme Sports Remixing 🛹🏀 🪂

we had a lot of fun remixing Flow TV prompts last week 🔥 so we're trying it again - this time, we're remixing prompts from Flow TV's "Extreme Sports" channel 🚲⚽🛼

How to get Started 📽️

1. Head to the "Extreme Sports" channel on Flow TV
2. Select "Reuse Prompt" on a prompt of interest
3. When prompted, create a new project with said prompt
4. Remix the prompt as you see fit, making it your own

📣 We love featuring creations shared in this channel (and across the server!) - if you're interested in potentially having your content be shared across Labs socials, be sure to check out the opt-in system in #start-here・📍.

[zenith marsh]

🔥 this week theme is an extremely fun one

A kayaker paddles through glowing molten rivers, each stroke splashing arcs of fire. The drone circles low, lens catching red-hot embers as they burst into the air and sizzle out against the kayak’s hull. *You can click on the link to reuse the prompt *

[willow gyro]

Extreme Sports Remixing 🐎

[torn coral]

As reminder we cannot create images and videos in the server!

[solar herald]

Yes

[hallow arrow]

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

[white comet]

i noticed that peoples prompt is hidden, so i wonder if it is a rule , well if it is i must wonder yet again but if not here is a one💯 /"Hyperrealistic, National Geographic style video. A majestic bald eagle with sharp, detailed feathers soaring gracefully in slow motion through a mountain valley with towering snow-capped peaks in the background. Dynamic aerial drone shot, tracking the eagle smoothly from the side. Shot during the 'golden hour' with bright, natural sunlight casting long shadows. Tiny ice crystals glinting in the air, wind ruffling the eagle's feathers. 8K, cinematic, professional color grading."

[zenith marsh]

Thanks for sharing , did you make sure that you follow the theme of the challenge for this video?

[misty seal]

💀

[ornate tinsel]

Hello, as a reminder we cannot create images and videos in the server and it will not generate. Thank you!

[left umbra]

I made from this:
https://labs.google/flow/tv/channel/extreme-sports/noJusHX4eCq1ZuD8oqsE

[acoustic mulch]

I made this.

[analog skiff]

Hello, as a reminder we cannot create images and videos in the server and it will not generate. Thank you!

[golden bronze]

This week's Flow prompt is: Film Noir, featuring Prompt Expanders! 📽️

We're celebrating a new feature within Flow: Prompt Expanders This new feature uses Gemini to enhance your prompts to create a specific style of video. Learn more ➡️ https://x.com/FlowbyGoogle/status/1970607925443076587

How to get Started 📽️

1. Head to Flow (flow.google) and create a new project.
2. Start prompting with a starter idea - feel free to get creative 🎨 Need inspiration? Check out #flow-videos・📹 & #flow-tips-n-tricks・✏️ 🙂
3. Click "Expand" on the bottom right, and select "Film Noir"
4. Remix the prompt as you see fit, making it your own

📣 We love featuring creations shared in this channel (and across the server!) - if you're interested in potentially having your content be shared across Labs socials, be sure to check out the opt-in system in #start-here・📍 🙂

[zenith marsh]

🐕 tested it on one of my fav gen prompt from many months ago, if anyone remembers 👀 🕵️

[feral tusk]

A film noir - This week's challenge.

[lean latch]

Hi

Hi

[ornate tinsel]

@obtuse trout
Hello, as a reminder we cannot create images and videos in the server and it will not generate. Thank you!

[willow gyro]

Film Noir

[west dove]

This was short, we demand a sequel to the movie.

[quartz socket]

Guys how I make vedeo veo3 here ?? Need help

[willow gyro]

Part 2: Do they ever notice? Coming soon… 🎬🐀

[fallen vortex]

Night Snail Bar

[acoustic mulch]

The Obsidian Hour
Week challenge "Film Noir"

[sullen oriole]

I know we were supposed to use the prompt expander...but I just wanted Bambi hunting for clues under mushrooms lmao

[hallow arrow]

Hello, a friendly reminder. You cannot create images or videos in this server or within channel. You need to use an external browser. Also 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.

[left umbra]

Here is Film Noir version, I had shared #perpetual-stew・🫕 with my prompt 🤣

[sour summit]

Where should i create the video which channel to use?

[hallow arrow]

Hello, a friendly reminder you cannot create videos within the channel. You must use an external browser. Thank you.

[ornate tinsel]

Hello, a friendly reminder you cannot create videos within the channel. You must use an external browser. Thank you.

[hoary zinc]

Okay will keep in mind but create an video of an snail black and white snail is producing another snail

[rose hawk]

Hello, a friendly reminder you cannot create content within the channel. Prompts can only be used in an external browser. Thank you.

[broken carbon]

Working through the plot. Who should the thief be?

[nova umbra]

Working through the plot. Who should the

[lime cargo]

how do you make longer videos?

[broken carbon]

Great question! I create a project within Flow, then start creating scenes. I’ll work to get the shots and actions I’d like to see within the film (getting a scene ‘right’ might require a few attempts with the same/similar prompts). From there I drop the ‘selects’ into an assembly (scene builder) that matches the story I’m trying to tell.

Or another way of saying, each video is still 8 seconds. The way it is longer is using the scene builder function within Flow.

[hallow arrow]

Hello, a friendly reminder you cannot create content within the channel. Prompts can only be used in an external browser. Thank you.

[hallow arrow]

Hello, a friendly reminder you cannot create content within the channel. Prompts can only be used in an external browser. Thank you.

[abstract quail]

two camera movements in same clip. can you tell me the prompt ?

[modest cloak]

Moon shine above in a rat house 😂😂🫠

[golden bronze]

the Film Noir entries really blew us away 😮‍💨 🖤 🤍 stay tuned for how we'll feature these... 👀

(‼️as a reminder - we won't feature your work unless you've opted in to content sharing here: #start-here・📍 message. this issues you the 📸 role, btw)

[cyan raft]

not what i was going for but i pay for only 3 videos a day no room for what i want and makes it boring to do videos, If I only get 3 a day why waste my time with 3 failures a day?

[hallow arrow]

Hello, a friendly reminder you cannot create content within the channel. Prompts can only be used in an external browser. Thank you.

[sullen oriole]

sorry for the wait, the prompt was simply "black a white film featuring bambi as a detective looking for clues under mushrooms"

[devout yarrow]

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

[primal summit]

look what i created

[graceful karma]

Just out of curiosity, what language is this? Sounds like Swiss played backwards. 😆

[azure sentinel]

have you thought about deconstructing the videos and seeing what went wrong ?

