flow-tips-n-tricks・✏️

[serene bramble]

Has anyone else credits change for fast veo 3 to 20? Quality is still 100 but there’s no sound??

[ionic palm]

Yep

[serene bramble]

It’s so dumb, can’t even make a decent video with sound now

[thick kayak]

[WIDE SHOT – inside coffee shop]
BARISTA is cleaning the counter. The COP enters. Door bell jingles.

BARISTA
(smiling)
Morning, officer! Right on time.

COP
(stretching)
Yeah… long night. Please tell me this coffee is stronger than the paperwork I had to fill out.

[MEDIUM SHOT – barista pours coffee]
Steam rises as he fills a cup.

BARISTA
(cheerful)
Extra strong. Just how you like it—black.

[CLOSE-UP – Cop’s face stiffens]
His smile fades. He looks at the cup.

COP
(flat, tense)
Black?

BARISTA
(confused)
Uh… yeah. No cream, no sugar. Like always?

[SILENCE – tension builds]
The cop slowly sets the cup down. Hand hovers near his gun.

BARISTA
(nervous)
You okay?

[WIDE SHOT – dramatic beat]
The cop suddenly pulls out his gun and FIRES at the cup. A loud bang. Coffee explodes everywhere.

[SLOW MOTION – coffee cup bursting apart]
Cardboard, hot liquid, and steam fly across the counter.

[MULTIPLE GUNSHOTS – the cop empties his clip]
He keeps shooting the puddle of coffee now on the floor.

[AFTERMATH – stunned silence]
The coffee shop is a mess. The barista is covered in splattered coffee.

COP
(to the coffee, breathless)
Black… stop resisting!

[CLOSE-UP – Barista, horrified]

BARISTA
What the hell?! It was just coffee!

[WIDE SHOT – the cop holsters his weapon]

COP
(grumbling)
Can’t be too careful these days…

[He walks out. The bell jingles again.]

[Final Shot – Barista just stands there, drenched in coffee, staring into the mess.]

[frank robin]

It should still be generating with sound on fast mode… if the sound isn’t working try tweaking your prompt

[ruby oasis]

Create a hyper-realistic, cinematic nighttime video in downtown Miami, with colorful city lights glowing, people walking in the background on the sidewalk, and a cool, urban vibe.
The camera is handheld in third-person style (the camera operator is not visible). The subject is a bald, muscular Black TikToker wearing a sleek black button-up shirt (top two buttons unbuttoned) and tailored Pronto Uomo pants. He confidently holds a Mini Mic Pro Wireless and walks up to two women on the sidewalk for a fun street interview.
The two girls are beautiful blonde women with dark-blended glam makeup.
• The first girl is wearing a gray spaghetti strap summer bodycon maxi dress — double-lined, with a flattering fit — and a custom gold necklace that says "Jessica" in cursive.
• The second girl is wearing a black matching crop top and shorts set, with large gold hoop earrings and a shiny gold bracelet.
The TikToker smiles and raises the mic to the first girl and asks:
“What’s the biggest red flag in a guy?”
She confidently replies:
“If he’s broke, he’s a joke.”
Both girls burst out laughing.
Then the TikToker turns the mic to the second girl and asks:
“How about you?”
She replies with a smirk:
“If he doesn’t own two cars, he’s gotta go.”
They both laugh again, and the first girl adds:
“That’s so real.”
The scene captures the confident energy, nighttime street vibe, and raw humor of a viral street interview.
Let me know if you want an alternate version for daytime, a different city, or a female interviewer instead.

[marble brook]

@ruby oasis wow that is really good, it slike reading a page of a book

[heady flare]

anything to stop captions?

[thick kayak]

“A tense, dark comedy scene: A cop walks in, a person says ‘Here’s your black coffee.’ The cop suddenly pulls out a gun, shoots the coffee cup repeatedly. The coffee explodes. The cop yells ‘Stop resisting!’ to the destroyed cup. Cinematic, exaggerated, 8 seconds.”

[merry mirage]

#flow-general・🌊 A 45 year old Vietnamese male dentist is in a dental clinic with a green and blue background, wearing glasses and a white blouse. The dentist says: "Tôi là AI do Kim Cương tạo ra, trông tôi có giống nha sĩ không hehe".

[warped loom]

“A tense, dark comedy scene: A cop walks in, a person says ‘Here’s your black coffee.’ The cop suddenly pulls out a gun, shoots the coffee cup repeatedly. The coffee explodes. The cop yells ‘Stop resisting!’ to the destroyed cup. Cinematic, exaggerated, 8 seconds.”

[thick kayak]

Anyone wanna help me out?

“A tense, dark comedy scene: A cop walks in, a person says ‘Here’s your black coffee.’ The cop suddenly pulls out a gun, shoots the coffee cup repeatedly. The coffee explodes. The cop yells ‘Stop resisting!’ to the destroyed cup. Cinematic, exaggerated, 8 seconds.”

[foggy bay]

#flow-general・🌊 a POV of a ferret in the jungle walking on two legs, hes energetic and very human like and saying " alright i dont see any human out here , im planing the trip to hawai you know

[azure seal]

I’m trying to generate a bigfoot video of my car could anyone help me generate a prompt

[smoky current]

Hey.. can some one please clarify teh difference between Google AI Pro and Google flow. Is it same or different. I am subscribed the AI pro and when I try to enter flow it says "Flow is not available in your country yet"

[smoky lark]

any way to keep them from "mouthing" words in VEO 2?
I've tried she's thinking, listening, and i get spotty results.

[mild nova]

"A cinematic, slow-motion video of a vibrant Holi celebration in a small Indian town. Wide shots of narrow streets filled with color powder in the air, people laughing and dancing. Close-ups of faces joyfully smeared with bright colors — pink, yellow, blue. Traditional Indian clothes, sarees and kurtas flowing in motion. Background of rustic homes, marigold garlands, and temple bells ringing. Smooth camera movements, golden hour lighting, warm color grading. Music is upbeat and festive, hinting at Indian folk percussion. Include drone shots above the crowd and intimate handheld shots of kids playing with water balloons."

[swift pulsar]

Its working fine using mixed scenes with the same context, still not able to have the same character in diferent prompts with audio (without using the extend function), any tips?

[molten parcel]

#flow-tips-n-tricks・✏️⁠Animated thermal camera-style footage of a fictional soccer match. Surreal, psychedelic color palette: cool areas (like grass) appear in deep blues and purples, while warm areas (like animated players) glow in bright reds and oranges. The background is pure black, creating a striking contrast. Smooth, continuous tracking shot follows the motion of the ball and players. No real teams, logos, or identifiable likenesses—entirely original characters and setting

[jade leaf]

is flow available in Portugal

[fringe badger]

Hey guys, I trying to make a scene fight 2 VS 1 like a 80s movie but sometimes Veo can't understand the prompt correctly, any prompt to make a gutpunching scene or a fullnelson grab?

[molten parcel]

I am trying to make a vlog style video with the main character traveling back in time to tell a story of a famous event in time.

I generated the first clip. He says “hey guys blah blah” - but the clip ends before he finishes his sentence.

1. How do I make the character talk a bit faster? Just the last two words cut off.
2. How do I make the next scene with the same character telling the next piece of the story?

Anyone have any tips for creating a video like this with the main character being the same but in different locations, talking about different things (events in the timeline of the famous historic event)

[mortal iron]

idk about 1 but for 2 use fixed seeds

you can use whisk for same main character

just upload a single photo ur main character on the character thingy

[lavish olive]

Anyone know some potential reason I am getting no audio in all of my Veo 3 generations? I am including dialogue for them to say..

[hardy prairie]

Double check to make sure its veo3.. sometimes it automatically switches back to veo2

[heavy pawn]

How to achieve a drone shots without showing the drone itself?

[wheat fog]

Anyone have any tips to get consistent dialogue ? most of my prompts never even get sound smh

[wheat fog]

anyone else get silence clips ???

[earnest cedar]

YES

damn bug

[wheat fog]

has to be because ive changed all of my prompts and they were not harmful or offensive at all

@timber estuary @olive cove im expecting some sort of refund or compensation

[earnest cedar]

Same here

@severe aspen what we should do? Do you know something?

Sorry if you’re not the right person to ask

[muted quartz]

Can someone help me with a prompt using veo-2? I have a start frame and an end frame and im trying to do 2 things

1: keep the video vertical, or atleast keep the text small enough to be in frame when I crop it to 9:16.

2. Make it one seamless transition. It keeps either doing a fade in for the next frame or just doing some weird transition.

Here is my current prompt:
The camera glides smoothly out of the forest and emerges onto a gentle beach shoreline at sunset. The camera transitions to a top-down view of shimmering waves washing over golden sand. As the waves recede, the sand is revealed with the words “road trip” and “6 / 20 / 25” in the Beach Resort font, written in the sand. Warm pastel light, Pixar-style 3D animation, magical and cinematic.

This is probably the best results I've gotten

[maiden glacier]

This short video created using veo3(fast)

[river panther]

I cannot create video with audio

[rancid iron]

I get alot of dms asking me how I made that video that gave the illusion it was the same character. but it wasnt 100% identical. I have got close with 96-97% identical match using this method and this method alone using only veo 3 Quality. I did not 100% this. I ran out of resources. this was what I tried. Hope it helps if anyone is tinkering like me. Just a experiment I tried on my off time.

This system is called MSSv7.5. It simulates memory in Veo 3 by locking everything visually and emotionally across scenes. The illusion of memory comes from repetition, not stored data. We built this by identifying what causes AI to drift between clips and manually locking it all. Each time it broke, we added a new layer. It now includes Character Lock, Audio Lock, Environment Lock, Camera Lock, Symmetry Lock, EchoDNA, NeuroSim, and TRFL (emotional correction). This is the full working template:

Subject: Real human male. Mid-30s. Olive skin tone. Short dark brown hair with slight wave. Trimmed stubble. Slight tension in brow and under eyes. Wearing a dark gray hoodie with neckline slightly stretched. Hood down. Standing still, facing camera. Calm but intense.

Environment: Natural daylight coming from the left at a 45-degree angle. Subtle shadow cast on the right side of his face. Neutral white background with no visible decor. Background must not change.

Camera: Front-facing fixed position. Medium-tight frame (head to upper chest). Eye-level. Subject is perfectly centered. No camera motion or lens distortion.

Symmetry: Head level. Chin neutral. Eyes aligned horizontally. No head tilt. Shoulders even. Posture must match original anchor frame with no slouching or lean.

Audio: Natural voice. Calm delivery. Slight breath between lines. No acting. Feels like he is continuing a thought already in progress. No sudden changes in volume or tone.

EchoDNA: This must feel like the same person in the same moment. Not reappearing. Not restarting. Continuing.

NeuroSim: Match micro-movements, eye blinks, lip tension, and breath rhythm from the previous clip. Maintain emotional momentum and body memory. No fresh reset. Output must feel like the scene never cut.

TRFL: If the output breaks emotionally, even if it matches technically, adjust the next prompt based on human intuition and reaction. Final validator is perception, not code.

This is how we simulate memory in a system built to forget.

HOW TO USE MSSv7.5 + EchoDNA + NeuroSim
Pick your anchor.
This is the first clip or frame that defines everything. It is the only visual truth. All future prompts must match it exactly.

Copy the full template.
Paste the entire MSSv7.5 block (the one you just got) into your prompt field or scripting doc. This is your foundation.

Update the subject and setting ONLY if needed.
If you’re using a different character or scene, swap the values in “Subject” and “Environment,” but do not change formatting or tone. Keep structure identical.

Write new scenes as a continuation.
Every prompt must feel like the same moment is continuing. Don’t say “He begins speaking.” Say nothing about starting. You’re picking up mid-thought. Imagine the cut never happened.

Do not let Veo freestyle.
Avoid vague prompts. Avoid energy shifts. Do not let Veo “interpret” — lock it down. MSS is a system, not a suggestion.

Visually compare every output to the anchor.
Check hair, face shape, eyes, hoodie fit, lighting, camera angle, and posture. If anything changed, reject it. Restart from the anchor.

Use your human perception as the validator.
Even if the AI says it's close, your eye decides. If it feels wrong, it is. Use TRFL — adjust your next prompt based on what was missing emotionally, not just visually.

Repeat the template every time.
Never reuse a broken output. Never drift forward. Always return to the anchor if you break. Consistency comes from precision, not speed.

Optional: How to continue a series
Use the same MSSv7.5 template for every scene

Only change the dialogue and maybe a tiny shift in emotion (if part of the story)

Make sure every new prompt sounds like a direct follow-up to the previous line

Keep NeuroSim running by observing micro-behaviors like blinking or tension

This is how you use the system.
You are not prompting.
You are simulating memory.

Every prompt is a continuation of one truth.

This is the scaffold. Nothing else matters.

[sacred sequoia]

What is prompt?

[slender spire]

"First-person, intensely shaky handheld POV, filmed on a low-light capable camcorder, mimicking found footage. Grainy texture, subtle lens flares from natural light sources, and minor digital artifacts (e.g., faint timecode overlay in corner - implied visually). A brief, chilling glimpse of something moving rapidly between distant trees. A massive, bipedal ape-like creature, approximately 9 feet tall, covered in thick, matted, dark reddish-brown fur. Its eyes are large, intelligent, and subtly reflective amber, set deep in a broad, powerful brow. Its gait is a heavy, deliberate stride, sometimes transitioning to a powerful, ground-eating lope. Its muscular build is evident, especially in its broad shoulders and long, powerful arms. Its features appear minimalistic, almost ghost-like in the distance. Deep, temperate forest, dense foliage, dappled light filtering through the canopy, a palpable sense of isolation and primordial wilderness. A sudden, sharp intake of breath from the cameraperson, conveying raw disbelief and creeping dread."

[sacred sequoia]

Perfect, brilliant 😮‍💨🤝👍

[obtuse bramble]

me too, did u fix it?

[viral timber]

I have the same problem

[brave idol]

Use a VPN and loads it

[fossil violet]

Helllo all ! Flow - 🤯 = but , are there any good sources to replicate what we see in Flow TV or at least get close to whats happening on there , i know i can just copy and paste and fill in the best i can but maybe some one has a nice birds eye view cheat sheet or the like ,. Thanks in ADVANCE! 🧸✨✨

[timber heath]

can someone try this prompt for me? its for my english test: a 1980s found footage in a security camera of a alien walking from behind of a space ship and hitting a dynamic dances moves, area 51, no color, night time, from above, far away

[fossil violet]

sorry to poke out of the blue - is this a workflow you created or is teh MSSv7.5+EchoDNA + NeuroSim system instructions or an outside plugin? sounds pretty good!

[rancid iron]

It was something I created with chat gpt(echo) and Gemini

[obtuse bramble]

did u fix it

[rancid iron]

Basically tried building a procedural brain, two ai. Me. Left brain, right brain, and the eye. The perception

[viral timber]

yes i did

dm me if u want i explai nto u

[whole zealot]

Can you make these vids in here

I’m based in uk but won’t let me sign up

[whole zealot]

What’s the prompt

[obtuse bramble]

i did answer pls

[visual night]

Can you please share your prompt so others can learn how to be more consistant?

[obtuse bramble]

can u answer me please @viral timber sorry for ping but i rlly want to fix it it’s been weeks like that

[bright anvil]

hi everyone! can I please get the latest guidelines for efficiently using VEO 3 prompts? I do not want to continue wasting credits, I seem to not get any real prompt adherence. thank you! I.E. I need the most efficient prompt structure and the does and don'ts

[prime sky]

I use chatgpt for prompt

[bright anvil]

hi! it would be a massive help if you can provide a prompt structure for consistency. Feel like I wasted a lot of credits with not much success.

[earnest cedar]

you can personalize gemini or chatgpt to be a prompt engineer

[viral ice]

https://www.youtube.com/watch?v=nEiwV12PmE4

[bright anvil]

I know how to write my prompts, but that is not what I was asking. I was asking for the VEO 3 prompt structure. Each model has its won structure, that if you folow you get prompt adherence

[ruby stag]

A cinematic scene in a chaotic traditional Indonesian market. The camera shakes as it follows an old grandmother, worn but fierce, running with a rifle through the panicked crowd and rising smoke. Stalls are overturned, people are screaming, small explosions echo in the background. The grandmother turns toward the camera and grumbles in Indonesian:

"Dasar Ambiya, gak ada otak! Gua udah tua gini disuruh perang. Emang anjing nih anak!"

Her face shows a mix of anger and resignation. The camera cuts to a close-up with high contrast color tones and dim sunlight, adding a dramatic and epic atmosphere.

[upbeat rover]

A small brown monkey wearing a red hoodie and sunglasses, shopping in a supermarket like a human. It pushes a cart, looks at fruit, and checks a list. Bright lighting, casual vibe.

[rancid iron]

Already shared above

Already shared not too far above

[hallow spade]

I need help with flow,
Im from Colombia, i have access to ai ultra, and i try to do a video with flow, I need my character to be able to speak Spanish, and i wrote all the prompt in english, but i wrote the speech in spanish and flow say to me than i cant speak nomore in spanish

[hybrid stream]

put the spanish to be spoken in quotations

How do I use the same character in Flow w/ veo 3 with Pro subscription?

[dusty bluff]

anyone know how to fix lip sync issue ( whenever i create character ) he is not talking or saying anything

[tranquil sky]

well
I like gorillas, and thought about me before this prompt

[marble brook]

I am having so much trouble getting aspecific prompt down

I'm burning through text to audio without reaching results that I want

[tranquil sky]

sounds silly but works - try less thinking, less trying and less working. keep it simple in keywords and just let chatgpt do the prompting for you. give him your base idea and keywords and give the task, to create a Master-Prompt for (veo3). and the task to give always suggestions at the end, next steps etc..., and short questions for the user to keep his design wishes.
I let gpt make the prompt, than, even if already good, I tell him he should redo it and make his best prompting work. SEO, viral, perfect for cinematic videos, [very detailed], [or what you want to include]. after he regenerated, I do this again, and if not already good, I just regenerate until gpt got the best for me.

secret tip: use "old"-words or more out-of-fashion phrases, old school phrases, or just like "less-used" words, for really good stuff! (example: I had some weird, ugly, acid styled scene with robots and human brains and aliens - which I descriped the scene as grotesk - its not an unknown or used word, I never would use it but for veo (or other) are such keywords = GOLD!

[marble brook]

I'm trying to make a sporting prompt and I cant seen to get the players to stand at the specific location . And do the right movement. Let's say im saying jumpsuit the scene ends up being a different throw and outcome

Thanks for the tip BTW

[tranquil sky]

yeah I cann see why it gets troubles with that :/
maybe try concentrate in the end frames to focuss on less, or just on one thing as workaround

(from chat here)

[some shots references:]
//PAN (Left / Right / Up / Down),
Horizontal or vertical pivot of the camera from a fixed position
//Low Angle Shot,
Camera looks up at the subject, emphasizing power or intensity
WHIP PAN,
//A fast, motion-blurred pan between subjects or scenes, often used for dramatic or comedic effect
WIDE SHOT,
//Establishes setting and scale, shows subject in full within their environment
//Cowboy Shot,
Frames the subject from mid-thigh up, often used in confrontational or cinematic contexts
//Over-the-Shoulder Shot (OTS),
Camera peers over a character’s shoulder, used in dialogue or perspective-driven scenes
//POV Shot,
Represents what a character is seeing, places viewer in their shoes
//Push-In,
Camera slowly moves toward the subject to build intensity or focus
//Pull-Out,
Camera withdraws from the subject to create emotional distance or reveal context
//CRASH Zoom,
A sudden, rapid zoom in or out used for shock or emphasis
//Vertigo Effect (Dolly Zoom),
Simultaneous zoom and dolly in the opposite direction to distort depth and create tension
//Camera Roll,
Rotates around the lens axis, often matches or contrasts with character movement or disorientation
//Tracking Shot,
Follows the subject’s movement, usually forward or backward
//Trucking Shot,
Camera moves left or right alongside the subject using lateral movement
//Arc Shot,
Camera circles around the subject using dynamic horizontal or vertical motion
//Motion Blur,
Can be applied to fast camera movements like whip pan or crash zoom for cinematic effect
//Handheld/Shake Effect,
Adds realism or chaos, can be subtle or dramatic depending on tone
//Rule of Thirds (Composition),
Subject is aligned along gridlines to create balanced, visually pleasing frames

[final tangle]

A professional real estate agent stands in front of a vast open green land under a clear sky. He confidently gestures with his hand and speaks directly to the camera. Tamil voice-over: "Indha nilam unga kanavu veedukka perfect! 100% vasadhi, vasool, vasandham." His lips move in sync with the dialogue. Add smooth camera pan, soft depth-of-field background, ambient village sounds with light breeze.