[tropic tulip]

I tested Veo 3.1 in Gemini and Flow

[tropic tulip]

Action Design: The key elements are a slight glance of the eyes and a single drop of rain. These two subtle movements are enough to break the stillness of the frame, suggesting that the character is intensely observing their surroundings and has been holding this pose for an extended period.

[primal saddle]

How did you make it 13 sec? Can we expand now 9:16 videos?

[nimble axle]

you can kinda manually extend them by using the last frame of your first generated video and use it as the first frame for your new prompt to extend

you have to put the clips together yourself though

they might have figured something better out but i thought I'd mention it cause i have a continuous 48 second video now im working on

[primal saddle]

Thanks for the quick reply

[fallen vortex]

Hi @undone cape Sorry, I missed your message. It was a prompt about a city square in the early 20th century. If it's relevant, I can look up the exact prompt.

[south ginkgo]

The transition are so smooth, nice!

[golden bronze]

This week's Flow Creative Challenge: Animal Adventures

we're celebrating the launch of Veo 3.1, our newest video generation model learn more about Veo 3.1 and the latest updates through our [latest announcement](#flow-announcements・📣 message)

With "Ingredients to Video," you can use multiple reference images to control the characters, objects and style. Flow uses your ingredients to create a final scene that looks just as you envisioned. for this week's Flow Creative Challenge, we'll be focusing on ingredients!

The Prompt:

You've been provided with two ingredients - featuring the [Labs leveling animals](#server-directory・☎️ message): Collective Capybara and Passionate Parrot these two best buds are headed out on an adventure that will rely on their individual strengths (Passionate Parrot's flight, Collective Capybara's swimming) to help achieve a common goal. where are Passionate Parrot and Collective Capybara going? the rest is up to you.

How to Start:

1. Head to flow.google
2. Select "New Project"
3. Above the prompt box on the top-left, select "Ingredients to Video" and upload both the provided ingredients.
4. Start prompting!

📣 We love featuring your creations that you share in this channel (and from across the server!) - if you're interested in potentially having your content be shared across Labs socials, be sure to check out the opt-in system in #start-here・📍.

[fallen vortex]

I was able to come up with situations in which they could be useful to each other.

[ornate tinsel]

Hello @livid rain , as a friendly reminder we cannot generate images in the server! Please do not post your prompts in the channels as it will not generate, unless you are asking for advice or tips

[ornate tinsel]

Hello @strong owl , as a friendly reminder we cannot generate images in the server! Please do not post your prompts in the channels as it will not generate, unless you are asking for advice or tips

[left umbra]

I just prompted "Generated freely" 🤣

[shrewd palm]

Prompt: Create a cinematic short video featuring a capybara named Capi and a parrot named Paco, based on the uploaded reference photos.
The two are best friends on a tropical jungle adventure.
The video shows them trying to rescue their favorite toy — a colorful coconut — that floats down a jungle river.
Show Capi swimming powerfully through the water while Paco flies above, guiding him from the air.
Include scenes where they help each other: Paco picks up a stick to free Capi from reeds, and later Capi catches Paco when he falls from a tree.
End with a beautiful sunset scene where they rest on a small island, happy and proud.
The tone is heartwarming, humorous, and cinematic, with dynamic camera movement and vibrant jungle colors.

[shrewd palm]

This time, i used a German prompt: "Der capybara und der Papagei aus den beiden hochgeladenen Fotos spielen zusammen am Ufer eines Flusses mit einem Ball. Der Ball rollt versehentlich ins Wasser und fängt an, Wegzuschwimmen. Der Papagei fliegt los und zieht den Ball aus dem Wasser, fällt aber wegen des Gewichts dann mit dem Ball selbst ins Wasser. Der capybara schwimmt zu ihnen und rettet den Papagei und den Ball zum Ufer"

[tropic tulip]

I used SceneBuilder to extend a video. The difficulty was to make the background music continue seamlessly. The result is just okay.

[sudden ember]

<@&1366831844502274180>

[small imp]

You've tagged us twice. How can we help?

[normal walrus]

The Video shared here? Does those have copyright for that particular user ?

[golden bronze]

we loved seeing what y'all made for the 9/24 challenge: Film Noir 📽️ so we featured a compilation on the FlowbyGoogle Twitter account ❤️ check it out here: https://x.com/FlowbyGoogle/status/1981103596805771352 thank you so much to everyone who participated!

📣 interested in having your works be featured across Google Labs socials? be sure to check out the opt-in system in #start-here・📍

FlowbyGoogle (@FlowbyGoogle)

Take a look at the stunning, shadow-drenched videos our Discord community created using one of our new prompt expanders Film Noir.

X · Oct 22, 2025, 9:01 PM

[rose hawk]

Hello @rugged prawn , 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 you are asking for advice or tips on your prompts do so in the appropriate channels.

[rugged prawn]

ok

[left umbra]

Here are my challenge results:

1. High-altitude flight
2. Spider-Man-style chase
3. Inspired by wingsuit footage
4. To the Moon

More details are in threads

[gilded linden]

hi

[keen jewel]

Hello

I'm a french person and my english it’s not so good but I'll do my best

So, I want to know the objectif of this "flow-creative-challenges".

[ornate tinsel]

Hello @jade lichen , 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 you are asking for advice or tips on your prompts do so in the appropriate channels.

[zenith marsh]

🫡

[zenith marsh]

also bonus blooper i guess ! this was generated before i realized there’s already a prompt theme for this creative challenges, LMAO.

[sullen oriole]

When Passionate Parrot attempts teaching Collective Capybara how to speak, the bird quickly learns that his furry friend is well ahead of the game

[analog skiff]

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.

[left cobalt]

<@&1366831844502274180> spam

[granite dove]

Hello friends. Choose a fruit and its texture. I am running tests.

[reef coyote]

OH, this is a result of sora.ai but i expected more but its fine as same as Veo mean see that below:

[reef coyote]

i didn't understand 2nd last word @granite dove by the way try something more interesting like making thing impect full if you have sora.ai pro subscription don't wast it

but if you are new so keep it up

[granite dove]

new ?

[devout yarrow]

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

[reef coyote]

hello <@&1366831844502274180> , can anyone tell me that what is the porpose and impect of this #flow-creative-challenges・🎨 channel. I am confused about uploading my experience and video shots here because all previous videos all around Passionate Parrot and capybara So, tell me that this channel allow me it's or not. plz guide about that!

[analog skiff]

each week there is a creative challenge posted by one of the google staff, we have #flow-videos・📹 for general videos made with flow!

[winged island]

Linking to this weeks challenge here: #flow-creative-challenges・🎨 message
Or view it in pinned messages!

[analog skiff]

but also reminder that mods are here for moderation issues so please dont ping us in future if there is no moderation issues/ rule violations

[faint lagoon]

Hi

[torn coral]

Hello @slow ruin , 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 you are asking for advice or tips on your prompts do so in the appropriate channels.

[versed bone]

3.1 fast lol nice

[hallow arrow]

Hello @pastel raft , as a friendly reminder we cannot generate images or videos in the server! Please do not post your prompts in the channel as it will not generate. If you are asking for advice or tips on your prompts please use the appropriate channel such as #flow-tips-n-tricks・✏️ Thank you.

[junior moth]

@sullen oriole @sullen oriole

[sullen oriole]

Yes?

[junior moth]

Good

[analog skiff]

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.

[ivory ether]

8-second cinematic realistic video showing the full process of online shopping for “Asankharid”.
Scene 1: Close-up of a man using his smartphone to order a product on the Asankharid app.
Scene 2: Quick cut to a warehouse where a worker packs the item in a brown box with the Asankharid logo.
Scene 3: A Tipax delivery van drives through the city streets.
Scene 4: The courier delivers the package at the customer’s door; the customer smiles and opens it.
Realistic lighting, smooth transitions, commercial style, shallow depth of field, 4K quality, no text on screen.

[rough sierra]

hello please what prompt did you use for that

[hollow rune]

Anyone have luck in creating different character accents? What's the secret?

[ivory sand]

https://youtube.com/shorts/2SX3UDZVGvw?si=ZATSMwcDkee4FOPw

Flow + Suno

[analog skiff]

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.

[past inlet]

Generally if you place the subject in a specific region, or city in a location with a specific known dialect of English.. you can reference, for example, to obtain an accent of a person who lives in Mumbai, India by specifying details that reinforce this desired attribute . I have found that outright saying 'speak with an indian accent' can induce failures via content restrictions, less Simpson Qwiki mart clones?

[past inlet]

I have used xml / json tags in other prompts when using the API, does anyone know if these types of input is acceptable/advantageous ?

[keen gate]

very nice effort, I am also trying to make a long video but through only flow lab

[analog skiff]

fwd to #flow-tips-n-tricks・✏️ !

[lean axle]

@keen gate nice. Flow is cool for certain things but it's too censored for my kind of cinematic story telling. It's also very expensive compared to the Chinese models that now do the same and the physics are bad. Lastly they charge you for pro and still leave a watermark on your creations. I use it for a few scenes tho since I have pro and alot of credits on there.

[keen gate]

I get your point, but I have my own Ultra Pro account, and it’s been working quite well for me so far no watermark issues and overall pretty smooth performance. 👍

[keen gate]

Hello Bee,
I’m new here, and I’m not sure if I might have shared the wrong video. Could you please let me know if sharing videos isn’t allowed in this group?
Thanks!

[analog skiff]

sharing videos is allowed! this channel is for our weekly flow challanges! we have #flow-tips-n-tricks・✏️ for advice on how to best improve your flow videos!

[lean axle]

@keen gate That's an expensive commitment to what I'm trying to build. If they weren't so sensitive and strict I would definitely commit to ultra. They have to lower the guardrails to allow for real cinematic story telling and not just cat videos.

@analog skiff what's this week's challenge please?

[keen gate]

@analog skiff yes whats this week's challenge ?

[keen gate]

Yeah, totally understand your point, they’ve kept the creative boundaries a bit too tight for cinematic storytelling. But from my side, the Ultra plan has been working pretty well so far. I think once they relax those guardrails a bit, it’ll open up a lot more possibilities for creators like you.

[analog skiff]

weve not had one for a while but this was the last one:
#flow-creative-challenges・🎨 message

[keen gate]

oh ok

[sweet finch]

can anybody help me why minors video are not allowed in flow

I imported a film image of child, where he is standing with a blank stare, and when I try to import it into Flow, it says ‘We do not allow the import of images of minors at this time. Please try with another image or send a comment,’ even though the image is fine. Frankly, we've been stuck since its release in Image to Video. It's tiring.

[inland cobalt]

@inland cobalt please where can I create a video here?

[hallow arrow]

Hello, a friendly reminder videos cannot be created in this server. For flow please visit https://labs.google/fx/tools/flow . Thank you.

[random dagger]

Question: How do I go about keeping character voices the same when transitioning into a different clip or extending an existing clip. I am getting some reall crazy mixtures from scene to scene.

[keen gate]

write your prompt perfectly you will get the good results, as I am creating many videos for sports

[keen gate]

The error message you received, "We do not allow the import of images of minors at this time," indicates that Flow (the platform you're using for Image to Video, likely a reference to a specific AI tool like Gen-2 by Runway or similar) has a very conservative and automated policy regarding images of children.

AI Detection: Even if the image is a film still and perfectly "safe" or "fine" by your standards, the underlying AI safety model in Flow is designed to be extremely cautious and likely flags any recognizable image of a child. This is a common practice to comply with Child Safety and Protection regulations and internal ethical guidelines, preventing misuse.

Context doesn't matter: The AI often doesn't consider the context (e.g., "It's a film image," "It's harmless"). It just detects the presence of a face belonging to someone below a certain age threshold. The "blank stare" might even be interpreted in a way that triggers an alert within their safety system.

[keen gate]

Sir you can create videos on labs.flows

[keen gate]

Generate Clip 1.

Extract/Clone the desired voice using an external tool (e.g., ElevenLabs).
Generate all dialogue for the scene/extension using that single cloned voice file.
Video Editor: Mute the new clip's audio, and replace it with your single, consistent audio file.
This ensures the voice is stable across all cuts and extensions.

[dense steeple]

how do you guys creatw these videos here

[torn coral]

Hello @dense steeple , 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!

[rough hound]

so then what i get
when i share any video here and guys like it ?

[keen gate]

@analog skiff Kindly advise is this allow to share video here ? #flow-creative-challenges・🎨

[zenith marsh]

This is the Creative Challenges channel it’s for those who actually want to take part in the weekly themes. Follow the theme , make your video based on the theme, and share it here.,

the actual channel for posting any kinds of video generation using Flow woul be #flow-videos・📹

[zenith marsh]

the theme are already pinned in this channel btw. so do check them out 🙂

[analog badger]

feed back'

[rapid flint]

Well its pretty cool, but these its under Flow Creative Challenge. Has a set of rules

[hallow arrow]

Forwarded to #flow-videos・📹

[frail plaza]

where are weekly prompt?