Create a hyper-realistic, cinematic nighttime video in downtown Miami, with colorful city lights glowing, people walking in the background on the sidewalk, and a cool, urban vibe.
The camera is handheld in third-person style (the camera operator is not visible). The subject is a bald, muscular Black TikToker wearing a sleek black button-up shirt (top two buttons unbuttoned) and tailored Pronto Uomo pants. He confidently holds a Mini Mic Pro Wireless and walks up to two women on the sidewalk for a fun street interview.
The two girls are beautiful blonde women with dark-blended glam makeup.
• The first girl is wearing a gray spaghetti strap summer bodycon maxi dress — double-lined, with a flattering fit — and a custom gold necklace that says "Jessica" in cursive.
• The second girl is wearing a black matching crop top and shorts set, with large gold hoop earrings and a shiny gold bracelet.
The TikToker smiles and raises the mic to the first girl and asks:
“What’s the biggest red flag in a guy?”
She confidently replies:
“If he’s broke, he’s a joke.”
Both girls burst out laughing.
Then the TikToker turns the mic to the second girl and asks:
“How about you?”
She replies with a smirk:
“If he doesn’t own two cars, he’s gotta go.”
They both laugh again, and the first girl adds:
“That’s so real.”
The scene captures the confident energy, nighttime street vibe, and raw humor of a viral street interview.
Let me know if you want an alternate version for daytime, a different city, or a female interviewer instead.

[turbid lintel]

As an option, re-edit the video in a video editing program (Adobe Premiere, Vegas, Resolve)

[elder wolf]

#flow-tips-n-tricks・✏️ A Robot toy that comes alive under the bed of a kids bedroom at night

/create prompt: A Robot toy that comes alive under the bed of a kids bedroom at night

I dont know how this works at all. 😉

[spice folio]

Lol

[spice folio]

https://g.co/gemini/share/d00c50579834

[spice folio]

[twilit widget]

Person over 50, seen from behind or profile, looking at a sunrise. The sun is stylized as a glowing network of digital connections or binary code. Holding a tablet. Colors of a sunrise merging with digital light. Hopeful and expansive. Text: 'Tu Horizonte Digital'."

[hybrid stream]

Are you using a model with Beta Audio?

[safe owl]

Guys, do you recommend in VEO3, first generating the images individually and then generating the video?

[sly wind]

dog selfie camera, shot from extended arm perpective in prayagraj sangam watching pooja and he says in hindi " Guys, ye mauka life mai bs ek baar aata hai, acha hua maine apne maalik ko nhi laaya, agar aapko mera mini vlogs pasand hai toh follow krre mere daily vlogs ke liye"

[minor crater]

Hey guys. sorry if this isnt the right place or structured well, however this is my first message in any discord server ever and I'm only doing this because I can't find an answer anywhere else. I am struggling to get Veo to output any videos in a 9:16 aspect ratio, I've tried several prompts with no luck. has any one been successful with this?

[round grove]

i have not been able to over-ride the 16:9 default aspect ratio.

also, i have not been able to over-ride the default lenght of 8s despite clear instructions to make shorter clips, they always come at 8s. anyone know how to change or over-ride the default?

[azure siren]

A beautiful, confident Black woman, wheelchair user, is in her spacious, accessible closet. She sits gracefully in front of a large mirror, wearing a stunning lingerie set — comfortable, sensual, and designed to embrace her curves. Soft lighting highlights her skin and the fabric’s texture. She looks at herself with pride and admiration, smiling with self-love.

She’s a successful entrepreneur, completely independent and in control of her life. With poised, elegant movements, she turns her chair, reaches for a light blazer and shoes from the nearby rack. The camera follows her gestures gently, reinforcing her autonomy.

Cut to an exterior scene: she’s leaving her home confidently, in her motorized wheelchair, and entering her adapted car with ease. She’s glowing with joy and readiness. The video ends with her adjusting the rearview mirror and smiling — ready to conquer another day.

Visual style: hyper-realistic, soft natural lighting, elegant and empowering ambiance. Focus on facial expressions and lingerie textures. Smooth, inspiring camera movements.

Emotional tone: empowerment, self-love, representation, independence, beauty in authenticity.

On-screen text (as transitions or voice-over):
“Dress in power. Dress in your truth.”
“Beauty is being fully, unapologetically you.”
“Vem de Renda — lingerie for every woman who owns her story.”

Final frame: brand logo and @vem_derenda

[muted quartz]

Can I get some prompt assistance?

[dusty bluff]

Yes

[lost robin]

I tried to create an actual EP 1 , this has now 2M views in meta

https://www.facebook.com/Komikis16/videos/1961619757978267/?mibextid=9drbnH&s=yWDuG2&fs=e

[dusty bluff]

Nice work mate

[tranquil solar]

Can I get some prompt assistance?

I want to generate an image to video, the image has a cars in the roads, how can I make a prompt to make all the cars stop from moving? Like i want them to just stay still?

[reef prairie]

hey guys, you guys watched the Rasta Monkey videos on Tik TOk? How can I recreate the same exact monkey for funny videos? need help

[heady flare]

Anyone have a good realism prompt for a yeti. 60% of the time they look real 40% it’s a waste of credit.

[inner quest]

Medium wide shot, interior of a humble and rustic house in the sertão of Bahia, Brazil. Natural light filters modestly into the room, highlighting the simplicity and worn textures of the impoverished surroundings. A journalist, wearing a suit and tie, is standing and holding a microphone towards an elderly man, "Seu João," who is approximately 95 years old with a time-worn face. Both men are standing to give a broader view of the simple environment. The journalist asks with a clear, professional voice: "Qual o segredo para viver 95 anos, seu João?". Seu João looks at the journalist and replies in a warm, aged voice with a strong caipira accent: "Muito caldo de mocotó, dry, ice e amassos na minha véinha." The audio should be a clear recording of their spoken lines with minimal ambient sounds. There should be no on-screen text, graphics, or subtitles. The overall feel must be authentic and grounded, as if capturing a raw, genuine, and exclusive moment.

[fading latch]

How do I make a prompt?

[weary pewter]

checkout Flow TV, there are several examples there, also you can use ChatGPT to generate VEO 3 prompts. https://labs.google/flow/tv/channel/dream-factory/m7VNi046MIbVQjddDzqI?random=true

[distant magnet]

blue cat playning football und pinbk sun and commenting in ukraininan as its scoring a goal against a goal keeper kanguru.

/create: blue cat playning football und pinbk sun and commenting in ukraininan as its scoring a goal against a goal keeper kanguru.

/create blue cat playning football und pinbk sun and commenting in ukraininan as its scoring a goal against a goal keeper kanguru.

/create blue cat playning football und pinbk sun and commenting in ukraininan as its scoring a goal against a goal keeper kanguru.

/create blue cat playning football und pinbk sun and commenting in ukraininan as its scoring a goal against a goal keeper kanguru.

/create blue cat playning football und pinbk sun and commenting in ukraininan as its scoring a goal against a goal keeper kanguru.

blue cat playning football und pinbk sun and commenting in ukraininan as its scoring a goal against a goal keeper kanguru.

/cre

[surreal merlin]

Amazing video what is the prompt you used?

[vague spear]

https://g.co/gemini/share/70c5583c28ad

[bronze thorn]

don't extend. reuse prompts and only change a few things on every jump

[surreal merlin]

When you make multiple scenes of a human character, how do you make sure the look and face is consistent?

[maiden glacier]

Keeping the look consistent across scenes mostly depends on how detailed and specific the prompt is.

[surreal merlin]

Thanks so for now have to describe very detail how each character looks to make it as consistent as possible right?

[maiden glacier]

Yeah, exactly.
To keep the look consistent, the more detailed your prompt is, the better the AI can reproduce the same appearance in different scenes.

For example, here’s the full prompt I personally used for the video above:

---

📽️ Visual Description (Prompt):

A young man stands alone in the center of a war-torn battlefield under a dark, storm-filled sky. His black wavy hair thrashes in the violent wind. Glowing white eyes shine through the darkness. Thin black lightning veins crack from beneath his eyes, crawling across his skin...
(...rest of the prompt as you wrote it)

---

You can see how I tried to describe everything — from lighting and camera angle to hair, eyes, skin texture, and even atmosphere — in as much detail as possible.
That level of specificity really helps keep the character’s look consistent, especially across cinematic scenes.

[fervent bridge]

Does anyone know why this keeps going to Veo 2 and not letting me generate with audio?

[odd birch]

Please use the appropriate channel when posting content. #flow-videos・📹 is for video content. Keep this channel for prompt discussion. Thank you kindly! 😺

[light ridge]

@maiden glacier @maiden glacier

[unkempt relic]

#flow-general・🌊

[untold wagon]

Samething happend to me all the time, I think is a glitch

[surreal merlin]

Looking for ideas how to replicate dance moves - for example if I want the AI video character to dance the same way as a certain dance, how would I do that, it’s kinda impossible to describe every dance move perfectly just by text

[verbal wyvern]

Anybody creating characters with photorealism in whisk and then importing them into Flow to create a series of actions ?
i want to make an instructional video for preschool teachers based on real teaching moments but i want to generate the humans in whisk and create the video in Flow so that I dont have to use images of real children in the training videos. The problem is i need to control details like the materials, signage on the classroom walls, materials being used etc so they all reflect best practices. For now im going to leave out audio and we will just see a caring teacher crouched down speaking to a child who has had a big upset and I can use subtitles or voiceover to explain what she is likely saying to be supportive. But even getting the sequence of actions of a child , ,for example trying to build a tower but bumping it and then sobbing is really hard to keep any consistency on materials child etc....advice?

[frank robin]

I’ve found trying to do anything involving kids produces wildly inconsistent results and most of the time no audio.

I’ve tried a scene where a kid says something to their parents and burned a lot of credits trying to do so. Got slightly better results when I changed the kid to an “18 year old” but if you are trying to get something with a young child I have a feeling it’s intentionally difficult (probably for the best tbh)

@verbal wyvern

[lean viper]

@verbal wyvern Anything to do with children returns no voice.
The camera opens from a father's point of view. His wife and their two young kids, a son and daughter, are in the kitchen smiling. Together, they hold out a plate of pancakes.

The wife says, “Happy Father’s Day! We made your favorite breakfast.”
The daughter exclaims, “You’re the best dad ever!”
The son shouts, “We love you, Daddy!”
All three of them are grinning, filled with love and excitement.

Any other variation of anything with the word child or young kid, etc. the video is ok, just no voice at all.

[timid pivot]

probably describe the name of the dance. it worked well for with Sora - name the dance and the music type and it should try to keep the beat - if you put more than one person doing it that also helps it get some type of cohesion. i.e: two buxom blonde women wearing green Lederhosen dancing one top of a table at Oktoberfest. *see what happens

I forgot to put “doing the Riverdance” or something like that…like I said, usually works well across the spectrum of AI TTVG’s when it’s more than one person and there’s a dance name attached

[ruby stag]

An old Indonesian grandfather, with a white beard, wearing shorts and a traditional headband, is racing against a large hippopotamus on a professional running track. Spectators cheer from the stands. The grandfather runs with all his might, laughing loudly, while the hippo looks serious and sweaty. The scene is funny, absurd, and full of energy.

[drowsy blaze]

Hey everyone! I tried using this prompt in Voe3 but the result didn’t come out the way I expected. Can someone help me figure out if the prompt is too long or if something’s wrong with it?

Here’s the prompt I used: The scene takes place in Buenos Aires at sunset, with the iconic Obelisco and the Argentine flag waving gently in the background. A blonde woman in her early 20s sexy figure and long straight hair comes running toward the camera in a hyper, excited way. She’s wearing a playful and stylish outfit that highlights her movement and energy. As she reaches the camera, she grabs it with both hands — we don’t see the camera itself, but the image shakes slightly to simulate her grabbing it. She looks directly into the lens, wide-eyed and breathless, and says with overwhelming excitement: “Oh my god… DJDTime’s new track is driving me crazy!” After saying the line, she quickly turns and runs off with the same energetic vibe, disappearing into the lively Buenos Aires backdrop. The mood is playful, sexy, and cinematic. Do not include subtitles or on-screen captions

[weary pewter]

I had the same experience, it did't even allowed me to generate that video. And in one case no audio

[lost robin]

Your prompt is too long to be cramped in 8 secs if you want her to stand still while saying your script

[dim gull]

Explain what missed the mark, you only mention it didn't come out the way you expected.

[lost robin]

Hes expecting for the girl to standstill and say the script before running, also instead of grabbing the cam, she is holding a cam

[weary pewter]

Hi everyone, I have been playing with VEO 3 since last few weeks now and learned alot, wanted to share some issues that i experienced and if anyone of you can give me solutions to any of these would love to learn from you guys. Thanks a bunch.

• Creating 9:16 ratio vidoes without getting left and right side of the video blurred.,
• Text display is often getting messed up, it either misspelled it or completely garbage text is displayed.,
• Character consistency in multiple generations.,
• Sometimes dialogue delivery is messed up b/w characters. Male dialogues delivered by female, and female dialgoue delivered by male etc.,
• When i am trying to extend a scene generated from VEO 3, I am not allowed to use VEO 3 for extended scene, only VEO 2 is avaialble?

[grizzled quarry]

Hey everyone! I tried using this prompt in Voe3 but the result didn’t come out the way I expected. Can someone help me figure out if the prompt is too long or if something’s wrong with it?

Here’s the prompt I used: The scene takes place in New York City at sunset, with the iconic Mississippi flag and the American flag waving gently in the background. A blonde woman in her early 20s sexy figure and long straight hair comes running toward the camera in a hyper, excited way. She’s wearing a playful and stylish outfit that highlights her movement and energy. As she reaches the camera, she grabs it with both hands — we don’t see the camera itself, but the image shakes slightly to simulate her grabbing it. She looks directly into the lens, wide-eyed and breathless, and says with overwhelming excitement: “Vaquero Artisan is the very very best ” After saying the line, she quickly turns and runs off with the same energetic vibe, disappearing into the lively New York City backdrop. The mood is playful, sexy, and cinematic. Do not include subtitles or on-screen captions

[grizzled galleon]

I have issue with the sound everytime is no sound

[weary pewter]

are you sure you are using VEO 3 model, it works for me. VEO 2 does't give you sound. Also if your prompt is related to children, it may not generate sound

[grizzled galleon]

I did try Veo 3 and same result with flow

[weary pewter]

hmmm that is weired, did't experience that for VEO 3. May be there is something in your prompt which is cuasing this.

[grizzled galleon]

I just testing some prompts that someone did

[lost robin]

Hope youll like it

[weary pewter]

@drowsy blaze you seems to be the hottest Dude in the town LoL 😁

[weary pewter]

As usual good job on this one 🙂

[dim gull]

Yikes man...

[drowsy blaze]

Hey KOMIKIS! Thanks a lot for making the video 🙌 Would you mind sharing the prompt as an example of how you did it? I'm also burning through credits like crazy 😂 Out of every 10 tries, maybe 1 turns out good!

[drowsy blaze]

Thanks for the info, Zacro! I’m going to give prompt-alchemist a try. 🙌

[dim gull]

Finalizing a few things, It'll be released soon ™️

[drowsy blaze]

LOL thanks ather.ali! , Big thanks to everyone here!

[grizzled galleon]

I got the sound working,the program does what it wants sometimes very close and mostly very far from what was the prompt

[sinful mountain]

@drowsy blaze A dramatic 4D cinematic night war scene. The camera opens from a high aerial angle, revealing a dark navy-blue sky pierced by a massive barrage of hypersonic ballistic missiles falling towards a partially ruined city. Trails of glowing fire streak across the sky. Each missile impacts with devastating explosions, sending up glowing fireballs and smoke clouds over the crumbling buildings.

Cut to ground level: dozens of anthropomorphic mice in ragged tiny clothes run in panic through the burning streets, dodging falling debris and flames. They rush toward underground burrows, pushing and shoving to hide. The camera tracks their desperate escape, showing close-ups of terrified faces.

🎧 Sound Design Cues:

High-pitched hypersonic missile whistling overhead.

Earth-shaking explosions with intense bass.

Overlay multiple mouse voices with expressive Arabic panic:

🐭 (screaming): "يوووو يـووووو!"
🐭 (panicked male): "Oh my God!!"
🐭 (frantic voice): "أريد حفاظة بسرعة!"
🐭 (female voice, high-pitched): "آآآآآآآآاااه!!"

Lighting: cinematic volumetric lighting. Orange-red flickers from fires and missile trails, mixed with blue backlight from the night sky.
Camera: dynamic tracking shots, crash zooms on mouse faces, wide explosion frames, slow motion as debris rains down.
Environment: detailed post-apocalyptic urban zone, cracked streets, glowing ash, shaking camera effects for realism.

[lost robin]

This may generate but your element composition is already too much and confusing for veo 3

[honest cipher]

[8-Second Script | Warhammer 40K | Thousand Sons | Ice Planet | Night Scene]

[Setting: Frozen tundra under a starlit sky. Ancient ruins jut from the ice. Wind howls. A Rubric Marine, glowing faintly with sorcerous energy, kneels before the towering form of Magnus the Red.]

⸻

RUBRIC MARINE (deep, echoing voice):
“The warp is silent, my lord… even the dead fear this place.”

MAGNUS THE RED (commanding, mystical):
“Then we shall make it scream, brother.”

[grizzled galleon]

This reminds of early stages of Midjourney

[keen mason]

inspired by @elfin aspen 's awesome prompt

[olive saffron]

Hey creators, I'm wondering how I can create videos with images to video from exiting image. At present I'm not getting the results I want. The image I'm using is a studio portrait of an AI influencer and I want to show that this influencer is having a drink in talking script to camera

[olive saffron]

Hahaha love the prompt

[twin walrus]

any video prompt tips for making awesome food commercial ads? I have a small food truck business and I have been experimenting. Its so much fun! 🙂 Thanks

[quiet goblet]

promt plsss

[blazing burrow]

/create बंदर जंगल में घूमते हुए कैमरे के सामने आता है और कहता है कि वह आज अपने फेवरेट फ्रूट यानी केला ढूंढने निकला है। वह पेड़ों को गौर से देखता है, कभी छलांग लगाता है, कभी पत्तों को हटाकर झांकता है। फिर उसे दूर एक पेड़ पर पका हुआ पीला केला दिखता है। वह खुशी-खुशी उस पेड़ पर चढ़ता है और केले को तोड़ लेता है। व्लॉग में वह दर्शकों को बताता है कि ये जंगल में फ्री में मिलते हैं, और यही जंगल का असली मजा है!

[molten rose]

hey, can somebody help me with a prompt that can make animals look more realistic?

[static moth]

Add macro semi realistic anthropomorphic to the beginning of the prompt

[molten rose]

what do you mean exactly by that? (I dont really get it :(

[static moth]

That’s supposed to make the animals realistic

[molten rose]

what is anthropomorphic?