[strange vault]

Photo ka video

[lime gulch]

hello, Why does Flow tell me that “portrait mode isn’t compatible in vertical and that I need to switch the screen to horizontal” when I try to create a vertical video? I checked the prompt to make sure it doesn’t mention portrait mode, but the same message still appears, even though it still generates the video.

[willow ocean]

🙂

[paper charm]

https://www.youtube.com/watch?v=ccJ-BiHnn4w

[chrome flint]

is this on your personal server?
or is it on a professional server?
Also would like to know what is it to you, in any way, as in before someone sees this media maybe they are of some work flow or some thoughts and would be or would have been during or after,
thus, the question here is no matter what the duration or maybe of the duration also what is it really to you for anyone or your main target audience, tell what is it that would be during and or after from this media

please do not take this the wrong way... it is possible that I unlike anyone else (possible).... would like to familiarize you with proper, possible or accurate criteria not only of readiness

[paper charm]

Hello @chrome flint, I own this youtube channel and the video is created by me using Veo 3.1. This is cretef for kids aged 2-4. Is this what you asked? I would be happy to assist if you want to know something else.

[willow ocean]

Did you did it all with flow?

[paper charm]

Yes with flow. I created my main character "Neini" using gemini and the animated everything using Flow. Mostly extending the video after each scene.

[willow ocean]

Nice bro, thanks!
Keep it up 🙂

[chrome flint]

can you further explain that

[paper charm]

1. Server Type: It's on my personal YouTube channel.

2. What it is to me: It's a passion project. I created this alphabet learning animation using Veo 3.1 to practice my animation skills and contribute something educational.

3. Target Audience & Impact: The main target audience is kids ages 2–4 (and their parents/guardians). During the video, I want them to be engaged and absorb the letters easily. After watching, the goal is simple: that they recognize and remember alphabets and associate them with animal names.

[chrome flint]

I am not stating that you are wrong sir, i was simply trying to understand your functionality or what it was about and without any background information at all.
Additionally, it is not that you are not valuable, it is that maybe there are further identifyings of all this to be monitored and also on your own behalf also... however, i may not be expert either

[chrome flint]

maybe you have some supervisions that you may further engage with or converse with more regularly of standard bare conversations not anything displaced

[chrome flint]

anyone read any of the Harry Potter series? Can you all chat about what is of the quittage games or chapters of quittage games? Anyoen ever get into any of that?

[zenith marsh]

wrong channel?

[swift crater]

Oh ur right lol, I'll post it in the correct channel. overlooked the name

[rose hawk]

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!

[nova pike]

Hi there, you Flow SME's how do you overcome encoding/font bug having an issue with Accurate Text Rendering in Veo Videos. How have you overcome this because I have seen some great veo vids with excellent text.

[quartz robin]

https://www.youtube.com/watch?v=eCYe5zwSCG8

O Gato - "The Cat"
Short film entirely made using AI.

[pale cradle]

🎬 Flow Creative Challenge | Character Consistency

TO FALL INTO YOUR EYES A fully AI-generated cinematic music video created entirely with Google Veo 3

Watch: https://youtu.be/dDpNDH0wX7U?si=Dpev8GT0WYWLvPRP
━━━━━━━━━━━━━━━━━━━
🎨 The Project: A 3:40 narrative music video about unrequited love—45+ shots maintaining perfect character consistency for two characters throughout the entire story.

🤖 Built With:

• Video: Google Veo 3.1 (via Flow)
• Character Refs: Ingredients to Video
• Music: ElevenLabs
• Post: CapCut

🔧 Technical Approach:

Character Consistency Strategy:
✓ Detailed 200+ word character descriptions
✓ Reference images in every generation
✓ Consistent prompting structure across all 45 shots
✓ Result: 85-90% first-gen character match rate
Prompt Engineering:

Each shot included:

• Character reference DNA
• Camera specs (lens, angle, movement)
• Lighting (color temp, direction, quality)
• Motion physics (subject + camera)
• Emotional beats • 200-400 words per prompt

━━━━━━━━━━━━━━━━━━━

📊 Project Scope:

• 45+ Veo generations
• 3:40 final runtime
• Every lyric mapped to visual • Cinematic narrative structure
• Also submitted to Chroma Awards

🎯 Goal Achieved: Prove that Flow + Veo 3 can create emotionally resonant narrative storytelling with consistent characters when directed with creative intention.

Feedback welcome! What challenges have you faced with character consistency?

[heady raft]

hi

[dry bolt]

Great job, and thanks for sharing details about your approach to help others.

[sick kernel]

there is alot of restrictions for veo 3.1, of alot of reasons, i also can hit alot of road bumps when trying to generate even the most innoicent videos, i let veo 3.1 generate a video of 2 children fully clothed stand on a beach holding hands an looking at the sunset, no problems it rendered fine, but when extending the clip i got the messege that it coould violate their terms, very strange the only thing i wanted to was raise the camera over the children so the full sunset came in view. The only thing you can do is report this things to google team, so they can fine trim the rules for better and more consistant renders with out this blockers

[drowsy quartz]

https://www.youtube.com/watch?v=lgXJ9GaFmP0

[analog skiff]

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.

[ornate tinsel]

Hi @wooden kite ,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.

[devout yarrow]

@steel pineAs stated in #start-here・📍; currently, our team can support engagement in English; the discussion in this server will largely remain in English for now.

[steel pine]

How can i generat vidéo

[vale totem]

To generate a video with Flow, start here: https://labs.google/fx/tools/flow

[wooden sundial]

@wooden sundial

[devout yarrow]

@weary lance Hello, just a reminder that there is currently no way to generate images or videos on the server, so posting prompts will not produce any results.
If you need advice or feedback on prompts, please use the appropriate channel dedicated to that purpose, for example #flow-tips-n-tricks・✏️.

[smoky swan]

Wow...the quality is top notch...is this Nanobanana Pro?

[plucky nymph]

Hello. I have questions about the rules for the upcoming Google Ai film contest. Is this where i can ask?

[strange nimbus]

Hi

[humble quail]

Hi

[modern vapor]

Sorry flow team, we stop used our Ultra account for useless video since we have to pay for every generations.

[fervent hinge]

we should get a refund this is a scam

[hallow arrow]

Hello @runic echo , a friendly reminder videos and images cannot be created within the server. Please visit https://labs.google/ and pick a tool to get started. If you need help please make sure to ask clearly and do so in the correct channel. Thank you!