[static moth]

Anthropomorphic means giving human traits to animals like walking on two legs, wearing clothes, or showing human emotions etc

[molten rose]

ah okay. thanks. but what about the apperance? because right now it looks like its 3d more than a realistic version

[static moth]

You might need to remove ‘anthropomorphic’ from the prompt, and you can try adding things like ‘photo-realistic,’ ‘high detail,’ or ‘natural lighting’ to make it look more realistic

[molten rose]

okay, I'll try it out. thanks for the help

[static moth]

You’re welcome

[worthy pine]

A video of "What if Jurassic park was located in central park?"

[keen axle]

is there a way I can make the videos come out in 9:16 portrait not 16:9 landscape?

[latent owl]

A black superhero with tribal armor and glowing blue eyes jumps from a rooftop in slow motion, comic book animation style, cel-shaded, dramatic shadows, vibrant colors, cinematic action framing, inspired by Marvel animated movies, detailed but stylized, cartoon hero aesthetic.

[sage cipher]

You have to go to Flow AI (google it).

[limber token]

holy

[dense rune]

My video didn’t have sound, used veo 3. Any reason why it wouldn’t?

[dapper river]

/imagine

[mortal iron]

use " for dialogues and sometimes it takes time to generate audio for a video so wait around 1 hour
if not then as its mentioned that its in beta and sometimes it cant have audio

[molten rose]

use veo 3 - quality

[dense rune]

I did, I’ll wait an hour and see thanks

A realistic Hello Kitty character, wearing a pink dress and bow, walks through a flower-filled meadow during golden hour. Gentle, magical background music plays throughout — soft piano and chimes. Birds chirp softly in the distance, and a light breeze flows through the flowers. Hello Kitty approaches a smiling little girl (around 5 years old, wearing a pastel sundress). She kneels down, smiles warmly, and clearly speaks in a soft, cheerful voice:
“Hi! I brought you something special!”
Her voice is audible, synced to her mouth, and stands out clearly above the music. All natural sounds — wind, birds, footsteps — blend into a calm, heartwarming atmosphere. Shot in 4K cinematic quality with soft-focus background, warm lighting, and a wholesome, magical tone.

Here is the prompt

[dim gull]

You have a child in your prompt, it will not generate audio.

[dense rune]

Ahhh thanks

[distant sinew]

If I write a slightly more complex prompt, the interaction will be problematic, and I will waste a lot of credits to select a slightly better result.

Then when can be able to generate videos of about 15-20 seconds? 8 seconds is too short...

[oblique copper]

#flow-tips-n-tricks・✏️ Spider-Man in full red and blue suit performs a straight vertical freefall, dropping directly downward from the top of a tall building in an Indian city. He holds a small vlogging camera in one hand, pointing at his own face in selfie style. The shot is entirely vertical — no horizontal motion, no web-swinging — just Spidey falling straight down with gravity. The camera view shakes naturally as he free-falls, with a stable lock on his face and upper body.
The Indian cityscape above him rapidly recedes as he drops — temple domes, billboards, rooftops, and hazy sunset-lit buildings blur upward past the edges of frame. His body is vertical, head-first or chest-first to the camera, suit slightly flapping from wind force. His expressive mask eyes show confidence and thrill. He looks directly into the lens and speaks excitedly in Hindi:
“Welcome to Spideyverse! Ye meri duniya hai — is mein main jo chahoon woh kar sakta hoon!”
The background motion is strictly vertical, fast downward acceleration. The mood is cinematic, bold, thrilling, and fun — like a superhero’s fearless vlog. Add golden hour sunlight, light wind sound, and slight motion blur on background. Duration: 8 seconds.

[quasi timber]

ahri?

that's sick

[cerulean kite]

This prompt generates the same characters pretty much every time. Although this video took me about 10 tries just because Evo 3 kept adding random things...

PROMPT:
A realistic, cinematic 8-second shot, no captions, no music, no subtitles: A Yeti, named Sam, has thick, shaggy white fur, with a slightly darker grayish-blue face, giving him a rugged yet noble appearance. His deep-set, intelligent eyes glow faintly with an icy blue hue, and his brow is heavy but expressive, hinting at both wisdom and strength. dressed in full riot gear, body armor, and shield - sits heavily on a public bus bench with his helmet on the ground like an exhausted football player, catching his breath. Sitting next to Sam on the bench is Juan, a Sasquatch. Juan has thick, shaggy brown fur with a lighter tan-colored face, He has a bloody head from being beat up. He has expressive amber eyes, and a wise, noble demeanor. His towering frame is massive yet dignified. Around them, a full-blown protest riot rages with flashing police lights, tear gas, screaming protesters, riot police clashing with shields, smoke, fire, and debris flying. The camera remains steady and intimate, medium close-up with shallow depth of field—blocking out the background chaos just enough to focus on the surreal stillness of the two beasts. Sam, the Yeti says to the camera: “I don't know why, ice, is trying so hard." he pauses and says to the Sasquatch, "If you order Uber eats, they'll come right to your door!" This causes the Sasquatch to laugh. The moment is absurd, quiet, and strangely grounded amidst the mayhem.

The only thing I change is the words really

[kind plinth]

How are people making consistent scenes? when the extend feature doesn't allow for audio to continue with Veo 3?

[dim gull]

Good prompting, direction, and knowing Veo's strengths and weaknesses. A lot of people are asking the model to do too much in a single clip/scene. Solid prompting is the biggest factor.

[slow moon]

Hi, so ive been trying to use flow to make scenes of the same person, but when i do this i keep getting the different people event though i mention explicitly to keep the same subject and have them in a different location. Do you have any tips on making the same person?

Damn, will this deffo work tho, dont mind signing up for the more expensive version, but when i try to upload a photo of even an already AI generated person as a prompt for a frame i get this error

yeah i had basically screenshot a clear frame of the AI generated person and tried using that as a prompt but still no luck, might just need to get a VPN to sort that out

[wanton imp]

so does image to video take that image into a video or will it take the people in the photo and make it AI?

[slow moon]

Na it doesnt take anything that looks human from the looks of it

agian i wouldnt might paying for the upgrade but just didnt want to do it if im gonna get the same error using image to video

that'll probs work but was going for some realistic movie scenes

[dim gull]

Generating people with any real consistency solely using text to video (the only model that produces dialogue) is basically impossible. You can do okay with really solid prompting (exceptionally detailed subject blocks), but there will always be noticeable inconsistencies because as humans we're trained to see and recognize people, and we know if anything is off. Best practice is to not expect consistency and prompt around that.

[slow moon]

yeah country is deffo going down the toilet hahahaha

[drowsy blaze]

@dim gull @weary pewter @grizzled quarry I think this one came out way better! What do you guys think? I spent about 500 credits testing things out — three didn’t work, but I’m really happy with how it turned out.

[undone ridge]

I need help, im using this prompt and i get no audio, what can i change? I tried using and old prompt from someone here and audio worked, so im thinking its a prompt issue, this is my prompt A wide classroom shot, lit by cold fluorescent lights and framed from a low angle, captures an old Argentine principal standing at the front. She has short, gray, rounded afro-like hair, pale skin, and wears a faded cardigan. Her face is stern, her tone sharp and irritated. With a thick Argentine accent, she declares loudly: "¡No se usarán más los celulares en clase!" Cut to a handheld close-up of a young white-skinned girl in the back. Her brown roots fade into blonde tips. She looks panicked, eyes wide, breath quick. Light from a window catches her face as she pulls out her phone and yells, trembling: "¡ESTOY LLAMANDO A LA POLICÍA!" Her voice is audible and emotional, with a scared, angry tone in a clear Argentine accent.

[dim gull]

Veo wont generate audio that includes children.

[undone ridge]

Ohh so thats the reason, thank you

Can i keep the classroom abient and make the student older? Or if it has a classroom it wont

[bitter stump]

A realistic, cinematic 8-second shot, no captions, no music, no subtitles: A Yeti, named Sam, has thick, shaggy white fur, with a slightly darker grayish-blue face, giving him a rugged yet noble appearance. His deep-set, intelligent eyes glow faintly with an icy blue hue, and his brow is heavy but expressive, hinting at both wisdom and strength. dressed in full riot gear, body armor, and shield - sits heavily on a public bus bench with his helmet on the ground like an exhausted football player, catching his breath. Sitting next to Sam on the bench is Juan, a Sasquatch. Juan has thick, shaggy brown fur with a lighter tan-colored face, He has a bloody head from being beat up. He has expressive amber eyes, and a wise, noble demeanor. His towering frame is massive yet dignified. Around them, a full-blown protest riot rages with flashing police lights, tear gas, screaming protesters, riot police clashing with shields, smoke, fire, and debris flying. The camera remains steady and intimate, medium close-up with shallow depth of field—blocking out the background chaos just enough to focus on the surreal stillness of the two beasts. Sam, the Yeti says to the camera: “I don't know why, ice, is trying so hard." he pauses and says to the Sasquatch, "If you order Uber eats, they'll come right to your door!" This causes the Sasquatch to laugh. The moment is absurd, quiet, and strangely grounded amidst the mayhem.

where am i suppose to prompt

[dim gull]

https://labs.google/flow/about

[bitter stump]

I click 'Create Flow' and it does this

[dim gull]

Well yeah, it's asking you to use a chromium based browser.

[bitter stump]

my bad

lazy me didnt read

[slim lynx]

looking for some help if anyone would be willing

[dim gull]

What are you needing help with?

[slim lynx]

looking to learn to master my prompts in order to get consistent characters in my videos

[dim gull]

Giving an example of your prompt and posting the video it produced is a good way to get feedback. The best way to improve is by looking at what others are finding success with. Otherwise I'm currently working on an app that creates and manages prompts tailored for Veo3.

[slim lynx]

amazing

ok

[agile dirge]

Having trouble keeping the camera from moving
A stunning, high-resolution view of a rotating spiral galaxy in deep space. The camera remains perfectly still, centered on the glowing galactic core. Dusty arms spiral outward, filled with radiant stars, glowing nebulae, and colorful cosmic clouds. Soft interstellar light scatters through space, with slow, majestic motion. No camera movement, no zoom, just serene, continuous galaxy rotation with a sense of cosmic scale and beauty

[dense rune]

@dim gull Let me know when your alchemist gpt is up and running. Please

[velvet inlet]

Anyone use a good gpt for their prompts ?

[limber tiger]

hello

i have google AI pro but it shows "You must be a Google AI subscriber to generate. Feel free to look around in the meantime." whats the solution?

[tall jay]

does anyone have a great prompt to create asmr videos? ive seen some on socialmedia with people cutting fruits in asmr style but it was ai generated and probably with VEO3

[viral tulip]

Make a snowman on the skateboard in South Florida

[naive tartan]

this is nice.

[silk anvil]

Vídeo 360 estilo 'tiny planet' com estradas e rodovias torcidas e interligadas, formando uma malha urbana circular compacta. Ruas urbanas minúsculas e redondas se entrelaçam na superfície do planeta. Um girafa gigante, em escala surreal, caminha majestosamente por este 'pequeno planeta', perfeitamente centrada no centro da moldura

[cosmic summit]

I don't know if this has been discussed before, does anyone use a tool to generate detailed prompts for Veo? Or some kind of workflows.

I am looking to provide a single liner and then get a detailed prompt that could be fed into Veo.

[patent flax]

https://youtu.be/0r2SlMm2lr4?feature=shared

[unborn cove]

This was a nightmare and burned quite a few credits on Veo being Veo.. but its about as consistent as Ive been able to get longer than a 2-3 scene clip... https://www.tiktok.com/@itschupaguey/video/7515563355793362219

[dim kiln]

un 16 - 1800
more_vert

PRO
User profile image
June 16, 2025
📷 Scene Description: A wide-angle shot sweeps across the bustling Piazza district of Addis Ababa, where thousands of peaceful demonstrators chant in unison, fists raised in rhythmic solidarity. Banners flutter above the crowd: saying “ጆን ዳንኤል ከእስር ይፈታልን!” which means free Jon Daniel, free Jon Daniel Cut to a roving handheld close-up: a young Ethiopian TV journalist in a blue PRESS vest approaches a random man amid the crowd. She quickly raises a mic. 🎤 Voiceover (Amharic) > “ለምን ጆን ዳንኤል ከእስር እንዲፈታ ፈለጋችሁ?” (Why do you want John Daniel to be released from prison?) 🎧 SFX Distant chants echoing: “ጆን ዳንኤል ይፈታልን!” Microphone crackle Light wind Urban crowd murmur 🧔 Response - Interviewee (confused but honest) He squints under the sunlight and chuckles nervously: > “እኔ አላቅም... ሰው ሲጮህ ሰምቼ ነው የተቀላቀልኩት።” (I don’t know… I just heard people shouting and joined in.)

#flow-tips-n-tricks・✏️ un 16 - 1800
more_vert

PRO
User profile image
June 16, 2025
📷 Scene Description: A wide-angle shot sweeps across the bustling Piazza district of Addis Ababa, where thousands of peaceful demonstrators chant in unison, fists raised in rhythmic solidarity. Banners flutter above the crowd: saying “ጆን ዳንኤል ከእስር ይፈታልን!” which means free Jon Daniel, free Jon Daniel Cut to a roving handheld close-up: a young Ethiopian TV journalist in a blue PRESS vest approaches a random man amid the crowd. She quickly raises a mic. 🎤 Voiceover (Amharic) > “ለምን ጆን ዳንኤል ከእስር እንዲፈታ ፈለጋችሁ?” (Why do you want John Daniel to be released from prison?) 🎧 SFX Distant chants echoing: “ጆን ዳንኤል ይፈታልን!” Microphone crackle Light wind Urban crowd murmur 🧔 Response - Interviewee (confused but honest) He squints under the sunlight and chuckles nervously: > “እኔ አላቅም... ሰው ሲጮህ ሰምቼ ነው የተቀላቀልኩት።” (I don’t know… I just heard people shouting and joined in.)

[steady hound]

Create a cinematic 16:9 video of a old filipina women around 40-45 years old doing a POV-style vlog. The video should appear as if it's recorded using a handheld phone camera, with a slightly shaky movement to give it a realistic selfie vlog.
The old filipina women saying something infront of camera:
"HAPPY BIRTHDAY ALBHENKHAIR SAUPI MORE BEMBANG TO COME LONG LIVE ALPHA KOKAK!"
The old woman laugh so hard because of happiness The background should be crowded and lively streets. Most streets are narrow and filled with tricycles, jeepneys, street vendors, and people walking. Houses and small stores are built close to each other. Many areas are busy with street markets, food stalls, and small businesses. Children often play outside, and neighbors usually know each other well. Despite being known as a poor area, Tondo has a strong community spirit.

[verbal wyvern]

thanks @lean viper I'm okay with giving up voice for now because I can just do an instructional voice over, but I want to keep the setting and teacher constant across several actions so we can see a child get upset, a teacher join him calmly, the teacher pointing and talking and helping the child reorganize the materials etc. advice on how to lock elements of a clip?

[floral vapor]

this is great. can you share the prompt? how'd you get the girls outfits?

[rocky needle]

Hi Hussam,

We cannot see the prompts ? Can you share them ?

Thank you

[limpid lodge]

Hey! I’m loving yall videos ! I have a question, what prompts are you using to look “realistic” in added “hyper-realistic, 4K, 8k, shot with Oscar winning camera, etc and it’s still coming out semi cartoonish for me .

[hasty grail]

what prompt did you use for this?

[drowsy blaze]

Hey Alan!
Thanks! Here's the prompt I used — I did mention a bit about the clothing style, something like sexy and modern vibes, but didn’t go into too much detail. and… well, VO3 decided to turn up the heat 😄🔥 I guess it all depends on how sexy the algorithm was feeling that day!:

"cinematic scene in Buenos Aires at sunset, with the iconic Obelisco bathed in warm golden light, two confident blonde women in their early 20s, both with curvy, well-defined figures, standing together on a bustling sidewalk, exuding poise and charisma. They should wear stylish, slightly provocative outfits that align with the trendy, playful, yet tasteful Argentine nightlife aesthetic. The first woman Say: “Did you see that girl? She went totally wild over DJDTime!” with genuine curiosity. The second woman should respond with a sly smirk, glancing at the camera, and say in a playful, knowing tone, “Yeah, and you know what? You can download the track for free — it’s right under this video.” Both should then share confident, in-the-know smiles and nods, reinforcing the scene’s cool, pop-culture vibe. Maintain a relaxed, fun, and distinctly cinematic atmosphere, No subtitles, on-screen captions, NO music. "

[urban dust]

generate the video of a beautiful black girl walking alongside of a black boy and talking about their love

[zinc spear]

I used this prompt but changed the details. I forgot to command to not include subtitle. haha

[hollow torrent]

Hi everoyne new here, learning flow. Anyone know how to make the scene extensions continue with the same sound and setting? is there a simple text prompt one could use or do you just create an entire new 8 second scene and copy/paste scene details forward?

[midnight rampart]

Where?

[fervent pagoda]

https://www.youtube.com/watch?v=jfDj-znPW1Y

The prompt is written in the description section of this URL.

[limpid lodge]

Let me grab it when I get back to the computer

[coral thicket]

Where are the prompts?

[cunning basalt]

Hey, do you guys find it better to do text to video or to start off with an image as the first frame?

[heavy totem]

Anyone having luck with character consistency?

I’m using jump to and last frame

One character loosely resembles , the other one no luck at all

Can we just get the same toolset Aronofskys team got 🤣

[green tide]

Hello everyone. Is there a guide somewhere on how a prompt for Veo3 should look?

I want to make videos in which several people talk and perform various actions. Unfortunately, I'm not having much success generating the prompt via Chat GPT. The results are not very satisfactory. How should a prompt be structured?

Many thanks in advance.

[tall swallow]

Using Veo 3.0 Quality — how can I suppress chewing/lip motion defaults in a hyperrealistic alpaca character? I’ve layered no-chewing prompts, suppressed blinking, tried alternative phrasing, even removed ‘alpaca’ as a keyword — but the mouth keeps chewing. Anyone cracked this?

[visual yarrow]

Does anyone have tips on how to make a character disappear/vanish in teleportation scene. I’ve tried several prompts even using Gemini but it doesn’t work. Is it a limitation on VEO itself?

[nocturne ibex]

https://www.youtube.com/watch?v=ZkXG3ZrXlbc

anyone have any tips to recreate the car compacting scenes in this video

tried a bunch of stuff but not getting very close

[lethal eagle]

I'm having very similiar issues myself. Trying to get a first-person-camera view of a car crash from the inside, and it just ends up looking like a GTA animation so far

Prompt: The camera is set to the driver's point of view and we drive extremely fast in a 2021 Dodge Charger, going at speeds over 127 mph as the car races down the road with very poor visibility because of the darkness and the speeds which are causing motion blurring of the commercial buildings and parking lots the car is passing extremely quickly. Out of nowhere, there's a left turn the driver can't possibly maneuver, and the car crashes VIOLENTLY into a tree head on, causing catastrophic damage as the car explodes.

I'm just gonna put it out there, I'm pretty sure the driver of this car didn't even care that they hit a random ghost object and exploded because they started speeding back up again. 🤣

[hasty grail]

your average bmw driver

[dense rune]

What’s a good prompt for selfie video? Mine always come up as the holding a phone

[valid rune]

Camera Style: Selfie POV, handheld and slightly shaky filmed on a GoPro with a slightly wide angle, duct-taped to a bent selfie stick. The camera moves a bit as the character talks giving it a vlogger feel. The right hand of the character holds the selfie stick and doesn't let it go. Character: A scruffy urban raccoon with matted, dingy fur, a faded dirty denim vest with a couple of tears and a missing button, he has one ear nicked with a scar like he’s seen some things. Grease smudges his paws and snout. He’s holding a greasy slice of pepperoni pizza in his left paw, halfway eaten. and he holds the selfie stick in the other. Environment: Dimly lit city alleyway at night flickering overhead light, puddles reflecting neon signs, graffiti-tagged brick walls. Next to him: an overflowing dumpster he clearly just emerged from, its lid still creaking shut. A few curious rats scatter as he strolls past. Action: He is standing in front of the trash can, he takes a bite of the pizza chews a couple times and he says: “You say trash can…I say mystery buffet”, as he points to the camera with the pizza. Glances back at the dumpster with mock reverence, then forward again like he owns the block. Mood / Expression: Cocky and streetwise—he’s clearly proud of his scavenger skills, playing it up for a vlog-style audience. Bit of a mischievous glint in his eye. Voice: the racoon has a Raspy, East Coast-style accent with a smirk in the tone. Audio: Occasional background sounds: distant siren, metallic clatter from inside the dumpster. Duration: ~8 seconds, Do not include any captions, subtitles, or on-screen text

[ocean patio]

use handheld selfie stick, or handheld selfie view

[dense rune]

Thanks I’ll give it a shot

[torn stone]

anyone knows why my videos not getting sounds. I am using the veo 3 highest quality 100 creds per video.

[dense rune]

What’s the prompt?

[torn stone]

This one I just chat gpt this

A realistic baby, about 1 year old, is sitting in a mini podcast studio modeled after a real gym podcast — think modern, black-accented setup with subtle lighting, a dynamic mic in front of him, headphones too big for his head, and a shaker bottle labeled beside him. The baby is dressed in toddler-sized gym gear — sleeveless hoodie, baby joggers, and tiny sneakers. Despite being a baby, his tone and delivery are edited with a deep, mature-sounding voiceover that makes it funny and surreal. 🎙️ Voiceover (baby’s voice, edited): “Yeah, people always ask me… how does a baby hit PRs before he hits kindergarten?” He smirks and takes a sip from his sippy cup. “Easy. Baby milk shake. That’s my secret sauce.”

[livid condor]

Does it make sense to have Chatgpt create prompts for Flow in a JSON format?

[dim gull]

Veo 3 doesn't support audio for generations that include children. Also use Veo Fast for 20 credits. Nail the prompt and then use Quality if you choose.

[torn stone]

Thank you for this. But I made my first one with a gym baby and it has voice. Just my next prompts that doesn't have.

[torn stone]

Thank you. Noted on this.

[acoustic rune]

trying to get a first person view of someone running and its only giving me third person or a stationary camera panning to them, can anyone help?

[hasty grail]

whats ur prompt rn?

[open spire]

any tips on how to stop making everything giberish

[prisma galleon]

video

[odd bough]

Same sound is not possible right now for scene extensions, just video, which will default to Veo 2 Quality. Until they bring Veo 3 into the same world where you can seed with ingredients/frames, your only option for video with sound is going to be a new, standalone 8 second video.

[hollow torrent]

This was very helpful thank you

[inner wren]

How do I get more cuts for these tried to switch up the prompt and it just glitches

[serene hawk]

Hi, how to put audio in Brazilian Portuguese in flow?

[spring saffron]

I just use : and name says in Portuguese Brazilian "Texto aqui em inglês dentro os parenteses" and he translates

For example in this one i used : An man in a construction site being interviewed by a journalist the journalist says in Portuguese Brazilian "What you think about arcanjo situation" and the man responds in Portuguese Brazilian "Yeah he is without gmail, in the construction site that's all we talk about"

[serene hawk]

Well, I've tried exactly this but generatede no audio

[spring saffron]

Are you using Veo 3 high quality ?

[serene hawk]

yeap

I used this sentence

“Oi! Eu sou o Koalino… seu amiguinho da hora de dormir!” (Translation: “Hi! I’m Koalino… your little bedtime buddy!”) •

[spring saffron]

Try using like this...

An chinese man speaking in a chinese stereotyped ascent in Portuguese Brazilian "Hya! I’m Koalino… your little bedtime buddy!"

[serene hawk]

it works!!!! thank u

[spring saffron]

Did it worked ?

[spring saffron]

Send here too, i wanna see lmao

[spring saffron]

Valeu kakakak

[keen mason]

Does anyone have suggestions for getting the AI to talk faster?

[strange spindle]

Anyone know how to extend the prompt window?

[knotty peak]

so if im trying to get my characters to talk in spanish, how exactly do i do that ?

[compact roost]

Anyone have an idea how to make prompts for the vlog style videos?

[lyric flint]

hi ¯_(ツ)_/¯

[earnest hornet]

Anyone having good results with keeping vocal consistancy?

[fierce prawn]

I do

How's everyone doing

[hardy quiver]

Allll goood

Could you send a prompt as an example for the vlog style videos???

[tropic vigil]

First 8 seconds created with veo 3 fast with audio. Then when I try to exter "jump to" or "extend" i get the dismiss notice and the only option is to create the remaining clips with veo 2, which has no audio and less quality. How do I overcome this so my entire video can be created with veo3 fast?

[patent flax]

https://youtu.be/9xOHj0L9XpY?feature=shared

[strange spindle]

I'm having trouble getting veo 3 fast and quality to create a scene with the proper camera angle. I am trying to get the camera angle to face the subject through the windshield as a close-up. No matter how I change the prompt, I keep getting the scene through the passenger or driver's door with the door open. Any thoughts?

[celest fog]

good resturant

[ancient lark]

this one came out good

[cerulean hound]

Thanks

[hollow crow]

Lol

[tiny patrol]

rip

[outer ether]

Can you help with the language and the pronounciation? We tried VEO3 with German words, but it sounds very Robo like. In flow it did not give a speach with the video. Do you have tipps for the prompt? (In VEO3 we prompt "Spoken Dialogue (German)".

[vale bloom]

Does anyone have any ideas why ingredients to video not working with Veo 3 Audio (BETA)?

[blazing lagoon]

Hello. I'm creating a video using Flow, but if the result turns out strange, is it possible to get a refund for the 100 credits used?

[outer ether]

Can you help with the language and the

[rare oxide]

[Location: The Heart of the Jungle – Sacred Rock Throne – Sunrise]

Birds cry in the golden dawn. Sunlight pierces through the dense canopy of the ancient forest. All animals – elephants, deer, tigers, eagles, monkeys – gather in silent reverence around a towering stone throne covered in vines and claw marks of legends past.

Atop the rock stands SIMHA, the majestic Lion King, his golden mane flowing in the breeze, eyes blazing with wisdom and fire.

The old Owl, jungle seer, speaks in a solemn tone:

OWL (Seer):
"From today, you are not just a hunter, not just a protector. You are the Dharma of the jungle. You are the Roar of Justice."

Simha steps forward, looking at the animals below.

He raises his paw toward the sky and his voice echoes like thunder:

---

SIMHA (The Lion King):
"Main Simha, vanraj, aaj sabke saamne vachan deta hoon..."
"Jab tak is jungle mein meri dahad goonjti hai, jab tak mere panje zameen par hain —"
"Na anyaay hoga, na bhay hoga!"

"Har nadi, har ped, har jeev mere parivaar ka hissa hai!"
"Jo kamzor hai, uski raksha meri zimmedari hai. Jo balwaan hai, usse dharm mein bandh ke rakhna mera farz hai!"

"Yeh mera vachan hai – yeh van ekta ka, maryada ka, aur sammaan ka prateek banega!"

"Agar kabhi is jungle par andhkaar chha gaya... toh main apna khoon baha dunga, par is van ki roshni kabhi bujhne nahi doonga!"

---

[A roar erupts – powerful, primal, divine. The entire jungle falls silent in awe. Even the winds bow to the promise of their king.]

The animals bow one by one – not out of fear, but out of love and trust.

FADE OUT with a thunderclap and the sound of the lion’s roar echoing through the mountains.

#flow-tips-n-tricks・✏️ [Location: The Heart of the Jungle – Sacred Rock Throne – Sunrise]

Birds cry in the golden dawn. Sunlight pierces through the dense canopy of the ancient forest. All animals – elephants, deer, tigers, eagles, monkeys – gather in silent reverence around a towering stone throne covered in vines and claw marks of legends past.

Atop the rock stands SIMHA, the majestic Lion King, his golden mane flowing in the breeze, eyes blazing with wisdom and fire.

The old Owl, jungle seer, speaks in a solemn tone:

OWL (Seer):
"From today, you are not just a hunter, not just a protector. You are the Dharma of the jungle. You are the Roar of Justice."

Simha steps forward, looking at the animals below.

He raises his paw toward the sky and his voice echoes like thunder:

---

SIMHA (The Lion King):
"Main Simha, vanraj, aaj sabke saamne vachan deta hoon..."
"Jab tak is jungle mein meri dahad goonjti hai, jab tak mere panje zameen par hain —"
"Na anyaay hoga, na bhay hoga!"

"Har nadi, har ped, har jeev mere parivaar ka hissa hai!"
"Jo kamzor hai, uski raksha meri zimmedari hai. Jo balwaan hai, usse dharm mein bandh ke rakhna mera farz hai!"

"Yeh mera vachan hai – yeh van ekta ka, maryada ka, aur sammaan ka prateek banega!"

"Agar kabhi is jungle par andhkaar chha gaya... toh main apna khoon baha dunga, par is van ki roshni kabhi bujhne nahi doonga!"

---

[A roar erupts – powerful, primal, divine. The entire jungle falls silent in awe. Even the winds bow to the promise of their king.]

The animals bow one by one – not out of fear, but out of love and trust.

FADE OUT with a thunderclap and the sound of the lion’s roar echoing through the mountains.

[tough sleet]

#flow-tips-n-tricks・✏️ fait une video avec veo3, un homme qui interview une grand mere dans la rue il dit, bonjour madame comment aller vous" elle repond "bien ou bien la mifa, sa traine dans quartier vasi la rue tu connais" il repond" ha dacors vous vivez dans le quartier?" elle repond" ba biensur que j'vit dans le tiercar cousin qu'est tu croit on vient de la street la vrai, pourquoi t'es toi tes venue pecho de la bonne frero? "

[green tide]

can someone help me with my prompt?

Unfortunately, I always get poor results in veo3. The prompt is generated via chat gpt.

It would be great if someone could help me.

[dire brook]

I have someone reading a script, but on each generation, when the script ends, they then look like they're about to say something else before the end of the clip. Has anyone got a prompt to have the subject read the script, then just stand there? Weird one, I know.

[patent flax]

https://youtu.be/wyDva8dBP2Y?feature=shared

[wanton dawn]

when will the support for tamil speaking prompts be supported for Google Veo3 in Flow?

[amber summit]

Hello, I have two questions

1. Sometimes videos are produced with black bars on the top and bottom. How do I prevent this? I want it to be in 16:9 format. Is there a definitive solution?

2. Sometimes 2 or 3 scenes appear in videos. I don't want this either. Is there anything I need to do to make it a single scene?

Thanks.

[tiny saffron]

Misty prehistoric forest at dawn. Wide shot slowly zooming toward a primitive man in front of a cave, striking two stones. Close-up of cracked hands, a spark lands on dry leaves — fire bursts with a crackle. Warm light hits his mud-covered face. He rises, lifts arms, cries out in joy. Wide shot shows tribe watching silently, uninterested. Ambient wind, forest sounds, and fire crackling. Cinematic 8K, natural dawn lighting, slow dolly push-in.

[muted quartz]

Im trying to get these birds to play and fly around but it just keeps generating this.

[blazing escarp]

I'm making a jungle video with Mowgli, and in the prompt I'm also providing a consistent character description so that the face comes out 70% the same. There's no violation in my video either, so why is it not generating the video smoothly? Can someone help me?

I'm making a jungle video with Mowgli, and in the prompt I'm also providing a consistent character description so that the face comes out 70% the same. There's no violation in my video either, so why is it not generating the video smoothly? Can someone help me? every time he say This generation might violate our policies. Please try a different prompt or send feedback

[tawny idol]

ask chatgpt

[solar igloo]

Any tips on how to create consistent looks of characters for multiple 8-second clips?

[viral skiff]

Hello, good morning. Can someone help me generate a prompt for an image? I can send it and it's not a problem because it doesn't come out.

[rocky needle]

Hi @hybrid lily ! Nice one. You're supposed to attached the prompts there

[charred nimbus]

Hey everyone, any tips on prompts for these realistic videos?

[wheat flume]

use same detailed description of the character in every prompt

[wintry prairie]

#flow-tips-n-tricks・✏️ A pair of black gloves slowly squishing a bright red apple on a white table. The apple breaks and juice splashes realistically. High detail, cinematic lighting, slow motion, ASMR style, clean white background.

[ebon kelp]

A pair of black gloves slowly squishing a bright red apple on a white table. The apple breaks and juice splashes realistically. High detail, cinematic lighting, slow motion, ASMR style, clean white background.

[wanton lichen]

https://www.youtube.com/shorts/r4iVSlacQMs

[autumn roost]

@hallow spade que tal hermano pudiste hacer videos en flow en spanol!?

[next socket]

Hi those with the another one is gone videos.. anyone got good prompts for them

[lean star]

https://youtu.be/jeHQ2yUhTmY?si=kuUbjgzNmOhhVsg7

[cinder crag]

Does anyone know how to specify timing in FLOW. For example, how would I specify in prompt that a character is to enter the scene at the 3-second mark? Also, all shots seem to be 8 seconds long: is there a way to specify the length of a specific shot?

[fickle gate]

Ultra-realistic video of Bigfoot in the woods. It's broad daylight with dense green forest in the background. Bigfoot is walking through the trees like a vlogger, occasionally looking into the camera with a goofy expression. He says: "Hi Guys. I heard Dragon loves it when you call him, Ah Long…Next time you see him say, Ah Long, Ah Long, Ah Long Ah Long Ah Long, and watch him turn red".
'. then laughs and gives a thumbs up to the camera. Include natural ambient forest sounds, shaky handheld motion, and a humorous, lighthearted tone.

[half orchid]

Does anyone know how to make the video into vertical (9:16 format)? I have already written the "generate a vertical video in tiktok video format", but the result still come out to be a horizontal one?

[molten parcel]

I edit mine in divinci by changing to vertical resoultion you will loose some of the video but you just need to adjust the centering etc

[rotund lantern]

🗣️ Stormtrooper (to camera):
„კახეთში ვარ, ღვინო ბომბაა… რუსულიც მირჩევნია, მგონი.“
(Translation: „I'm in Kakheti, the wine is amazing... I think I prefer Russian wine though“)

[Sudden silence. Locals freeze mid-toast. One man slowly stands up from the supra table.]

🗣️ Local (seriously):
რას თქვი, შე თეთრო ჩაფხუტიანო?“
(Translation: „What did you just say, you white-helmeted?“)

[Stormtrooper looks around nervously. Locals reach for churchkhela, shovels, and qvevri lids.]

🗣️ Stormtrooper (panicking, running):
„ჩემი აზრით შემდეგში რუსები არ უნდა ვახსენო!“
(Translation: „I think I shouldn't remind them about russia next time!“)

why doesnt this work?

[rotund lantern]

I am also getting an error that only english is supported

[cloud vessel]

Is it possible to generate a video with a 9:16 aspect ratio in Flow?

[onyx kite]

maybe you should add that he is saying that phrase like stormtrooper says

and then “”

what he is saying

[rotund lantern]

it used to work

now Georgian is not working at all

[onyx kite]

i mean i am confused by ur prompt cuz i dont know what ur putting exactly

[rotund lantern]

it does everything in english but it doesnt do it in georgian anymore for some reason

[inner wren]

I got it to do 2 cuts now which is not bad need to figure out a way to make it do possible 3-4 cuts

[edgy rampart]

How do I create

Here is my prompt

⸻

🎥 [SHOT 1 – HOOK | TEXT ON SCREEN]
🎙 Voiceover (VEO 3 tone):
“I used to dread taking supplements… until I found these.”

📷 B-roll: Close-up of the Re Lierre ACV Gummies bottle spinning slowly, gummies spilling out like candy.

🖋 Text overlay:
“Apple Cider Vinegar just got delicious 🍏✨”

⸻

🎥 [SHOT 2 – BENEFITS | VOICEOVER + TEXT]
🎙 “These gummies don’t just taste like green apple candy — they’re packed with B12, multivitamins, and grapefruit extract.”

📷 Clip:
You holding one gummy between your fingers and popping it in your mouth.

🖋 Text overlay:
✔️ Supports digestion
✔️ B12 energy boost
✔️ No harsh vinegar taste

⸻

🎥 [SHOT 3 – TESTIMONIAL STYLE | FACE TO CAMERA or VOICEOVER]
🎙 “I’ve been taking 2 a day for a week and already feel less bloated and more energized — it’s part of my morning routine now.”

📷 Clip:
Morning routine aesthetic: brushing teeth, taking 2 gummies, sipping water, smiling at mirror.

🖋 Text overlay:
“Morning ritual ✔️ Wellness win 💚”

⸻

🎥 [SHOT 4 – CTA | PRODUCT SHOWCASE]
🎙 “They’re just $9.99, and there’s a $3 off coupon right now. Plus — free samples? Say less.”

📷 Clip:
Pan across the product + pop-up of TikTok shop “Buy Now” button.

🖋 Text overlay:
“Grab yours → Tap Buy Now 🛒”
🔥 Limited offer! 🔥

[tall grove]

Hey smal tipp your prompt has to be like 80% text in english and the rest can be a diff language. Tell that chat gpt/gemini and they fix it, if you still get the error just tell chargpt that u need more english words

[rough falcon]

X2

[topaz comet]

@severe aspen @alpine sorrel

[fluid lynx]

A monkey riding a bike through an Indian village, cinematic lighting, 4K film quality, vlog style camera movement

[coarse matrix]

Any updates on audio, seems no to be working

[split veldt]

A white ex marine muscle man wearing black shades walking in a city under the rain, tracking camera, the man speaks angrily: "This is where it ends" camera zoom to his eyes

[compact raptor]

can you share your prompt?

[tribal widget]

took me a bit to make. support me on tik tok and insta!

[midnight rampart]

prompt ? 🙏

[royal quartz]

Goku beats superman and superman knows it

[hybrid lily]

Cuidado em! Quem fuma Maconha. o chupa Cabra vai pegar kkk

[fierce cape]

I’m trying to get it to generate a horde of people cartwheeling at speed, toward a tower with a belt of black smoke circling ‘round it.

I can never get it to get the cartwheeling right, they always move in a completely unhinged way, or are doing front flips

[jaunty goblet]

Tried to create a swat action prompt

[fierce cape]

Are there any pro prompt engineers that can help me with this? 😄

I think it’s might be that there’s no footage of anyone doing this to base it on

[dim gull]

That's very likely the case, it has trouble accurately depicting quick movement especially with many subjects. As you can see with the clip above and the spawning and dissapearance of people in action sequences.

[orchid jay]

been trying negative prompts but it keeps on putting subtitles.... in gibberish

[tribal widget]

i used chatgpt to create a prompt that describes the setting first before putting in any characters. i made it as descriptive as possible so it can remain consistent. afterwards, i put in the characters, again using chatgpt to describe their posture/expressions

[compact raptor]

Prompt please

[odd atlas]

Hi! How did you get the video to be longer than 8 seconds? I’m trying to learn how to make a 30-40 seconds video.

[dim gull]

You'll need to stitch clips together, or you can use flows extend features but you won't be able to generate audio.

[muted quartz]

How do you prompt for camera movement?

[worthy ether]

Has anyone figured out how to get Veo 3 consistently pick correct text dialog for off-screen camera holding interviewer? It makes the on screen character to say it instead. I tried everything possible already. Here is the example below where “Hey John! Welcome! Wait till you see what I have prepared!” should have been said behind the camera by Tamara, but instead John says it. Sometime it still shows Tamara's side of the face as well, sometimes "Hey John. Welcome" is said by Tamara and the rest by John.

[CHARACTERS]

• John: young man in early 30s.

• Tamara: OFF-SCREEN CAMERA HOLDER. NEVER RENDER HER FACE, HEAD, OR BODY. Her perspective is the camera. ONLY HER HAND may appear. She speaks via voiceover (V.O.) ONLY.