[formal wind]

I've hit the limit on my Ultra account. What is the daily generation quota?

[rose hawk]

Hi! A friendly reminder videos and images cannot be created within the server. Please visit https://labs.google/ and pick a tool to get started. If you need help please make sure to ask clearly and do so in the correct channel. Thank you!

[drowsy raven]

Add smooth animations

[rose hawk]

Hi! A friendly reminder videos and images cannot be created within the server. Please visit https://labs.google/ and pick a tool to get started. If you need help please make sure to ask clearly and do so in the correct channel. Thank you!

[analog skiff]

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.

[scenic gulch]

fr

this is literally scam

[quiet moss]

did you get an answer?

My ultra account ran out of Nanobanana Pro image generation after like 30 min of use.

[prisma parrot]

chat! im trying to get a UGC person to look at the camerea without saying anything, or moving their mouths. any tips? can't get it to work they seem hell bent on talking / moving their lips.

[lilac hull]

I did a whole video where I needed a lot of this. Variants of this worked:
Locked frame. Static camera. The character gives a controlled, weary, internalized performance. He listens to the hum of machinery with suspicion. Eyes drift downward as if following a vibration beneath the deck. A slow, subtle head tilt. Jaw tightens once. No talking. No gestures. Keep movement minimal and restrained. Match the provided image exactly.

I think controlled, internalized, restrained, minimal are probably what do it.

[fresh tusk]

Hey everyone! I’m trying to generate an animated scene in Veo / Whisk / Nano Banana with a flat, front-facing 16:9 screen so I can drop a live-action video into it later.

My issue: No matter how clearly I prompt it (“perfectly flat,” “no tilt,” “no perspective,” “front-facing”), the model keeps generating a screen that is angled, tilted, or in perspective instead of straight-on.

Has anyone found a reliable prompt pattern or technique that forces these models to render a perfectly rectangular, camera-facing screen—like a TV or monitor seen directly from the front?

Would love any advice or example prompts that have worked for you! 🙏

[keen gate]

#flow-creative-challenges・🎨 Hey , I have just worked on your given task and see the below prompt and work on it and share the result whats happening ?

An orthographic projection view of a 16:9 aspect ratio display screen. The screen is perfectly rectangular and centered in the frame. Zero perspective distortion, purely 2D flat geometry. Architectural front elevation view. The camera plane is perfectly parallel to the screen surface. A minimalist, glowing placeholder screen against a neutral background. Symmetrical composition.

[fresh tusk]

Thank you so very much!!!

[keen gate]

Your welcome always

[fresh tusk]

Thank you--the screen was being placed in an "animated room"--it kept trying to place the screen on the wall which tilited it, but when I added this to your prompt it worked: "Place the screen directly to the right of the character, on the exact same depth plane and foreground level as the character. They must appear side-by-side, aligned horizontally, both at the same distance from the camera, just like the example composition on the right-hand reference. The screen should rest on a short stand or two small caster legs placed on the floor, similar to a mobile smart board"

[warped lagoon]

When I design a photo using Flow, the quality is very low. It comes in at 311x553 pixels. Is there anything I can do to make it 1080p or higher quality?

[hallow arrow]

Hi there, your message has been forwarded to #flow-tips-n-tricks・✏️ . That is the appropriate channel for prompt help. Thank you!

[quartz robin]

https://www.youtube.com/watch?v=lvO7MqwE2N0

Tribute to Phil Lynott - Animated short film
Finalista no Festival AI Creative Festival (@aicreativefestival) na categoria animação.

[fickle hare]

Anyone else having issues with assignment of dialog with two people in a scene? I have tried 6 different prompts using a dialogue block, voice direction, banning the other speech and by adjusting timing. I have even tried rewriting in XML and JSON. Any other suggestions?

[keen gate]

That’s fantastic news! I’m glad that adjustment solved the tilting issue. Specifying the 'same depth plane' and the details about the stand was a smart way to stop the AI from trying to mount it on the wall. I'll make a note of this refinement for future reference. Let me know if you need anything else!

[swift crater]

Hii! I'm new in this channel

what is this channel for? I would love to participate if it's some challenge related stuff! 😆

[frail grail]

First time making Rap Mv using veo 3.1
https://youtu.be/ZhC1cbtmwnQ?si=qKisMrcWsxK2zANR

[red latch]

Hey everyone!
My first post here. I’ve created a short film using Flow and a few other tools.
I’d really appreciate any feedback or ideas for improvement.
Greets!

https://youtu.be/cC9iHfZ0Xx4?si=Iy4eNuGIspVv63Xl

[fringe sorrel]

Amazing work bro!!!

[still summit]

Flow can’t generate 1080p yet, 311×553 is the normal output.
If you need higher quality, just upscale the image with an external tool.@warped lagoon

[warm hinge]

https://youtube.com/shorts/shRpu1m1SeM?si=j7gJRLGpGTh7YK9W hii

[ornate tinsel]

Hello @pallid spade 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!

[warm hinge]

IS that true? U deleted my Art because of the Ai actor get arrestes of drug abuse and use this Nickname?🤣

[strong peak]

I love the back ground music, proffesional grade. relfecting the emotion of the visuals.

[small forge]

Whats your preffered upscaler?

[clever lantern]

720p is the output.

[clever lantern]

It's not cheap, unfortunately, but Topaz is pretty universally understood as the best.

[small forge]

Your right about that. Im also considering higgsfield.ai

[clever lantern]

Higgsfield's upscaler is run by Topaz, and has been down for a while. I check every day, because I have a backlog, and don't want to pay for yet another service just for this.

The Higgsfield one went down at the same time, but it's cheaper. It apears to be some kind of streamlined version of Topaz, but for stuff that just needs simple upscaling, it worked for me in the past, before this 'Service disruption'. If you want to upscale and smooth things out a bit (creative upscaling), then 16 credits per clip (for 1080p or 24) is still pretty reasonable.

[small forge]

@clever lantern Thats good info. Thanks for that. I tried using a self hosted Video2X + Real‑ESRGAN/RIFE. Its still not up to par.

[clever lantern]

None of the local models are near as good as the best private models. Eventually they'll all be so good the difference won't be a big deal in regular use, but we aren't anywhere near that yet, except for basic LLM use and some uncomplicated image generation stuff, if you have a really beefy PC.

[small forge]

Good thing there this channel for people who are into creatives. Ive tried automating soo much stuff with different tech stacks. But, still unable to find the golden ratio for output quality without the licensing issues.