[CAMERA POV]: FIRST-PERSON POV FROM TAMARA — audience is Tamara. DO NOT show Tamara's face/head/body at any time.

[SHOT 1]

📷 VISUAL: Wide shot. John walks in from the left.

👋 ACTION: Tamara’s hand (ONLY) enters from bottom of frame and waves hello to John.

🎙️ AUDIO:

Tamara (V.O.)

“Hey John! Welcome! Wait till you see what I have prepared!”

[scarlet kernel]

How do i use people to upload images

[modern totem]

https://www.youtube.com/shorts/04BA9F4ufog

I was surprised by music generation ability of veo 3. Not perfect still great.

[frozen coyote]

@rain coral can you share a promt?

[inner wren]

I think I’ve perfected it finally

[jaunty saddle]

Awesome 😎

[modern totem]

any suggestions for using consistent narrator voice across clips

https://youtube.com/shorts/t8od0Hp5F-E?feature=share

Had to use audio from eleven labs as different clips had different narrators

[stiff root]

can someone review prompt. We keep getting where bigfoot turned in a real lady. i’m also having issues where our business name is pronounced wrong. it’s Geaux Styles but sounds like GO.

Create an 8-second vertical cinematic video with professional lighting and smooth camera movement. The tone is humorous but polished and fashion-forward, suitable for a boutique ad.

Scene: A tall, well-groomed lady Bigfoot — confident but slightly shy — walks into a trendy, upscale boutique named Geaux Styles. She pauses at the entrance, scanning the modern interior filled with stylish clothing and accessories. Two fashionable shoppers glance over, surprised but clearly impressed by her presence.

The boutique has warm, elegant lighting and soft bokeh in the background. Use subtle cinematic lens flare or lighting transition as she enters.

Voiceover (female, confident, playful):
“She came in hiding… but Geaux Styles knew how to make her shine.”

[hearty perch]

Sorry if this has been asked already! Whats the best way to create 4K-8k prompts?

[fickle orchid]

using the word female or male almost always results in humanoid outcomes. instead of using it as a physical attribute use " female voice actor"

[normal abyss]

hey @rain coral your making some INSANE videos... I was wondering where I can find more AMAZING VEO makers like yourself? I work for icon.com and we trying to find some elite story tellers... would appreicate any advice you have

[jaunty saddle]

A glamorous hollywoods 1960s style, long brown hair, olive-tone skin woman sitting on a velvet sofa in a moody lounge, soft golden light from a vintage lamp lighting one side of her face, holding a glass of whiskey with jazz playing softly in the background, blurred warm bokeh lights, elegant retro evening vibe, semi-realistic digital illustration

Negative prompt
blurry face, extra arms, distorted proportions, watermark, cropped head, low detail, poor lighting, text, logo, cartoonish, modern fashion, weird anatomy, eyes closed

[rancid harness]

prompt ?

[autumn roost]

A question, what's the best way to make a video? inside the flow website or gemini video!? and Can I use my face? anyone know how to use my face to make a videos?!

[jaunty saddle]

There are websites and apps to train your model.

[autumn roost]

Yes but, How Can I use my face inside the VEO!

[jaunty saddle]

You need an accurate prompt. Use Flow.

You can convert photo to video using Veo3.

[autumn roost]

but how I use the image, but only generate without audio!

[jaunty saddle]

Click on the filter and choose veo3

[kindred laurel]

has anyone tried recreating videos using the prompts on flow tv?

[full ember]

Would you have prompt for this

[hallow dune]

could someone generate videos with veo3 of asmr fruits that cut out?

[full ember]

Could you please provide the prompt.

[chrome snow]

yo can someone help me

i just got google veo and like im tryna make a video of a dude fishing in alaska just cause like idk alaska is sick yk, and like, i made it make one video where this dude says like where fishing today in alaska in this beutiful lake blah blah blah, and then i asked the ai to keep the video from above and add another scene to it and then i asked for it to show a new angle from the camera like on a trypod and blah blah blah and anyways it just created a whole new video so like can someone help

that prolly made no sense but wtv 😭

[jolly forge]

Has anyone been using ChatGPT’s Custom Veo-3 prompt maker? Is it good?

[chrome snow]

NVM I GOT IT

YOU HAVE TO HIT ADD SCENE

[chrome snow]

pretty sure everyone here uses googles veo 3

[jolly forge]

Thats not what I asked…

[chrome snow]

wdym

oh

i have no clue if its good

there

[jolly forge]

Respect 😂🤝🏽

Ima try it out and see how good it makes prompts

[chrome snow]

is it free?

[jolly forge]

Oh ya it should be I mean I have gpt pro so it was free for me. But try going to ChatGPT and look at its GPT library and search for Veo 3 prompt maker

[chrome snow]

i already have google one for a year so im good

i dont even gotta buy/spend creditrs lol

credits*

[jaunty saddle]

Prompt: Ultra-realistic, cinematic wide-angle view of a massive computer keyboard with a giant 'Start' button. The 'Start' button is semi-transparent and glows faintly blue, like futuristic glass. Inside this oversized button is a full-fledged miniature room, fully detailed and functional — cozy bed, desk with laptop, shelves, and a soft morning light coming through a small window. A working woman in pajamas is just waking up in bed, stretching as sunlight hits her face. The contrast between the techy, digital keyboard surface and the warm, lived-in room inside the button creates a surreal, dreamlike atmosphere.

[chrome snow]

why is it such terrible quality after the first clip? 😭 and like why did the 3rd clip fail lmfao

oops downloaded it wrong

bro what the heck

why is it downloading from that clip

this is so weird

[crimson kernel]

Why not just have it say "Go Styles"? if there's no subtitles it will sound the same. then overlay a graphic to reinforce the name Title. Just a thought, not sure the complexity of what you're going for, so take it with a grain of salt.

[crimson kernel]

Awesome.

[wraith lake]

Any advice on how to get the same background for every character? I used the same description but it generated a different one every time.

https://youtube.com/shorts/4N_AJKhb7Fs?si=UHTjDCGGThhL3JkV

[normal abyss]

SICK

[jaunty saddle]

We are learning 🌚

[empty coral]

creating a hard seltzer brand using ai stacking agents to create petty story vidoes. check out sippettypearl for some entertaining pettiness.

[safe bridge]

video

[gritty rock]

Any good prompts for the glass cutting asmr?

[prime patrol]

In a dreamy morning of autumn, under a gentle rain, the camera reveals a vast blue sea with a few islands in the background, 8k. The scene centers on a man holding a golden heart close to his face, his expression abstract and mysterious. The atmosphere is filled with uncertainty and ambiguity, capturing a sense of surreal beauty in this cinematic moment.

[vapid garden]

can you please share the prompt : ))))))))

one more question these videos have different size is there any way to create exact size of youtube short frame videos directly from VO3 without editing

[dim gull]

No, Veo does not support vertical format generation yet.

[lost robin]

Check my latest VEO work, focusing with character continuity, theatrical dialogue delivery, and realism.

https://youtu.be/jCeY-fcHNXE?si=8xXC_aO5kPHndhUN

Although its in Filipino language

Ive spent around 6k AI credits for this

[sick ibex]

My videos are limited to 8 seconds, how do I remove this?

[full ember]

Prompt

[full ember]

Bruce

[vapid garden]

https://youtube.com/shorts/ORbQmu64tZM

[mint pecan]

A terrifying, pale-skinned monster woman with long, tangled black hair and hollow glowing eyes peers through a rain-streaked window at night. Her stretched, twisted face is pressed against the glass, and she wears a tattered, moldy nightgown. Her sharp claws gently tap the window frame as she whispers with a sickly sweet voice: “Let your mother in, sweetie.” Inside the dimly lit house, a frightened young girl in pajamas stares wide-eyed, backing away in fear as shadows loom around her. The atmosphere is tense, with cold blue lighting, fog outside, and eerie silence broken only by the whisper. Cinematic horror style.

pls someone generate ts for me I just want a video for it😭

[lilac forge]

Message from Gemini "I can't generate that video. Try describing another idea. You can also get tips for how to write prompts and review our video policy guidelines."

[keen mason]

any tips to get prompts to do things at the same time rather than sequentially?

[red slate]

This is so annoying I'm multiple cuts in and I can't get any audio at all on this prompt:
Camera is focused on a short haired brunet boy wearing a green polo, holding a silver desert spoon, sitting at a picnic table with a red gingham tablecloth on it. Behind him is an old beige and brown Winnebago. The subtle sound of lava popping and hissing, distant sound of birds chirping, children laughing, indistinguishable adult conversation. Action: A woman's hand places a bowl of lava on the table in front of the boy. The boy exclaims: "Basaltic! My favorite!" The boy begins to eat the lava. Do not include any captions, subtitles, or on-screen text.
I was getting background audio last night but character wouldn't talk. Now it's just silent. I'm on Text to Video and Veo 3 Quality. I don't get it the same prompt format did all the other scenes just fine.

I've wasted 1200 credits trying to get this to populate

[red slate]

Why can't it just put up an error. 🙄 😂 My credit wallet is crying.

[cedar frigate]

A small child selling souvenirs, including a scenario where a man gets angry, with an Urdu voiceover element:
Story Time: The Little Seller
Imagine you are a very kind and brave little boy or girl! You are sitting in a beautiful market with lots of colorful things all around. You have a small basket filled with sparkly, special souvenirs that you want to sell to people visiting the market. They are pretty little toys and trinkets that make people happy!
You see lots of friendly faces walking by, and you smile and gently say, "Hello! Would you like a special souvenir today?"
Uh Oh! An Angry Man
Now, imagine a big, grumbly man walks up to you. He doesn't look happy at all. He starts to shout at you in a very loud, angry voice in Hindi urdu

[plucky fox]

Hello, how are you? What was the prompt used in this video?

[wide ice]

can somebody help me create a prompt for flow, where a side tracking side following only the feet of a little body running

[fierce meadow]

Sure! Here's a short dramatic English story prompt based on what you described:

---

🎬 Title: "The Neighbor’s Truth"

Ali and Zara are neighbors in a quiet Pakistani town. Ali falls in love — a one-sided love filled with small smiles, short greetings, and occasional rooftop glances.

One day, rumors spread: Ali forced himself on Zara.

Shocked and heartbroken, Ali denies it. Zara stays silent… but why?

Was it a misunderstanding? A lie? Or pressure from society?

Innocence has no voice unless truth speaks.

---

Let me know if you want a version for video, Instagram reel, or TikTok voiceover — I can make it more emotional or realistic.

[vocal fjord]

prompt: Camera Style: Selfie POV, handheld, arm’s length with slight hopping bobble—shot on a frosty GoPro mounted on a stick made from a snow-covered branch, occasionally fogging up from the kangaroo’s breath • Character: A tough-looking kangaroo with a wool scarf wrapped around his neck, snow goggles pushed up onto his forehead, and a thermos clipped to his hip pouch. His fur is slightly frosted at the tips, and he blinks against the cold. • Environment: Snowy alpine hillside, windswept and sparkling in late afternoon light. Behind him: a makeshift camp with a tarp tent staked with ski poles, a campfire pit with faint smoke, and a snowboard jammed into the snow like a flag. • Action: He bounces slightly in place for warmth, panning the camera to show off the camp. Occasionally adjusts the scarf or knocks snow off his tail with a slap. He gestures proudly toward the tent like a home renovation reveal. • Mood / Expression: Playful and proud—like a chilly but upbeat outdoor vlogger showing off their mountain crib. Slightly winded from hiking up the hill. • Voice & Audio: Moroccan darija -accented voice, mid-pitch and cheerful, synced to lip movement. Ambient wind whistling, distant avalanche rumble, occasional crackling from the fire pit. • Duration: ~8 seconds Dialogue Sync: "مرحبا بيكم ف جبل تيرت، حي القرود! أما دوار الحلوف راه غير بلاصتكم فالخلاء، هنا ما مرحبش بكم هههه!"

#flow-general・🌊 prompt: Camera Style: Selfie POV, handheld, arm’s length with slight hopping bobble—shot on a frosty GoPro mounted on a stick made from a snow-covered branch, occasionally fogging up from the kangaroo’s breath • Character: A tough-looking kangaroo with a wool scarf wrapped around his neck, snow goggles pushed up onto his forehead, and a thermos clipped to his hip pouch. His fur is slightly frosted at the tips, and he blinks against the cold. • Environment: Snowy alpine hillside, windswept and sparkling in late afternoon light. Behind him: a makeshift camp with a tarp tent staked with ski poles, a campfire pit with faint smoke, and a snowboard jammed into the snow like a flag. • Action: He bounces slightly in place for warmth, panning the camera to show off the camp. Occasionally adjusts the scarf or knocks snow off his tail with a slap. He gestures proudly toward the tent like a home renovation reveal. • Mood / Expression: Playful and proud—like a chilly but upbeat outdoor vlogger showing off their mountain crib. Slightly winded from hiking up the hill. • Voice & Audio: Moroccan darija -accented voice, mid-pitch and cheerful, synced to lip movement. Ambient wind whistling, distant avalanche rumble, occasional crackling from the fire pit. • Duration: ~8 seconds Dialogue Sync: "مرحبا بيكم ف جبل تيرت، حي القرود! أما دوار الحلوف راه غير بلاصتكم فالخلاء، هنا ما مرحبش بكم هههه!"

[bright bolt]

i said something like "make something crazy and wacky"

[graceful marsh]

can you use a video to reference a character or is it enough that im generating the new video in the same project folder

[red slate]

I've got 8 iterations of the prompt that are variations of this. I've switched to having the dad say it but I'm having trouble replicating the style of lava it gave me in this clip. It looks like he's eating strawberry jam now.

[red slate]

I'm assuming that was inferred to under "Audio might not generate in certain cases." in the FAQ as I'm not able to find anything specific on the Flow site regarding children specifically. It would be helpful if it popped a warning rather than eating your credits though.

And now having the same prompt adjusted for an adult to say it with no children in the clip at all, it won't create any audio

[upper moon]

This image convert to realistic image

[dim gull]

Whats your new prompt? And did you create a new project?

[red slate]

It was a series of prompts. RV to the lava lake, scooping the lava, walk back and deliver line, person delivers line and eats lava. The final prompt was the same, I just changed the character description from the boy to the dad. Took about 4 tries on the same prompt for there to be audio. Finally got it to work thankfully. It's a dorky bit, but I'm only 3 days into using the program and wanted a challenge.

[graceful marsh]

thanks again just trying to make a continous character im picky i see slight variations even if I give it a very detailed prompt this will help

[gloomy spoke]

Show a camera movement from the ocean away from earth 🌎 into the ends of the known observation zone passing through different planets as the camera zooms by

[clear cape]

Muscular man, 30s, Caucasian, shirtless, performing a TRX exercise in a gym. Focused, determined expression. Sharp gaze directed slightly upwards and to the right of the frame. Strong, well-defined physique; visible muscles on arms and chest. He wears dark blue/grey shorts. Black and yellow TRX straps are prominent, used for resistance training. Modern gym setting; neutral light with large windows, bright lighting. Blurred background features numerous exercise equipment. Composition is centered, showcasing the subject's posture and focus. Mid-range shot, natural lighting. Dramatic lighting highlights the muscles and textures. Sharp detail in the subject's face and physique. Health and fitness aesthetic. Realistic portrayal. Athletic style. Gym-focused photography. Sporty atmosphere. Serious and determined mood.
The text of the speech that the model says:
احس جسمي فاصل، التمارين ذبحتني والله.
يساله صديقه هل جربت العسل الذهبي؟
يجيبه: ياولد والله فرق من يوم جربت العسل الذهبي وانا ماشي زي الصاروخ

[jaunty saddle]

Did you succeed in the Arabic language in the video?

[clear cape]

No

[graceful marsh]

i see that i can add in images but not the reference video

[uncut wedge]

mate might sharing the prompt

[graceful marsh]

thanks again youre the boss i'm gonna go do some tutorials its only my second day

images to video i got it to work , im gonna try video to video

[charred garnet]

#flow-tips-n-tricks・✏️ Prompt:
A digital mascot in the shape of a yellow oil drop, with a friendly face and black sneakers, walks toward the camera on a plain background. The character waves gently and says in Georgian: "უჩა, მე ვარ ქონექთის ახალი ციფრული ავატარი" (Ucha, I am Connect’s new digital avatar). The avatar smiles warmly.
Video should be 8 seconds long, 16:9 ratio, no background music.

id:guide

[coarse smelt]

Hey Guys, I accidentally deleted the folder with some of my prompts on flow. Is there a way to recover them?

[chilly tusk]

anyone willing to help me with my prompt? idk why it feels like a videogame lol

prompt:POV video-vlog at night in the Socotra Archipelago. The cameraman is holding the camera from behind with one hand, and a flashlight with the other. he is slowly walking through a forest overly filled of gnarled, dead Dragon Blood Trees. the ground is cracked and twisted roots stretch across a narrow path. there is a lot of fog and mist (silent hill vibes), The scene is dim, scary, off-putting and shadowy, with faint ambient light breaking through the trees. the cameraman whispers quietly: “This place feels… forgotten. Like something’s still watching from the dark.” Real-life like 4K, shallow depth of field, dark fantasy tone, soft color grading with muted greens and browns.

[kindred laurel]

Did you have gemini help with the prompt?

[chilly tusk]

i usually use chatgpt, is gemini better for suggesting veo3 prompts?

[kindred laurel]

Not sure I havent messed with Veo3 yet, but gemini been great for Veo2

[feral cloak]

I want to use this so bad but audio does not generate for me!

[ornate walrus]

This is good to know as my prompt which included "girl" word never generated audio

[wanton dawn]

Currently while using tamil text for dialogues via Google AI Pro, I can generate videos via VEO 3 with tamil dialogue without fail where as using same text in Google FLow fail to generate video with tamil dialogue using Veo3. How come same platform and model but doesnt support language support on one platform for same veo3 model whereas it works on Google AI Pro platform? attached screen shot of failure on Flow platform

[covert frigate]

Generate a video (aspect ratio: 1:1) where a rhino is munching in a busy street of Gulshan, Dhaka and saying this: ভাই, এতো কথা দিয়া কাম কী? দুই দিনের দুনিয়া, আপনে আমি সবাই মেহমান।
The rhino looks sarcastic and funny. It has a manly voice that follows old dhaka people's Bengai accent.

[nocturne reef]

#flow-tips-n-tricks・✏️ Prompt: generate a video the two sport girls

[mental dune]

A happy child is at a birthday party. Behind the child, the character Stitch waves cheerfully. The child turns slightly toward the camera and speaks clearly. The voice should sound sweet and soft, like a toddler learning to speak but articulating correctly. Perfect Brazilian Portuguese lip-sync and crystal-clear audio quality are essential. Natural facial and body movements are present. The child says in Portuguese: “chá eventos a sua festa começa aqui”. Background is colorful and festive, with balloons, confetti, and a joyful atmosphere.

[marble star]

ungle mein tantrik log black magic kar rahe hain
Prompt:

"Deep jungle in Aarey Colony at night, small fire in center, a group of shadowy tantriks doing black magic ritual, chanting, blood offering, dark red glow from fire, symbols drawn on ground, extremely creepy"

A dark, dense jungle at night in Aarey Colony. In the center, a group of hooded tantrik men are sitting around a small ritual fire, chanting in deep, demonic tones. The fire glows red and orange, casting moving shadows on trees. Skulls, bones, and black magic symbols are drawn on the ground with blood. One tantrik offers blood to the fire. The air is thick with smoke, fog, and low-pitched whispers. Bats fly overhead. A female ghost-like figure watches from between the trees. Horror cinematic style, slow camera movement, ultra-realistic, intense atmosphere, creepy jungle sounds.

Scene 1: Establishing Shot

A dark and foggy night deep inside Aarey Colony forest. The moon is barely visible through thick trees. The jungle is silent except for the distant sound of chanting.

prompt:In the middle of a deep, dark jungle in Aarey Colony at midnight, a group of five hooded tantriks are seated in a perfect circle around a flickering ritual fire. The forest is silent except for the low, demonic chanting echoing through the trees. Skulls, bones, and blood-stained symbols are scattered on the forest floor, glowing faintly under the firelight. One tantrik lifts a bowl filled with blood and pours it slowly onto the ritual symbol. The fire glows brighter, casting moving shadows across the twisted trees. Suddenly, a ghostly female figure appears between the trees. She has glowing white eyes, pale cracked skin, and blood dripping from her mouth. The tantriks continue chanting, unaware of her presence. As the ghost moves closer, the fire bursts high and one tantrik begins to scream uncontrollably, his body shaking violently as if possessed. Bats fly overhead. The camera pans upward through the fog as the ghost walks into the circle, and the fire suddenly dies out. Silence. Only the sound of one final scream echoes as the screen fades to black.

[lost robin]

Check out my veo prompts focusing character continuity, theatrical line delivery AI film making

Although its in tagalog language

[mild nova]

who has any example video of cutting glass fruits?

[hollow cedar]

@viscid gull POV-style cinematic sequence of a middle-aged professional baseball player walking up the dugout steps and onto the field of a fully packed professional stadium. He’s in his early 40s with short brown hair, a short scruffy beard, and an athletic build. He wears a red and white baseball uniform with high socks and a navy blue cap.

The camera follows him from behind in a casual, shoulder-level tracking shot – like a sports documentary crew capturing the moment. The sunlight bathes the field in a golden glow. As he reaches the top step and steps onto the field, the camera swings smoothly around in front of him, capturing his face as he looks out at the roaring crowd.

The stands are full of cheering fans. You hear scattered applause, distant crowd chants, the buzz of anticipation, the echo of announcers over the stadium speakers, and the crack of bats during warmups in the background.

He pauses, takes a slow, deep breath, smiles with visible relief, and says in a calm, grounded voice:

“Finally.”

The moment feels emotional, like a long-awaited return. His eyes are slightly glassy but focused.

No subtitles. No text on screen. No music. Just pure ambient stadium sound and voice – 100% documentary realism.

Camera movement is smooth, steady, and natural – like operated by a shoulder rig. 16:9 aspect ratio. No filters. Ultra photorealistic style.

Try this

[viscid gull]

I'll try it... I started with a much more complex prompt like this and kept trying to simplify it thinking it was to complex and maybe that could have been the problem.

[abstract pine]

please tell me how can I increase the duration of my video which I create with the help of veo2 in Gemini pro from8 second to 3 4 minutes

[viscid gull]

That got the audio working. The dialog is a little jacked. But, you got me back on track though. Thanks. I'll mess with it some more to try to get the dialog correct. Thanks.

[abstract pine]

please tell me how can I increase the duration of my video which I create with the help of veo2 in Gemini pro from8 second to 3 4 minutes hey guys can you solve this problem

No.how to increase the duration of video created with veo 2

👍

[silent pumice]

Photograph of a black, open wireless earbuds case on a wooden surface. The earbuds inside are black with a textured surface and a small red light on one.

POV video-vlog at night in the Socotra Archipelago. The cameraman is holding the camera from behind with one hand, and a flashlight with the other. he is slowly walking through a forest overly filled of gnarled, dead Dragon Blood Trees. the ground is cracked and twisted roots stretch across a narrow path. there is a lot of fog and mist (silent hill vibes), The scene is dim, scary, off-putting and shadowy, with faint ambient light breaking through the trees. the cameraman whispers quietly: “This place feels… forgotten. Like something’s still watching from the dark.” Real-life like 4K, shallow depth of field, dark fantasy tone, soft color grading with muted greens and brown.

[hollow cedar]

thats what u want? 🙂

[hollow cedar]

ah here np

[rough breach]

HOW:(

But I can put prompt here and they gonna make video?

For me?

Yes but who make video here

Like sluffy??

He creat?

[hollow cedar]

i don't, just helping with prompts 🙂

[rough breach]

So what else?

So how huhu

[vapid inlet]

My Flow Credits are not Showing i purchased it WHY?

[hollow cedar]

u have to buy Google Gemini Pro Ab00, and test ur prompts in Google Ai App 🙂

[hollow cedar]

Because its not supported in your country

[vapid inlet]

In INDIA?

so can i do that with USA gmail Account

[hollow cedar]

Check the info for Flow, (70Countrys atm supported)

VPN isnt working

it doesnt matter

[stark sphinx]

Can anyone try this prompt? Cinematic, a drone cam footage, in the sky localizing a medieval house the ui/HUD of the cam system says "alvo localizado", the setting is a medieval Village, at a dark night, during a outrage, explosions, people running desperate, fire shots, very intense scene

[hollow cedar]

u need more details (no one will test it 4free for you)

[solid yoke]

Is this veo 3 fast or quality?

very nice work btw!

[daring quail]

I want to turn this image into a video. the promt i am trying is: pan out to a small red-headed man wearing a similar outfit to the genie, except with a black belt, walking into frame and picking up the lamp. The redheaded man then rubs the lamp. a swirl of the purple smoke should begin to come out of the lamp and the genie should begin to emerge through the smoke, with no legs. as the genie appears, the redheaded man falls back and drops the lamp. the genie, now fully emerged, should look down upon the redheaded man.

but it is not coming out very well. the genie always has legs or there is just no coherece to what should be happening. an thoughts on alternative prompts for the desired outcome?

[viscid gull]

Finally got it done... thanks for the help! https://youtu.be/-x8AbXrzMPA?si=QALf-D5Zr4hiqNxJ

[viscid gull]

I tried both... I'm starting to think it was the prompt. Simple is not always better. And, I kept making the prompt more and more simple and it just wouldn't do it. Then... someone here gave me a really long complex prompt and it works.

https://youtu.be/-x8AbXrzMPA?si=QALf-D5Zr4hiqNxJ

[hollow cedar]

This server is for help, i help where i can 🙂

im also trying to do my prompts better and better ever

yda

[dim gull]

Simplicity leaves too much open for interpretation with Veo 3. Having a solid prompt structure really helps tame the beast.

[desert tundra]

how do I get started with VEO2?

[solid yoke]

I tried to use the same prompt and it came out so cartoony for me. I used veo 3 fast

any help would be appreciated 😄

[viscid gull]

I'd talk to @hollow cedar He helped me.

[abstract pine]

please tell me how can I increase the duration of my video which I create with the help of veo2 in Gemini pro from8 second to 3 4 minutes hey guys can you solve this problem

Hi guy s

Please tell me how can I in crease my video duration from 8 second without veo 3

[hollow cedar]

Its just 8 Seconds prompt 🙂

[abstract pine]

Can it be increased

[hollow cedar]

Nope

but u can do your prompts in 3 - 4min, in 8seconds duration clips 🙂 its super simple

[abstract pine]

Can you please clarify@hollow cedar

[hollow cedar]

type it in chatgpt, use ur brain. And u will get it

[abstract pine]

👌

[green tide]

Does anyone know how I can incorporate the accent of another language into the prompt? For example, German with a Dutch accent?

[hollow cedar]

maybe try this "german but with a dutch accent" 😄

[crimson plover]

anyone know how to get real good character consistency? Or is it just not there yet? I'm trying to upload a reference image for my character in Flow.. Gemini says that feature exists but I can't find it anywhere.

[humble scaffold]

Its the ingredients

[odd birch]

Please move on from this topic thank you.

[stark sphinx]

Then they say they defend free speech, but anyway

[crimson plover]

Ingredients where?

[dim gull]

Character consistency is not perfect, especially with human subjects.

Ingredients is a Google Ultra feature in flow, but it will only use Veo 2, so no audio if that's what you're going for. Best hope for consistency is having extremely detailed character prompting with unique features that the model can latch onto each time.

[empty flame]

Is it possible to take a generated video and prompt on top of it to create a second one? I know we can create a scene which uses Veo 2 and I don't want that. I want to continue from an already generated video

[lost robin]

Check my veo flow prompt, focusing with character continuity, story telling, appropriate emotional dialogue

[bleak bane]

hey guys i'm trying to generate videos with voices in french language with a hard "quebec" accent, anyone can help?

i'm always getting a french accent from france

but the first prompt I did worked perfectly

but I can't make it again idk why

[rotund stratus]

Does Flow have a prompt policy posted on a website? My prompt about kids playing in the street keeps getting rejected. I'm not able to find anything about prompt word/character limits or safety considerations.

[fallen tapir]

It will. Kids and political figures is a No No

[rotund stratus]

No kids at all? How can you create videos with families?

[fallen tapir]

🤷‍♂️

[fallen tapir]

A lot of Ai sites are very restrictive when it comes to children. For obvious reasons.

[rotund stratus]

Gemini allows you to create prompts with children, but Flow won't let you use them LOL w+f

[fallen tapir]

Had one Ai site not even generate a video for a mom holding a baby

[rotund stratus]

I guess ComfyUI is the place to go. No restrictions.

[fallen tapir]

The only restrictions there is your Vram

[rotund stratus]

I use RunningHub to bypass Vram.

[fallen tapir]

As long as your renders are innocent, there's no problem with it.

[rotund stratus]

There are no restrictions on RunningHub. Also, Veo 3 does allow children. This Veo 3 on YT (BABIES IN THE WILD Veo 3 Movie Trailer) has kids in it, so there must be something else in my prompt that's creating the issue. Thanks for your comments.

[heavy totem]

I’ve gotten children in there but you need to be very careful with how you prompt it, has to be G rated practically

This one took some creativity though and the model took some liberties

I didn’t end up using it because of the weird gun light

And the character wasn’t meant to have a gun in the first place

If it’s within context you can get some pretty wild violence

This one I had to remove any mention of children for it to run

[mint zodiac]

How can I shot constantly same character for Veo 3

[tame grove]

Hi guys just a quick question, would it be possible to put references to the prompt when generating videos for VEO 3?

[heavy totem]

You can’t get consistent in veo 3 yet , the best you can do is a hyper detailed character prompt and include it in every shot even if it’s a behind the back etc

[dark flame]

Can they say cuss words

[frail gate]

it really depends... sometimes it would stop the generation from what I've seen

[fallen tapir]

Any tips on getting people to speak at the right times?
Failed example.

[dim gull]

What was your prompt?

[fallen tapir]

@dim gull

[eager citrus]

hey, Im from India and want to create these videos as well. How can i create such video ?
https://www.instagram.com/reel/DLO-F-BxYea/?igsh=MTBxZnN2ZzA0M2p2MQ==

pillow_talks__

☠️💀☠️

#models
#ai #insights #viral #indian #podcast #iran #intimates #romance

Instagram

[sage cipher]

go to Google Labs - Flow, use your google account to log in and you can try it writting a prompt.

[quartz timber]

Galing idol

[leaden token]

hello how do i get in dialogue on prompt, thank you!

[jaunty saddle]

Awesome

Sorry.. how are you 😊
Write a dialogue between a teacher and a student discussing artificial intelligence

[maiden marsh]

hey! please keep conversation respectful to all users please :)

[dry hornet]

Audio problem:

I have a problem with my veo3 quality, where my video doesn't produce any sound. An error appears saying that they were unable to generate the audio. I tried to fix it in several ways and it didn't work. Can anyone help me?

[jaunty saddle]

Use gimini 2.5 pro
From flow Make sure you choose veo3

[light topaz]

yeah having this issue for the past hour too along with a few others

[orchid jay]

veo3 on flow scenebuilder?

[noble lintel]

We need image to video also in veo 3 fast 😁

[jaunty saddle]

Available

[noble lintel]

It literally says text to video next to fast, unless it got changed recently

[jaunty saddle]

Yeah I've never tried that before veo2 frame to video

[viral ice]

https://www.youtube.com/watch?v=S4ZSoTPuWf8

[prisma schooner]

how people making such good bigfoot vlogs lol

[tulip imp]

hey

[stark dagger]

I need a prompt for those vlogs too

[cerulean geyser]

Results are looking great, but does anyone know a prompt that has the subjects eyes open?

[dim gull]

What prompt are you using and what results have you gotten?

[hallow topaz]

hey guys, any good articles out there on how to use prompts more effectively?

[dim gull]

It's a lot of trial and error. It takes time to learn the model and what it's capabilities are. You'll find it has difficulty with handling many characters, highly interactive movement, and timing dialogue with actions. Some people like using prompting tools to flesh out an idea, but the best results will come with your refinements.

My suggestion is to test your prompts using Veo 3 Fast so you dont burn through credits as quickly and refine your prompt based on what you get.

[hallow topaz]

gotcha, thank you!

have some trouble with subtitles being in the video, despite stating not to have any in the prompts

[dim gull]

If you'd like to try my prompting tool, you can find it in my bio. Happy to help in DMs as well!

[hallow pollen]

I keep getting this warning: This prompt might violate our policies about generating sexual content. Please try a different prompt or send feedback. Even though my content is just a person using their phone browsing a generic web page...any idea why this flag happens or what I can do it about it? It totally seems like a bug.

[true igloo]

Hello, I am trying to make multiple clips with flow. Question is, you can change outputs per prompt, but what do I say in order to make sure it splits them up. A lot of times I will select multiple outputs, but it will just do different variations of the same clips. When I try and type out a long prompt, it will just shorten it into one clip.

[olive temple]

yes

[lost robin]

Why everyone here is still playing around with their prompts, I already structured a Veo prompt that can be used in AI filmmaking in the future,

Im already teaching AI how to become an actual human

[dim gull]

Veo can output up to 4 clips per prompt. If you want different scenes then use a new/altered prompt for each new scene.

[feral storm]

does anyone have a prompt for veo3 for when i "jump to" in a scene that gives me a better chance to have audio?

[stark dagger]

There is no audio for that option

[dim gull]

Jump to does work with veo 3 with audio. Its extend that will switch you to veo 2.

[modern crow]

i need like big foot and yeti prompts but cant get good ones from ai

[lime meteor]

What about AWS EC2 located in USA

[wild wyvern]

Ok, sooo.
I got approved for the student free 15 months.
But can only create 3 videos a day and are 8 seconds long.

[hallow topaz]

your prompting tool is a big help, thanks!

[dim gull]

I appreciate you trying it out, thank you!

[sharp ridge]

🎬 [OPENING SEQUENCE — ANIME INTRO STYLE PROLOGUE]

FADE IN — COSMIC DARKNESS.

NARRATION (V.O. — deep, ancient voice, half-whispered like wind through ruin):

“In the deep yesterdays, when the world was shaped by will and wrought by song, it was whole — a circle unmarred. Then came the fire in the dark… and the Sundering, and the Bending of the world, and the Great Silence thereafter. But not all was forgotten.”

[VFX: Starlight dots the void. A great celestial map slowly rotates — constellations shifting, lands drifting apart across millennia like molten plates. Music swells. Sound of echoing chanting in Quenya and Valyrian blended.]

“From the fall of the Elves and the fading of the Dwarrowdelve…
From the turning of the White Trees, and the last ship westward…
From the wars of wizards, the fire of dragons, and the breaking of the towers of men…
There came now… an Age unnumbered.
The Seventh.”

Can anyone generate this

[lost robin]

I pushed veo to its limits and made it like an AI film making tool, although its in tagalog language, what I focus is character continuity, line delivery with appropriate emotions, basically Im teaching AI to act with emotions

[lost robin]

My first ever prompt

[keen mason]

is there an effective way to shorten the legnth of the video outputs? Like I need all of these to be aprox 6 seconds not 8.

[drowsy sluice]

paturo po

Thank you

[reef path]

Are you adding dialogue audio after video generation? or is it in the prompt?

[jaunty saddle]

follow the same sequence for any niche video you want to create in veo3

Subject Description + Camera Perspective / Style + Setting + Dialogue / Voiceover

PROMPT:
"[Subject Description]
A molten gemstone is poured into a cube mold, solidifies instantly, and then is sliced diagonally with a steel blade.
[Camera Perspective / Style]
Smooth transition from overhead pour to close-up side slice. Warm cinematic tones, slow motion.
[Setting]
Polished white marble slab, glowing studio lighting from both sides.
[Sound]
Molten pouring hiss followed by a soft solid crack sound during slicing. Ultra-clean cutting audio."

[lost robin]

Do you see any difference my friends? Thank you for honest opinion

Im still in the process of learning and improving

[daring quail]

hey yall - looking for some help refining a prompt here.

I want to turn this image into a video. the prompt i am trying is:
"pan out to a small red-headed man wearing a similar outfit to the genie, except with a black belt, walking into frame and picking up the lamp. The redheaded man then rubs the lamp. a swirl of the purple smoke should begin to come out of the lamp and the genie should begin to emerge through the smoke, with no legs. as the genie appears, the redheaded man falls back and drops the lamp. the genie, now fully emerged, should look down upon the redheaded man. "

but it is not coming out very well. the genie always has legs or there is just no coherence to what should be happening. an thoughts on alternative prompts for the desired outcome?

[amber hatch]

Hey! Quick question — is there a way to keep the same character (face + voice) across scenes in Veo 3?
When I use “Expand” or “Jump to…”, Flow changes the character or the voice, even if I use the same prompt.
Any tips to make a consistent talking character?

[broken current]

Hey all, Ingredients to video, doesnt it have audio

[jaunty saddle]

Put negative prompt

[jaunty saddle]

Not yet, but train your character on one of the specialized sites to always get a good result. This requires a lot of explanation.

[daring quail]

As in at the end be like

Negative prompt: legs

?

[jaunty saddle]

Negative Prompt Then things you don't want let Chat try it in Google Studio first so you don't lose credit

[wraith lake]

https://youtube.com/shorts/9EuZW62fWGQ?si=ejllcOOGr7gQ0dCt

[fallow frost]

I mean these are excellent. Whatever you are doing is working really well. I don't know how many rolls you did, but getting good results also imporves your ROI, like many others I've spent lots of $

[coarse crane]

PROMPT:
"[Subject Description]
A molten gemstone is poured into a cube mold, solidifies instantly, and then is sliced diagonally with a steel blade.
[Camera Perspective / Style]
Smooth transition from overhead pour to close-up side slice. Warm cinematic tones, slow motion.
[Setting]
Polished white marble slab, glowing studio lighting from both sides.
[Sound]
Molten pouring hiss followed by a soft solid crack sound during slicing. Ultra-clean cutting audio."

[boreal adder]

a fish speaking to fisherman, is possible?

[midnight rampart]

prompt 🙏

[lost robin]

All within prompt

Check this titanic parody with character continuity, I thought Veo 3 was designed to ne an AI film making tool, but it seems that it was not and I just happen to push Veo to its limits

Rose is the real villain

[junior dust]

Do we know a way to get a 9:16 framed video?

[lost robin]

But its difficult to maintain

[languid pendant]

A man sleeping peacefully in a dark room, soft moonlight coming through the window, floating dollar bills and digital coins appearing in the air, futuristic holograms of AI tools working on a laptop beside the bed, cinematic lighting, hyper-realistic, 4K

[empty coral]

Check out @sippettypearl for some fun Ai pettiness!

[junior dust]

Yes this framing. What's a good way to get videos generated in this format?

[lost robin]

Komikis band

[modest zodiac]

I want to make ice cream scoop but never work.... any idea?

[odd birch]

hey, Im from India and want to create

[wraith lake]

I think I did pretty good on this one 😂

https://youtube.com/shorts/TyuJYACBnQ0?si=MOJw-bW3O9S4t0t-

[split falcon]

How?