[small forge]

I completely agree. Were atleast waiting a few months time for this to happen unlike before, that we have to wait for years.

[clever lantern]

I don't see the point when it's so much cheaper and faster to use far better models like Nano Banana Pro, Leonardo and Midjourney. You'd have to pay for those for years to make up the cost of a 5090, for lower quality results, generated a lot slower.

[small forge]

Thats unfortunate. At the moment im trying to play around with the diffusion models to get a great balance.

[small forge]

Thats right. US has a lot of compute power while China has better algorithms. Im trying to create one atm but restricted to 5090 RTX.

I should be able to get the good parts of the diffusion models with the help of AI which gives me lesser time to recompile them.

Proteus, Iris, Nyx, Rhea, Artemis, Gaia, Theia, etc.,

[clever lantern]

If China had better algorithms, they wouldn't be months behind. The US is in the lead because their models are better. China's rely on using US models to improve their own, which is why they're stuck months behind.

[small forge]

Thats also correct. If you would be asking a China model for its checkpoint. It actually says a US model.

Mainly Anthropic and OpenAI

They just have ways to finetune llm's or get the good parts.

[clever lantern]

They collect training data from huge amounts of responses from the latest SOTA models. Some of those replies have that stuff about them being Claude or GPT_, and with so many replies in their training data they can't filter them all out manually, so they get baked in.

[small forge]

They be doing that in batches for sure.

[gaunt cradle]

I just found out this channel, I want to share also my video, this one was made with a lot of tooling, including veo 3.1, hope you'll like it: https://www.youtube.com/watch?v=7mk1zVqr2HI

[fickle hare]

Such deep imaging.. and much respect for the lip sync.. I havent found a tool that does that well and I always end up in an editor.

[wraith yarrow]

how can i make a short movie like 2-3 minutes using flow

[pliant shoal]

Can someone help or assist me. I have been tasked to make a short animation and wondered if i can use Flow to create it.

[pliant shoal]

Urban Storyline Concept: From Streets to Sky

Scene 1 – The Hustle (2–3 sec)
• City backdrop: graffiti walls, subway trains rolling by, sneakers dangling on power lines.
• A young person sitting on the stoop with friends, looking tired of the day’s grind.
• They pull out a pack of Flying Papers.

⸻

Scene 2 – The Spark (3–4 sec)
• They roll up smoothly, flick the lighter → the flame ignites, and suddenly the smoke starts forming wings.
• The paper itself starts glowing, then literally lifts off their hands like a jet.

⸻

Scene 3 – The Flight (5–6 sec)
• The character grabs the paper as it transforms into a mini plane made of smoke + paper.
• They hop on, soaring over the city.
• Cutaway shots: flying past neon-lit billboards, rooftops, basketball courts, murals.
• A quick shot of jealous people on the block looking up like: 👀🔥

⸻

Scene 4 – The Message (2–3 sec)
• They land back on the stoop, smooth and calm.
• Pack of Flying Papers glows in their hand.
• Tagline pops up on-screen:

✨ “Flying Papers – Take Your Vibe Higher” ✨

⸻

This keeps it urban, aspirational, and visually trippy, while staying short (15–20 seconds).
⸻
https://www.flyingpapers.com/ - the characters are already made but need them in the story. Thanks

[clever lantern]

I'd split that into four clips, then trim and edit them together. Use keyframes or ingredients for the characters.

[pliant shoal]

Thanks for your reply. Is this something someone can kindly help me - just starting out in AI video creation and want to learn.

[clever lantern]

Then I'd just aim for grtting the first clip right first. Make the first frame of what you want using whatever AI image app you prefer, then use it as the first frame in Flow, describing what you want to happen.

[pliant shoal]

Ok, and this is all doable in Flow at a high standard?

[ornate tinsel]

Hello, everyone as a reminder this channel is for the weekly flow creative challenge prompts. While we have not had one for a while we ask for your help in maintaining the channel’s intended purpose. If you are unsure were to ask a question our #server-directory・☎️ has a list of our channels and their uses. Thank You!

[golden bronze]

This week's Flow Creative Challenge: Liquid Textiles 🫗

we are so back and we're starting with an open-ended prompt for this week's Flow creative challenge: liquid textiles. think of everyday objects and concepts, but turn them into liquid 🌊 showcasing their movement and versatility. what will you create?

How to Start:

1. Head to flow.google
2. Select "New Project"
3. Start prompting!

📣 We love featuring your creations that you share in this channel (and from across the server!) - if you're interested in potentially having your content be shared across Labs socials, be sure to check out the opt-in system in #start-here・📍

[golden bronze]

hey there! images and videos aren't generated within this Discord server. to get started, head to our website at labs.google.

[hallow arrow]

Hello! 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!

[fervent dune]

I'll be working with frames until summer. ☄️ 🫧

[last cosmos]

Oooh I'm digging the concept!

[fervent dune]

Ehh, idk between image 3 & 4 there shouldve been a proper break from the crystals. I'll take it though.
The R&D for this was a bit of a loop.

[fervent dune]

🌊 but in ☄️

[zenith marsh]

NEED ME WATER CAP

[zenith marsh]

Aqua Cap 🙂 get yours now

[zenith marsh]

The workflow for both videos followed the same process 🙂

First, I generated storyboard frames.

These frames were then used as inputs in the Ingredients to Video pipeline to produce the final outputs

[hallow wadi]

https://cdn.discordapp.com/attachments/1331469853646651403/1486377359807480030/Flow.gif

[golden bronze]

oh this is rad

[molten monolith]

Cup of coffee pouring out from a mug turns into liquid velvet. I started with nano banana to prompt, but the image was already mid-pour, so I then just inputted the full video prompt into flow, negating the start image. Sometimes image prompts for start images are not as effective for me.

[devout yarrow]

@dim cradleHello 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!

[indigo fable]

Nice

[fervent dune]

Nice jacket.

[warm hinge]

👐

[golden bronze]

This week's Flow Creative Challenge: Props 🎒 📱

last week, Flow shared a prompting tip around props and how they can shape the scene you're trying to create. this week, we're trying that tip out ourselves for the Flow Creative Challenge 💡 what will you create? find a prop of your choosing and shape a clip or scene around that prop.

How to Start:

1. Head to flow.google
2. Select "New Project"
3. Upload or generate an asset that features a prop of choice
4. Start prompting!

📣 We love featuring your creations that you share in this channel (and from across the server!) - if you're interested in potentially having your content be shared across Labs socials, be sure to check out the opt-in system in #start-here・📍

[fervent dune]

I sense an oil leak on this one.

[fervent dune]

If you squint you can still see the streetlight.

[last cosmos]

Such an odd place! Maybe it got forgotten when the city expanded/got rebuilt

[fervent dune]

Yeah i heard the streetlights are for walking pedestrians & bots in this timeline. Ehh you should see my other outputs some interesting ones for today for sure!

[quiet horizon]

Aliens APC remote control 🙂

[upbeat reef]

This was difficult because it did not want to give me the exact back pack when I extended the clips to what happens next.

[hallow arrow]

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!

[hybrid abyss]

nice job

[fervent dune]

Interesting concept, needs some more work(I could see this as a google pixel ad, but with more refinements eventually), but the creative direction is there, nice stuff @upbeat reef!

[shrewd bronze]

Scary box

[pallid wagon]

Uh oh two went missing

[copper shuttle]

W

[glossy musk]

Decided to try one clip here using a prop and video prompt. Nanobanana 2 for the image ref, prompt for the video below [Fast - Low Priority]:

A close orbiting shot of the shattered antique lantern.

A female ballerina of azure flame composed entirely of dancing blue fire executes her slow, precise dance inside the fractured glass of the lantern.
The erratic, sputtering blue light from her form casts fluttering, shifting shadow patterns onto the surrounding heavy dust motes.
The flame itself is volatile, but its human form remains controlled and constrained by the prop.

Music: Haunting wind up music box ballet tune.

[sick stag]

how do i have watermark VEO on the videos i created in flow and how do i take it off?

[calm ermine]

only on Ultra

[oak gulch]

rate this

[boreal oyster]

Did you use reference image for this? Look real enough

[woven obsidian]

Can someone please tell me why the "Extend" feature from Flow is so broken? The one issue I keep running into is with the Extend feature. Whenever I extend a scene, it never feels like a true continuation — there's always a visible cut, and the second clip rarely matches the tone, or feel of the first. For a tool this powerful, it's a real limitation.
I'm not sure if it's something in my workflow or if there's a better method I'm missing. Have you or anyone else in the group found a reliable way around this?
Would love to hear what's working for people. Thanks!

[oak gulch]

no refrence image just a single pormpt

[boreal oyster]

Yeah it does usually look better without reference.

[glossy musk]

very interesting result on this. You got a good performance and great lyrical timing in this. Text only prompting definitely yeilds unexpectedly good results, often surprisingly good, but any hope of location and character consistency goes out the window making it a bit of a fever dream and supremely hard to wrestle into a narrative. But a very good result.

[oak gulch]

thanks for the appreciation bro

[devout yarrow]

@raw sundialHello 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!

[edgy basalt]

Hi, I’m still getting this error in Flow:

“Failed. We noticed some unusual activity. Please visit the Help Center for more information.”

This has been happening repeatedly, even after waiting and retrying. Sometimes reopening the page lets me generate once, but the next generation fails again with the same message.

This issue is seriously affecting normal use. I’ve also seen other users reporting the same problem, so it does not seem to be an isolated case.

Is the team aware of this issue? Has anyone found a reliable solution?

Is anyone else seriously frustrated with Flow right now?

For days, I keep getting:

“Failed. We noticed some unusual activity. Please visit the Help Center for more information.”

At this point, it doesn’t even feel like a user-side issue anymore.
People in this community have already reported:

• no VPN, still getting it
• waited hours or even a full day, still getting it
• reopened the page, maybe one generation works, then it fails again
• different accounts, different prompts, same result

So what exactly is triggering this?
Because right now it feels like normal use of Flow is being treated as “unusual activity.”

This is not just annoying — it makes the product unreliable and basically unusable for actual creative work.

How many of you are still getting this error right now?
And has anyone found a real fix that works consistently?

[analog skiff]

hey! best place is #1238621511401734254 for any bugs you find! these threads are actively monitored by the team and so you may be able to find assistance in these threads or create your own for the team help!

[hallow arrow]

Hey everyone! Just a friendly reminder to keep all conversations in the supported language for this server. Content must have English subtitles. 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!

[devout yarrow]

@thorn granite@craggy ospreyHello 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!

[rancid relic]

anyone who experiences bad results and inconsistent environment while usinf extend feature

[south yoke]

meh, no, idk

maybe someone will have it

[golden bronze]

This week's Flow Creative Challenge: Share Links Remixing 🔀

We're celebrating the newest Share Links & remixing capabilities in Flow, [announced yesterday](#labs-announcements・📣 message## )

The Challenge:

Behold: The rare deep sea pug in its natural habitat.

With our new Share Link feature, you can see how a video was created and remix it your way! Just click the link, select "Add to Project," and hit "Reuse Prompt" on the video to get started. Then, simply swap our pug for something new to make it yours.

Share your generation here by hitting the three dots > "Share" to continue to the chain 🙂

📣 We love featuring your creations that you share in this channel (and from across the server!) - if you're interested in potentially having your content be shared across Labs socials, be sure to check out the opt-in system in #start-here・📍

[golden bronze]

I'll start ✋ I subbed the 🐶 for an 🦦

[merry quiver]

<@&1366831844502274180> @lyric cedar sry for pinging again but scam msgs

[lyric cedar]

No worries, thank you caring about the community! To give credit where credit is due: @analog skiff was faster than me in responding. Thanks, Bee!

[merry quiver]

Just doing my part 😄 thanks for the quick actions to keep this place safe

[fervent dune]

You took my idea...

😃

[golden bronze]

it's a party! 🦦 🐌 🐶 🦄

unicorn in place of the narwhal, since there's no emoji, sigh

[fervent dune]

I got the edit right here, but the MB is to big, and my compressing sucks this morning. 😆
https://x.com/Gaston_Pay_Up/status/2047349603326910759?s=20

GasMan (@Gaston_Pay_Up)

Testing the boundaries of Veo 3.1 Fast vs. Lite. One handles the speed, the other handles the crunch of 2,000+ character prompts.

Check the tusk glint, managed to hit the threshold on the second pass!
#WeeklyCreativeChallenge @GoogleLabs
#FlowRemix @Google

▶ Play video

X · Apr 23, 2026, 4:19 PM

A narwhal is just the unicorn of the sea, as I tend to hear.

[golden bronze]

pointy horse is 😂

btw - Flow has their own X account if you'd like to tag 'em!! FlowbyGoogle

[fervent dune]

Ahh, I kept getting the X tag mixed up with the gaming flowlabs.io
Which I've used in the past as well, I went the safe route on this one.