[jaunty saddle]

follow the same sequence for any niche video you want to create in veo3

Subject Description + Camera Perspective / Style + Setting + Dialogue / Voiceover

PROMPT:
"[Subject Description]
A molten gemstone is poured into a cube mold, solidifies instantly, and then is sliced diagonally with a steel blade.
[Camera Perspective / Style]
Smooth transition from overhead pour to close-up side slice. Warm cinematic tones, slow motion.
[Setting]
Polished white marble slab, glowing studio lighting from both sides.
[Sound]
Molten pouring hiss followed by a soft solid crack sound during slicing. Ultra-clean cutting audio."

[severe moss]

https://www.tiktok.com/t/ZTjcEPPWj/ anyone know the prompt for this type of video? Ive wasted lots of credits trying to figure it out

[idle plover]

anybody know how to generate a prompt for a cyberpunk drone jet flying over a medieval kingdom launching missiles at the buildings? I wanna make it a top gun sort of shot but every time I do the video either has the jet floating doing nothing while fire is going around or the jet is flying backwards.

[torn lichen]

I’m using veo2 pro and I can’t seem to make my ai talk. Everytime I add dialogue no sound comes out in video
How do I fix this ?

[deft owl]

audio is only in veo3 what i understand.

[keen jacinth]

https://youtube.com/shorts/zs9qMsJ1-TM?feature=share

Checkout my first daiy routine video generate from flow

[hidden crow]

why the sound is disable after you create 2nd prompt?
may I know the reasone why the sound get disable?

[hidden crow]

so i guess Veo 3 doesn't support continuous prompt scene

[keen jacinth]

Yes

In veo3 my credits are finished I create a Gemini pro account for 15 month student but my flow credits remain 10 and not updated give me the solution

[hidden crow]

i try creating in Veo 2 same prompt but still not audio

[hidden crow]

the character not look the same even you use the same detail for the character
this not good for creating a film animation

[zinc spear]

is this the prompt for a lava mukbang?

[zinc spear]

how do you turn it into 9:16?

[lost robin]

Ive split the west PH sea , one for PH one for CH , yes to peace.

[modern totem]

how do you maintain same voice across videos?

[oblique heath]

up

[gleaming root]

#flow-tips-n-tricks・✏️ A muscular, intelligent-looking monkey wearing a black winter jacket, woolen cap, and aviator sunglasses is vlogging with a smartphone and a DSLR camera in his hand. The scene is set in the snowy mountains of Kashmir. In the first scene, the monkey walks through snow-covered pine forests, speaking into the camera excitedly.

In the next scene, he stops at a beautiful wooden roadside tea stall surrounded by snow. Steam rises from a hot cup of pink Kashmiri noon chai as he takes a sip, smiling warmly. The monkey looks around at children playing in snow and sheep grazing nearby. The background shows majestic Himalayas under soft sunlight and floating mist.

In an emotional shot, the monkey sits alone on a wooden bench beside the tea stall, looking at the mountains in silence, his camera hanging from his neck. He remembers his jungle life and dreams big.

The final scene shows the monkey standing on a hilltop, taking a selfie with the vast snow-covered valley behind him. He waves at the camera, ending the vlog with a wink and a smile.

Soft cinematic lighting, snowflakes falling gently, warm emotional mood. Camera includes close-up, slow-motion shots, drone-style landscape pans, and POV vlogging angles.

[tight path]

"A sterile, meticulously organized room. Begin with a wide shot emphasizing the stark order and isolation of the space. A man with calm, piercing eyes meticulously cleans a set of antique knives. Transition to an over-the-shoulder shot as he works, showing his precise movements. Reflections of disturbing images – a fleeting scream, a terrified face – briefly appear in the polished blades, captured in a tight close-up. His hands move with deliberate precision. A faint, almost imperceptible smile plays on his lips, revealed in a dolly zoom that pushes into his face. Soft, classical music, slightly distorted, plays in the background."

[sour sphinx]

https://youtube.com/shorts/rCXB-5qpNzg?si=P7wjB3piVieZkmN5

[honest flicker]

hey arsheen. When you say next scene, and final scene- are you giving this within the prompt to flow? does it spit out multiple scenes? or is that sort of just internal notes?

[stark pier]

Enhanced Video:
The resolution is significantly increased, making details like hair and facial features much sharper. The colors, while maintaining their original stylized pink and cyan hues, are now richer and more vivid. The lighting is balanced to be brighter and more dynamic, removing the muddiness of the original and giving it a high-quality, "ultra-clear" feel.

[ruby pendant]

Hi. I'm using VEO 2 - Fast to seamlessly loop a character's isolated head in front of a chroma blue screen and the character simply refuses to keep its mouth closed. I've tried using "This character only speaks through telepathy, no mouth movement, no facial movement, but the words can be heard clearly" in the prompt to no avail. Any suggestions?

[zinc spear]

Oh thanks i'll try your advice

[mild sparrow]

I've already used 500 credits trying and trying but they all come out without audio

A hyper-realistic, live-action, 8-second scene in a 2024 Korean high school classroom, filmed from the side so that both the teacher at the lectern and two students in front of her are in frame. The teacher is a Korean woman in her fifties, with streaks of white in her dark hair, glasses, and a formal suit, fixing her gaze on the two boys, lips pressed together as she struggles to control her anger. The two 17–18-year-old Korean male students in navy school uniforms (blazer, white shirt, black pants, school badge) stand side by side, sneering and grinning at her. One student boldly says directly to the teacher, in clear, audible Korean, “응 우린 촉법이야~”, his tone mocking and defiant. The other snickers, his laugh echoing in the tense, silent classroom. The teacher’s sharp, suppressed breath is audible, and faint, uneasy shuffling is heard from the rest of the class. All dialogue and background sounds—including the student's voice and the teacher's breathing—must be fully and realistically audible in the final video. Cinematic, emotionally charged, and true-to-life. Duration: 8 seconds.

[placid vortex]

is there any reason that veo 3 fast is actually bad? it never generate the right audio even tho.... the video was also a big bugging

[gleaming root]

it's just a details.

[odd birch]

Thread

[vale field]

i am not able to get the ice block to slide, this is my 6th attempt the is the last prompt i tried. anyone have any suggestions?

generate a video a man wearing winter jackets and gloves, standing on a frozen lake.

the man is holding a sheet of ice that used to be part of the frozen lake

The camera angle is from behind the person holding the ice from over his shoulder.

He forcefully throws the ice he is holding with one hand, as the camera follows the ice block till it makes contact with the frozen lake surface.

It falls and slides and shatters completely on contact cracking and shattering on a frozen lake with all the tiny broken particles still sliding on the surface as the throw was full of force. intricate fracture patterns, splintering ice sliding on the frozen floor, high detail, sharp focus, natural lighting, winter landscape.  No part of the original ice sheet is remaining.

The shattering of the ice sheet make a great shattering sound and then followed by a ringing till the ice particles are sliding on the frozen lake.

[odd birch]

Enhanced Video:

[pearl peak]

Try icecream instead of ice cream

[split falcon]

What was your prompt setup?

[wraith lake]

i break down each part itno different scenes, i start by asking chatgpt "Create 3 bigfoot vlog video scripts where he interacts with modern human technology or trends." then i pick out of them and tell it to add a voiceover line for every storyboard, using the format: He says: "...". then i have chatgpt break it down into story boards. for this video it was "create a story board for vlog 2" and it gives me the video scene by scene. so the first one was Scene 1 – “The Discovery”
Visual Description:
Bigfoot, sitting in a mossy forest clearing, discovers an iPad left behind at a campsite. He pokes it curiously. The screen lights up, showing AI-generated art of Bigfoot.
Camera Direction:
Over-the-shoulder shot of the iPad screen → zoom on Bigfoot’s wide-eyed reaction.
He says:
"Why do I look like a gym rat in all these?! I don’t even know how to spell protein!" then just keep running that prompt until i get one i liked. then start with scene 2 and repeat that process.

[split falcon]

7 videos later veo 3 Quality still isn't generating any dialogue for me. I've tried with quotes and without quotes

[wraith lake]

i use veo 3 fast for my videos to 1 save the credits and 2 ive had the most susscess with it/

[split falcon]

Ok. I've been trying to use an image to begin the frame but that action doesn't seem to support Fast.

[random nacelle]

"Realistic video of a small empty bedroom with no windows during the night. The TV was playing some 2000s commercials and there was a bed and a brown carpet and the lights were dim. Suddenly, about 3 seconds in the video the power goes out with a robot voice saying 'POWER FAILURE', then 1 second later a red cartoony letter Z walks in the bedroom and says 'Zeeeeeeeee' in a robot voice and then when the Z stops walking he power goes back on. When it goes back on, a celebration noise plays, a sign in white letters on a blue background saying 'IT'S HERE' appears and there was colorful confetti. With small subtitles."
I want to see a Veo 3 video with this prompt.

[glacial sail]

After wasting so many credits figuring out why my videos showed a "Video audio generation failed". I decided to switch from Chrome to another browser & now they have sound. Just an FYI for those with the same issue, change your browser.

[split falcon]

what browser are you using now?

[glacial sail]

Unfortunately, Microsoft Edge. Haven't tried other browsers. It's odd because Chrome was working fine with many other generations.

[lost robin]

Coz you all here prompting like kids ,

You have to learn ghe correcct prompt structure

[glacial sail]

How do i make a character scream? I wrote "Then, in a loud startled male voice, Kevin screams: "Ahhh!" "

[lost robin]

Also your imagination are too much for Veo thats why your videos dont have any sound even if you choose Veo 3, but your prompt is too much for Veo, it will automatically generate you a veo 2 format that has no audio

Ive learned that the hard way I already spent 50k AI credits within this month

You have to learn how to balance the tone and elements of the video , Veo is smart but its not magic it has limitations too

So im giving you an advice from someone who burned 50k AI credits for trial and error

[random nacelle]

What's taking so long?

[split falcon]

What is the correct structure my friend? I want to prompt like an adult!!!

[heady shadow]

Anyone can help?

I have tried more than 10 videos on Flow to generate an orang utan interview a human, however Flow doest understand and only the human talks instead of the orang utan.

However, I tried the same prompt in Gemini it works with some flaws. but I am not able to reproduce the same result after fine tuning it.

doest anyone knows how to tune tthis?

[grizzled galleon]

I use brave and firefox both work fine for me.just in case

[tulip nacelle]

Hi

[reef sorrel]

Handheld amateur footage : highly detailed shot opens over the shoulder of a man wearing a heavy winter jacket and gloves, standing on a vast, frozen lake under crisp natural winter light. He throwing ice block with one hand across the slick surface. The camera executes a dynamic tracking shot, following the thrown ice sheet as it skims rapidly across the frozen expanse., the ice sheet dramatically slides and continue to skid outwards, across the ice with high-definition clarity. sounds of the sliding ice, all captured within the desolate beauty of the winter landscape

[sinful spoke]

It's amazing how veo managed to get his fingers and keystrokes completely in sync with the music... How can I get him to continue playing the song in the next scene?

Bonus points for anyone who recognizes the scene 😄

[sullen kestrel]

#flow-tips-n-tricks・✏️ 🎬 Scene 1: Chaos of the World (0:00 – 0:30)
“Cinematic urban street at night filled with chaos — people yelling, protest signs waving, sirens flashing red and blue, phones glowing in people’s hands, a single man standing still in the middle of the street frozen in time, dramatic lighting, emotional tension, wide angle, dark gritty tones.”

[tired spear]

#In to the son shine full of the light keeping and flashing from morn till light

[buoyant crane]

hay need some feedback please i follow this prompt structure here for veo 3

1. Subject (who's in the shot)

2. Action (what are they doing)

3. Setting/Environment (where are they or where is this happening time of day weather, so on)

4. Camera (How should it be filmed or how should the camera be moving)

5. Lighting/Mood (What lights are present and what is the atmosphere)

6. Audio and Dialogue (What do we hear or what is being said)

7. Visual Style (What should this feel like visually (E.g. should this be realistic, etc.)

Note: the more detail and the more clearly you describe thing's the more control you have so add as much detail as possible for the best result.

and im writing prompts like this:

1. Subject A woman in her early 30s, Subject 23, sits strapped into a heavy-duty, padded sleep experiment chair. She has pale skin with visible sleep deprivation under her eyes — deep purplish eye bags and light facial twitching. Her head is shaved at the crown to expose scalp electrodes that are tightly suctioned to her skin. Her mouth is clamped open with a plastic dental gag. She wears a light blue hospital gown stained with old blood around the neckline. Her arms are secured to the chair with thick black leather restraints. Wires lead from her chest to machines off-screen. Her fingers twitch subtly, and tears streak down her face — but her eyes remain unnaturally wide open. 2. Action Subject 23 sits motionless except for involuntary twitching and slight breathing. Her pupils are dilated and locked onto a fixed point ahead. Her jaw vibrates slightly against the gag. Behind the safety glass, a male technician in his late 20s (medium build, messy hair, lab coat) watches with increasing concern, pressing a button on the intercom that emits no response. The monitor behind Subject 23 displays violently spiking brainwave activity. Her body trembles gently, as if fighting paralysis. 3. Setting/Environment The scene takes place in a dimly lit underground sleep research lab.

The room is a sterile, windowless concrete chamber with metallic blue-gray walls. Exposed pipes run along the ceiling, and condensation drips occasionally. Multiple monitors, medical devices, and an old ventilator hum quietly in the background. The time is unclear, but a red digital wall clock reads "02:44 AM." The subject is placed in the center of the room beneath a surgical light array. The control booth behind reinforced glass contains blinking monitors, papers, and spilled coffee. The room temperature is visibly cold — condensation fogs around the subject’s shallow breath. 4. Camera The camera slowly dollies in from a medium-wide angle toward the subject’s face, passing slightly to her left side. As it approaches, the depth of field narrows, and focus tightens on her twitching eye and sweat-drenched cheek. Subtle handheld motion is introduced at the end to create a feeling of instability and dread. 5. Lighting/Mood Fluorescent surgical ceiling lights above Subject 23 flicker inconsistently, casting sterile white-blue light with harsh shadows. Occasional surges cause momentary darkness. A dull red hue from the emergency standby light pulses in the background. The overall mood is cold, clinical, and unnervingly quiet, with moments of intense visual strobing that highlight the subject’s terror. The contrast between harsh lighting and deep shadows emphasizes discomfort and tension. 6. Audio and Dialogue We hear the low hum of machinery and a soft, glitching beeping from the brainwave monitor. Subject 23's restrained breathing and faint choking sounds are audible. A distorted intercom crackles: Technician (muffled through glass): “Subject 23... do you copy? We’re seeing spikes—are you awake?” A heartbeat monitor accelerates unnaturally, syncing with a rising droning sound layered beneath. 7. Visual Style This should be highly realistic and cinematic, using filmic textures, grain

overlays, and subtle lens flaring. The image has a desaturated blue-green tint. Emphasis on extreme detail — sweat, skin texture, reflective glass. The scene should evoke David Fincher’s industrial tone with the eerie sterility of a medical sci-fi horror. Camera movement should feel precise but increasingly unstable as the tension escalates.

is this right for veo 3 ?

and heres my result

[sinful spoke]

Can anyone help? Got this video as a small intro for something i'm working on.

Its always Glitch talking, i tried 5 different prompts already, results always good but i can't get it to have both characters introduce themself 😦

Generated with this prompt:

subject: Two AI characters — Glitch, a human-sized humanoid hologram with neon glitch effects, and Byte, a small floating metallic cube about the size of Glitch head with glowing eyes.
action: The characters are introducing themselves to the audience in a fun and stylish way. Byte floats beside Glitch’s shoulder. Glitch is sitting on a modern, black, leather couch. The camera pulls back to reveal the full studio.
setting: A sleek, modern TV studio — clean wood and metal surfaces, LED accent lighting, podcast-style layout.
camera: Starts with a medium shot on the characters, then smoothly zooms out to a wide studio view.
lighting/mood: Bright and polished with a futuristic edge. Soft key lights and cool LED accents (blue/purple). Friendly and tech-forward mood.
dialog: The characters each are delivering playful, charismatic introductions to the audience.
visual style: Cinematic realism with subtle sci-fi flair. Glitch has expressive holographic features with flickering glitch effects; Byte glows softly and moves smoothly beside him.

[quasi condor]

Hey everyone, can I ask you if you know how to create this vidéo where you see a yéti vloging

[prisma schooner]

Write a script like big furry white yeti vlogging with wooden style selfie stick in Walmart or something and create dialogue

Make sure you describe it well

[quasi condor]

I tried so many prompt is this way, but it look always like 60s first animation

[prisma schooner]

I can help you check dms

[oak briar]

An anime boy saying " hello my name Michael"

[blissful prism]

how do you guys get 30 second videos you see on socials? most I can get is 8, do you combine them somehow?

[crystal sand]

I have the same question: how do you create longer videos?

[jaunty saddle]

Hello, it is not currently available. You can divide this by the number of seconds available to get a long video, but that will cost a lot.

[nova siren]

the same behaviour, on gemini works, but on flow doesnt.

[crystal sand]

Is it possible to maintain character consistency somehow? Like create an 8-second video and then continue the "story"??

[royal void]

hello has anybody tried making videos with consistent setting/ characters? did it work?

[cerulean kite]

has anyone tried to prompt an audience where the first row is all specific characters in a specific order?

I have 5 character prompts that pretty much produce the same audience member every time i just can't get them to sit in the right seats in the right order

this one was almost perfect but the black girl morphed into the white girl

This is my last try..
Focus is tightly held on the first row of patrons, each with highly specific features and clothing, while all other rows behind them are shrouded in soft shadows and indistinct silhouettes.

Sitting in the first seat on the right side of the front Row is a Middle-aged white male, 50s, balding with a fringe of silver hair, wearing a bright red Hawaiian shirt with white floral print, round wire-frame glasses.

Sitting in the seat to the left of the white male, in the front row, is a Black woman, professional, fit, early 40s, tightly braided hair pulled into a bun, wearing a black turtleneck and gold hoop earrings. She sips a glass of red wine with her left hand.

[cerulean kite]

my prompt is actually working now.. not every time but well enough

[stiff root]

This is the promt we are using and they keep saying our name wrong its Geaux Styles ("Geaux" is pronounced the same as the English word "go". It's a Cajun French spelling used in Louisiana, particularly in the context of LSU (Louisiana State University) and its sports teams, like "Geaux Tigers" which means "Go Tigers".

Prompt:Time + location + environment Inside GEAUX Styles boutique (pronounced “Go Styles”), Mandeville, Louisiana. Afternoon sunlight filters through the front windows. A mannequin in a breezy summer dress is staged next to a smiling woman dressed in complementary colors. 2. Detailed scene action The woman gestures to the mannequin, then faces the camera with warmth. She raises a branded shopping bag and delivers her line with a confident smile. 3. Dialogue (timed for 5–6 seconds) Geuax Styles "(pronounced “Go Styles”)— styled me perfectly. Shop local, Mandeville!" 4. Camera style and movement Medium shot with a soft rack focus from mannequin to the woman as she speaks. Slight handheld cam adds authenticity without distracting from the subject. 5. Audio/musical elements Charming acoustic strum or gentle indie pop beat. Ambient boutique sounds like soft footsteps or subtle jazz in the background. 6. Mood, lighting, and cinematic style Fresh, light, community-focused. Pastel tones, soft sunbeam highlights, and a cozy editorial boutique vibe. Feels personal, local, and effortlessly stylish.

I keep wasting credits

[graceful marsh]

I'm getting alot of achatacter incosistency even with the most descriptive prompt - even when every single feature is defined

[cerulean kite]

So spell it GO in the prompt..

for the speaking part remove any reference to the proper name

[reef sorrel]

The camera, in a medium shot with a subtle handheld feel, begins on a mannequin dressed in a breezy summer dress, soft rack focus then shifts to a young adult woman with long, dark hair and light skin, wearing a pastel yellow top and white shorts, standing beside it. She gestures gracefully towards the mannequin, her expression warm and inviting as afternoon sunlight filters through the front windows of the Geaux Styles boutique, casting soft sunbeam highlights across the scene. The woman then turns to face the camera, her confident smile radiating as she raises a branded Geaux Styles shopping bag. She says, "Go Styles styled me perfectly. Shop local, Mandeville!" The lighting maintains a fresh, light, and community-focused mood with pastel tones, enhancing the cozy editorial boutique vibe. The shot ends with her confident smile. Accompanying the scene is a charming acoustic strum or gentle indie pop beat, subtly mixed with ambient boutique sounds like soft footsteps.