Would've tagged you too, but I don't want to intervene too, too much.

[golden bronze]

no worries!! just thought I'd let ya know since I saw the tags 🤓 ty for sharing!

[fervent dune]

Now wheres that #todays-special・🥣 at Miss Otter-lady? 😃

[golden bronze]

https://tenor.com/view/door-hide-close-bye-cat-gif-22820275

haha, working on it!!

[kindred robin]

Uhhmm.. instructions unclear.. https://labs.google/fx/tools/flow/shared/video/99c047ee-0047-49c1-9bd7-138d2db3189e

[fervent dune]

smoooochesssss

[golden bronze]

😂 perfection

[kindred robin]

-1pt. Gibberish text on back of Tshirt

[fervent dune]

But +2 points for how the pink shirt really brings out her eyes. 👁️ 👁️

[devout yarrow]

@lost frigateHello 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!

[lilac dirge]

i am so frustrated with the output of image and video, i use gemini to make the promt and it gives short and simple despite asking it for extensive precise prompt. and then veo build and building halfway and then breaks down the whole building just to match the last image i uploaded. its so painful coz the credits get eaten up for nothing again and again, somebody help and guide me

[dire nebula]

Try Chat GPT, it is better for prompt in my opinion, Gemini is amazing for other things.

[blissful hearth]

why not try to learn how to prompt yourself, you're far more consistent than an AI. Granted, if you lack the language skills to vary and develop a story, then it'll be a challenge, but at least try to be involved in the process a bit more

[keen flint]

Tattoo ideas AI generated via Perplexity AI.

[still jetty]

plz prompt

[blissful hearth]

@still jetty so, a very shortened and "raw" version but just for this explanation. As always start with your initial reference image, then create your final or stepped images. The prompt, which you can be quite flexible with needs only be something like "Create a hyper-lapse of the room, reference image 1, is decorated by a single person. The result will be reference image 2" (I really did simplify it, but you get the picture). What you see here is a single generation - I've not tweaked it at all and I think it did ok. To do better you'd prompt in the stages of decoration - aka paint walls, install flooring, lighting, furnishings, finish

[fervent dune]

This is like visual asmr @blissful hearth.

[blissful hearth]

The servers here do not support any generations. You'll need to do that directly

[analog skiff]

Hi there, images and content cannot be generated in this server. Head on over to https://labs.google/ pick a tool and start generating.

[clever crater]

“A small peaceful village in a lush green countryside, warm sunlight, a curious young boy named Robi

[devout yarrow]

@willow ferryHello 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!

[karmic coral]

https://labs.google/fx/tools/flow/shared/video/e14293ce-9ed0-4966-9c85-10674166b799
Hun, I guess we can remix stuff on flow! As a Sora refugee, this is very cool, but the process is not quite as friendly as Sora. In fact, I had to bang my head a bit to get this done. On the one hand, it was interesting and got me to try some new things in Flow, but it also exposed some pain points.

If I can ask, why do we have insert, remove and camera as separate functions? I somewhat understand the camera one because then you don't have to teach a new user all the Hollywood jargon for camera movement (sure, we can type in "quarter circle orbital tracking shot with the image subject as the central focus", but having a big button with a picture of a camera doing that probably increases accessibility). But the other two? Why just combine them into "Change", with the option of using the marquee select tool?

This "remix" topic exposes why this would be ideal rather nicely actually. What I want to do with the shared video is not really remove or insert, I want to replace. I tried the simple path of taking the video of the underwater dog here and using "insert" with a prompt asking it to re-imagine the background as outer-space and replace the diver with an astronaut, coral reef with Earth, etc, but that failed horribly. What did work here was taking the input ingredients and then remixing them to look like what I wanted, then making the video from that. But then I have to write a new prompt for the video - sure, I can reuse the one that was shared with the video, but now it's describing a new scene in space.

I have no proof of this, but I'm 95% sure that Sora took the user's remix input prompt, considered it against the previous prompts in the chain, did some LLM thinking, and internally used that generated script input to actually make the videos.

[west birch]

https://labs.google/fx/tools/flow/shared/video/80a5c70c-d179-44d5-b1a5-f5d8cc3437e1 remix this

[fervent dune]

https://labs.google/fx/tools/flow/shared/video/50a3e8c7-c604-4392-a80f-731e73d15b75 Re-remix this.

[narrow galleon]

https://labs.google/fx/tools/flow/shared/video/c413edee-a145-4ba5-873c-61d6b81b0032

[drifting pawn]

Thank you so much for your suggestion. I made this for my wardrobe design that I eventually want to do but the results were great. Thank you for your suggestion.

[fervent dune]

<@&1366831844502274180>

[spiral shale]

hello . can anyone suggest which are top 2-3 tools to be used for converting ppt content or or voice over content to animated video ?

[tawdry eagle]

Eleven labs

[upper zinc]

<@&1366831844502274180>

delete this from all places

[spiral shale]

thanks . will it support indian languages ? i am not sure though

[golden bronze]

This week's Flow Creative Challenge: The Race 🏎️ 🏁🏃‍➡️

Inspired by "custom racetracks" shared in #todays-special・🥣 yesterday.

The Challenge:

Try to create a racetrack, a racing scene, or a scene that shows movement and continuity - you pick the setting, the characters, the speed. ready, set, go

📣 We love featuring your creations that you share in this channel (and from across the server!) - if you're interested in potentially having your content be shared across Labs socials, be sure to check out the opt-in system in #start-here・📍

[golden bronze]

otter takes off and thinks he's doing great, until he sees that he started running the wrong way. no worries though because he's racing against a snail. easy win, right?

[tender kayak]

Intergalactic Race 🚀
This is my first attempt at combining multiple video generations into one scene. Focusing on the high-speed flow and portal transitions.

[zenith marsh]

https://labs.google/fx/tools/flow/shared/video/bdb662cd-1eb2-40d9-8a34-ce01c5c06f33

[gentle tide]

<@&1366831844502274180>

[fervent dune]

Hi @proper ember be sure to check the #rules-and-guidelines・📋 when communicating in the discord. Thx!
Oh @devout yarrow

[proper ember]

thanks

[devout yarrow]

@proper emberHello, as Gaston mentioned, the rules channel is a great place to start. Please keep all conversations in a supported language for this server. This helps the entire community engage effectively. Additionally, videos and images are not generated within the server. Please visit https://labs.google/, choose a tool, and generate them there. Thank you.

[upbeat niche]

• Whaaaaooooooo !!!