[reef sorrel]

trying out your core idea concept. i will be testing the exact thing (referring to your video) later.

A medium shot shows a male orangutan, with rich auburn fur and intelligent eyes, seated at a professional desk, intently interviewing a human. The orangutan, dressed in a small, tailored suit, leans slightly forward, holding a tiny microphone up to a female human with long, straight brown hair and light skin, who is seated opposite him. The human, dressed in a smart business casual outfit, gestures subtly with her hands as she speaks. The orangutan's mouth moves in sync with his dialogue, "And what are your thoughts on the ecological impact of human development?" The background is a blurred, professional office setting with soft, natural lighting coming from a nearby window, casting a gentle glow on the scene. The camera slightly pans back and forth between the two, emphasizing the unique interaction.

[reef sorrel]

A photorealistic, slightly whimsical scene unfolds in a comfortable, open-air conservation area bathed in warm afternoon light. A wise-looking orangutan sits across from a strong, friendly older man affectionately referred to as "uncle". The orangutan leans forward slightly, gesturing with a hand, and clearly appears to "speak" the line "Wah, uncle, you look so strong! What's your secret?" in a distinct Malaysian accent. The man smiles kindly, about to respond. The camera maintains a static medium shot capturing both subjects equally throughout the brief exchange. The audio prominently features the orangutan's voiced question with its specified accent, accompanied by faint ambient jungle sounds and perhaps the man's initial chuckle.

[reef sorrel]

LMFAO.

The camera slowly tracks forward, focusing on an outdoor interview scene in Kuala Lumpur, capturing the vibrant cityscape blurred in the background. A male orangutan, with expressive eyes and rich reddish-brown fur, sits comfortably in a small, ornate chair. He holds a microphone in his left hand, occasionally gesturing with his right as he speaks. Opposite him, an elderly Chinese Malaysian man with a fair skin tone, long gray hair and a long beard, sits calmly. He is dressed in a traditional ornate red and gold robe and a matching hat, his hands resting gently in his lap. The orangutan, animated, leans slightly forward and, with a slight Malaysian accent, asks, "Wah, uncle, you look so strong! What's your secret?" The scene is brightly lit by natural daylight, emphasizing the intricate details of the man's robe and the orangutan's fur.

A wide-angle tracking shot follows a male orangutan with reddish-brown fur, light brown skin, and expressive dark eyes, wearing a tiny, perfectly tailored gray suit, as he walks alongside a Chinese Malaysian elderly man. The man has a fair skin tone, a long gray beard, and long gray hair, and is dressed in an ornate red and gold traditional robe with a matching hat. They are strolling through a bustling Kuala Lumpur street market, with blurred city lights and distant Petronas Towers in the background. The orangutan, holding a small microphone, turns his head towards the man, his lips moving as if asking, "Wah, uncle, you look so strong, what's your secret?" The camera maintains a low angle, emphasizing the orangutan's perspective, with natural, bright market lighting.

[rapid orbit]

Un faux prophète en habit religieux crie dans un micro devant une foule excitée, sous un abri de fortune

A Nigerian street preacher dressed in flashy robes, shouting into a microphone under a canopy, plastic chairs around, a crowd claps in excitement under the sun.

Ggg

[upper magnet]

A cinematic, emotional short film about an old man named Abdul Hakeem, living alone in a small, broken house in a remote Pakistani village. He’s in his 70s, with deep wrinkles, white beard, and kind, tired eyes. Every day he walks to the fields with a wooden stick, still working hard despite his age. Show his struggles — eating alone, looking at old photographs of his late wife and children who are now distant or abroad. The story unfolds slowly with soft background music, village ambient sounds, and heartfelt narration (in Urdu with subtitles).

One day, he finds a lost child crying near the fields. Despite his own loneliness, he takes care of the child like a grandfather — sharing food, teaching him how to read from old books, and giving him hope. This moment brings life back to Abdul Hakeem’s dull routine.

In the end, we see the child’s parents return and thank him with tears. As they leave, Abdul smiles with moist eyes, feeling fulfilled. The story ends with a message on screen: ‘Kabhi kisi ki zindagi mein roshni ban jao — woh tumhari zindagi bhi roshan kar dega.’

Use realistic visuals, soft light, traditional Pakistani village setting, emotional expressions, and focus on storytelling over dialogue. Add background score with flute, sarangi, and village nature sounds

[rapid orbit]

@cerulean kite A Nigerian street preacher dressed in flashy robes, shouting into a microphone under a canopy, plastic chairs around, a crowd claps in excitement under the sun.

[upper magnet]

A cinematic, emotional short film about an old man named Abdul Hakeem, living alone in a small, broken house in a remote Pakistani village. He’s in his 70s, with deep wrinkles, white beard, and kind, tired eyes. Every day he walks to the fields with a wooden stick, still working hard despite his age. Show his struggles — eating alone, looking at old photographs of his late wife and children who are now distant or abroad. The story unfolds slowly with soft background music, village ambient sounds, and heartfelt narration (in Urdu with subtitles).

One day, he finds a lost child crying near the fields. Despite his own loneliness, he takes care of the child like a grandfather — sharing food, teaching him how to read from old books, and giving him hope. This moment brings life back to Abdul Hakeem’s dull routine.

In the end, we see the child’s parents return and thank him with tears. As they leave, Abdul smiles with moist eyes, feeling fulfilled. The story ends with a message on screen: ‘Kabhi kisi ki zindagi mein roshni ban jao — woh tumhari zindagi bhi roshan kar dega.’

Use realistic visuals, soft light, traditional Pakistani village setting, emotional expressions, and focus on storytelling over dialogue. Add background score with flute, sarangi, and village nature sounds

[solar raft]

I need help. I'm using Flow, but even though I make it clear in the prompt to use the Brazilian Portuguese voice, it's creating a video in Portuguese from Portugal. What do I need to do?

[reef sorrel]

Did u include the actual dialogue in the prompt

A dynamic, fast-paced shot, opening with a wide-angle view of a bustling, sun-drenched outdoor market in Brazil. The camera rapidly glides forward, weaving through vibrant stalls overflowing with tropical fruits and handmade crafts, capturing the lively atmosphere with a slight motion blur to emphasize speed. The scene then transitions to a medium shot, focusing on a young adult woman with light skin and dark hair, wearing a casual, brightly colored sundress. She is enthusiastically engaging with a vendor, speaking clearly in Brazilian Portuguese, her hand gestures animated as she points to an item. Her voice is crisp and distinct, showcasing the unique phonetics of Brazilian Portuguese. The warm, natural light illuminates her face, highlighting her expressive features against the slightly blurred, colorful backdrop of the market. The camera continues its forward motion, past the woman, and then smoothly pans to reveal a diverse group of market-goers of various ages and skin tones, all conversing in authentic Brazilian Portuguese, their voices blending into a rich, energetic soundscape that unmistakably represents the vibrant culture of Brazil. The overall aesthetic is one of high contrast and saturated colors, reflecting the lively Brazilian setting.

[reef sorrel]

The camera slowly pans across a rustic, sun-drenched Pakistani village as a drone shot follows an elderly man with a long white beard and a weathered face, wearing a traditional shalwar kameez, as he walks with a wooden stick towards lush green fields. Close-up shots show his calloused hands gripping the stick, his kind, tired eyes reflecting a lifetime of toil under the soft, golden hour sunlight. Later, inside a small, sparsely furnished mud house, the same elderly man is seen eating a simple meal alone, the camera focusing on his solitary figure amidst the quiet hum of village sounds. He then slowly picks up a faded photograph, his gaze lingering on the smiling faces of a woman and children, a melancholic expression settling on his deeply wrinkled face. The scene transitions to the man walking through the fields when he spots a small child, aged around five, with tear-streaked cheeks, crying softly near a cluster of tall grass. The old man, his face etched with concern, gently approaches the child, extending a comforting hand. The next sequence shows the man and child sharing a modest meal, followed by the man patiently teaching the child from an old, worn book, their heads close together under the warm glow of a flickering oil lamp, the child's initial apprehension slowly replaced by curiosity and joy. A wide shot captures the child's parents, a man and a woman, both appearing to be in their late 20s, with expressions of profound relief and gratitude, embracing the elderly man, tears streaming down their faces as they express their heartfelt thanks. As the parents and son walk away into the distance, the camera zooms in on the old man’s face, a gentle smile gracing his lips, his eyes moist with a sense of peace and fulfillment, the setting sun casting long shadows behind him. The scene concludes with a message appearing on screen: 'Kabhi kisi ki zindagi mein roshni ban jao — woh tumhari zindagi bhi roshan kar dega.' The entire film is underscored by a soft, emotional background score featuring flute and sarangi, interwoven with natural village ambient sounds. Sorry Flow does not accept non english prompt atm

[solar raft]

I used this prompt, what i need to change?

Create a realistic and humorous video. Scene: A sunny outdoor town square during the day. Main character: A white man, about 40 years old, short brown hair, wearing black pants, white sneakers, and a white dress shirt with the word "VAZOLI" written on the left side of the chest. He is tied to a wooden pole in the center of the square. Extras: A group of elderly people (men and women, around 65+ years old), dressed casually. They are angrily surrounding the man, holding lit torches (flaming), some waving handbags or walking sticks. The scene should feel exaggerated and funny, not threatening. Dialogue (must be spoken in native Brazilian Portuguese – português do Brasil): Crowd (angrily shouting in sync, holding torches): "Queimem o correspondente bancário! Queimem o correspondente bancário!" Tied man (shouting loudly, slightly desperate but calm): "Eu consigo desbloquear os benefícios para empréstimo!" Crowd (suddenly happy, cheering in sync): "Soltem o correspondente bancário! Soltem o correspondente bancário!" Then a few elderly people untie the man, who smiles and pulls out a tablet to show them something. ❗️Voice & Audio Instructions: All lines must be spoken in native Brazilian Portuguese (PT-BR). Absolutely avoid European Portuguese (Portugal) accent. Use a voice actor or AI voice with authentic Brazilian accent and pronunciation. Do not include any subtitles or captions in the video. ✅ Visual & Style Instructions: Characters must look like real humans (no avatars or cartoons). The scene should look natural but exaggerated in tone (comedic realism). Torches should have animated flame effects to enhance the humor.

[reef sorrel]

hmm i need to figure it out how to .. make Flow generate the video but the challenge is that flow doesnt accept non english prompt

i tried but

nothing came out exactly like your core idea

@solar raft try generate the video with that prompt u wrote in Gemini

https://gemini.google.com/veo

[solar raft]

I've already tried but it didn't work either. How did you get your video to come out with audio in Portuguese without putting the dialogue in the script?

[reef sorrel]

I’m not sure how i did it and how it works either. All I know is Veo 3 has prompt enhancement turned on by default.

[solar raft]

I'm saying about this prompt

[reef sorrel]

yeah i didnt add the dialogue

it was the model itself doing it

its prompt enhancement in the model

[solar raft]

ok i got it

[solar raft]

That's it, but this accent is from Portuguese from Portugal and not from Brazil lol

[glossy wigeon]

I need help. I wanted to create an AI video, and I used Gemini, which did a good job, but it didn’t have audio, so it’s not good enough for my needs. Then I wanted to try Flow, but it doesn’t allow me to use it. It says I need the Google AI Pro subscription, but I already have it for free for this month. For some reason, Flow doesn’t recognize that I have the subscription.

[reef sorrel]

i gave up lol

[solar raft]

That's it, now teach me lol

[reef sorrel]

oh

i didnt use flow for this one. it was generated in google vids (using veo 3)

A dynamic medium shot captures the unfolding chaos and resolution in a brightly sunlit, bustling Brazilian Portuguese town square. The camera subtly pans and zooms to follow the action. Initially, a white man, approximately 40 years old, dressed in black pants, white sneakers, and a white shirt emblazoned with 'VAZOLI', is prominently tied to a weathered pole. He is surrounded by a large group of elderly individuals, around 65 years and older, who are actively waving lit torches, brandishing handbags, and gesturing with walking sticks. Their faces are comically contorted in exaggerated anger as they shout in unison, "Queimem o correspondente bancário!". The tied man, displaying a mix of desperation and calmness, loudly exclaims, "Eu consigo desbloquear os benefícios para empréstimo!". Immediately, the crowd's angry expressions dissolve into beaming smiles, and their shouts transform into joyous cheers of "Soltem o correspondente bancário!". A few individuals from the crowd quickly step forward to enthusiastically untie him. As he is freed, the man grins broadly and pulls out a tablet, its screen casting a soft glow on his face. The scene maintains a lighthearted, comedic tone, punctuated by the distinct sound of the crowd's angry shouting transitioning into cheerful acclamations, all set against the ambient backdrop of general town noise.

https://docs.google.com/videos/

[solar raft]

oh ok, now I'm getting the hang of it, man, thanks a lot

[rapid mason]

Anyone can help me on this prompt. I want to achieve aerial view of tiger running but i kept getting rear view after many attempt. 500 credit wasted lol. Kindly seeking help

[severe aspen]

can you paste your prompt?

[reef sorrel]

A highly detailed tiger, blazing tiger running fast in middle of vivid blue and
orange light neon city landscape. The tiger's fur is engulfed in flames,
crackling and sparking as it runs. Camera angle aerial view neon city.
Cinematic film

[gleaming root]

#flow-tips-n-tricks・✏️ "विक्की का क्रिकेट वाला Vlog" कहानी के लिए सीन-बाय-सीन टेक्स्ट-टू-वीडियो प्रॉम्प्ट। इन्हें आप सीधे वीडियो जेनरेशन मॉडल में इस्तेमाल कर सकते हैं।

टेक्स्ट-टू-वीडियो प्रॉम्प्ट: विक्की का क्रिकेट वाला Vlog
Scene 1: स्टेडियम का माहौल

Prompt: एक भारतीय क्रिकेट स्टेडियम का पैनोरमिक वाइड शॉट दिखाएँ, जो मैच से पहले दर्शकों से खचाखच भरा हो। माहौल ऊर्जा और उत्साह से भरपूर है। भारतीय टीम की नीली जर्सी पहने प्रशंसक नारे लगा रहे हैं और झंडे लहरा रहे हैं। कैमरा धीरे-धीरे भीड़ पर घूमता है।

Dialogue: (कोई डायलॉग नहीं, सिर्फ बैकग्राउंड में भीड़ का शोर और "इंडिया! इंडिया!" के नारे)

Scene 2: विक्की का परिचय

Prompt: कैमरा भीड़ के बीच एक इंसान जैसे दिखने वाले बंदर, विक्की पर फोकस करता है। विक्की ने भारतीय क्रिकेट टीम की नीली टी-शर्ट और स्टाइलिश काले चश्मे पहने हैं। वह एक सेल्फी स्टिक पकड़े हुए है और सीधे अपने फोन के कैमरे में देखकर पूरे जोश के साथ व्लॉग रिकॉर्ड कर रहा है।

Dialogue:

विक्की (उत्साह से): "तो हैलो दोस्तों! स्वागत है आपका मेरे चैनल पर! मैं हूँ आपका अपना दोस्त, विक्की! और जैसा कि आप देख सकते हैं, हम आ पहुंचे हैं स्टेडियम के ठीक बाहर! माहौल एकदम जबरदस्त है, और आज तो हम अपनी टीम को जिताकर ही रहेंगे! चलिए, अब अंदर चलते हैं और आपको दिखाते हैं असली एक्शन!"

Scene 3: गेट पर टकराव

Prompt: विक्की आत्मविश्वास से स्टेडियम के मेन गेट की तरफ बढ़ता है। जैसे ही वह अंदर जाने की कोशिश करता है, एक 30 साल की खूबसूरत महिला पुलिस ऑफिसर अपनी वर्दी में विनम्रता से उसे हाथ उठाकर रोकती है। विक्की को देखकर आसपास लोगों की भीड़ जमा होने लगती है, जो अपने फोन से वीडियो बना रहे हैं।

Dialogue:

महिला पुलिस ऑफिसर (शांति से): "माफ़ कीजियेगा, आप अंदर नहीं जा सकते।"

kya koi in prompt ko use karke mujhe video clips de sakta hai ..yah mujhe inhe banane ka sahi way bataye..veo se generate nahi rahe 1 scene ke baad vo error de de raha hai jabki paid subscription hai..

ek scene ke baad aage nahi ban raha

[reef sorrel]

generating atm

it laso gave me rear view

🙂

gonna have to iterate your prompt many times now brb!

imagen 4 model seems to understand , veo 3 straight up want to focus on tiger's rear view

[hearty estuary]

A dynamic medium shot captures the unfolding chaos and resolution in a brightly sunlit, bustling Brazilian Portuguese town square. The camera subtly pans and zooms to follow the action. Initially, a white man, approximately 40 years old, dressed in black pants, white sneakers, and a white shirt emblazoned with 'VAZOLI', is prominently tied to a weathered pole. He is surrounded by a large group of elderly individuals, around 65 years and older, who are actively waving lit torches, brandishing handbags, and gesturing with walking sticks. Their faces are comically contorted in exaggerated anger as they shout in unison, "Queimem o correspondente bancário!". The tied man, displaying a mix of desperation and calmness, loudly exclaims, "Eu consigo desbloquear os benefícios para empréstimo!". Immediately, the crowd's angry expressions dissolve into beaming smiles, and their shouts transform into joyous cheers of "Soltem o correspondente bancário!". A few individuals from the crowd quickly step forward to enthusiastically untie him. As he is freed, the man grins broadly and pulls out a tablet, its screen casting a soft glow on his face. The scene maintains a lighthearted, comedic tone, punctuated by the distinct sound of the crowd's angry shouting transitioning into cheerful acclamations, all set against the ambient backdrop of general town noise.

[shy fog]

Does anyone have a rock solid prompt for a gorilla that looks the same every time?

[gleaming root]

"A cinematic, high-quality, front-facing shot of an anthropomorphic, human-like monkey named Vicky. He is in the middle of a large, cheering crowd of cricket fans outside a huge stadium. Vicky is wearing a blue Indian cricket team t-shirt and stylish black sunglasses. He is holding a selfie stick and vlogging enthusiastically into his smartphone camera with a wide, excited smile. The atmosphere is electric and celebratory. He gestures towards the stadium behind him, then starts walking towards the entrance, all while looking into his camera. The shot is from the perspective of his smartphone camera."

What's wrong with this prompt, Veo it's not generating it? please help me

[dry hornet]

My videos are coming out with strange, very robotic voices, does anyone have any idea why?

[dim gull]

Yeah I've tried every variation of prompt for this type of shot and Veo 3 just doesn't do it. Top-Down, Aerial, Drone-View, Crane Operated, Subject directly below.. etc.

[sinful spoke]

Can someone help me with this prompt? Whatever i do, the character is a magician just summoning the vest onto himself, or the cap flies in from nowhere etc

Jack Merrow, a mid-40s investigative reporter.
He has just stopped deep inside a fog-covered Montana forest, far from any visible road, surrounded by towering pine trees and damp undergrowth.
He steps out of the Jeep, walks to the rear hatch, and opens it quietly. He pulls out a black bulletproof vest labeled “PRESS,” straps it on over a dark hiking jacket, slings a backpack over one shoulder, and puts on a black cap while scanning the treeline.
Photorealistic cinematic style with subtle documentary tension.
Slow, top-down drone shot hovering steadily above the vehicle, with a slight drift to the right for dynamic movement.
Jack is clearly visible at the rear of the Jeep, framed in the lower third of the shot, with fog-covered forest filling the upper two-thirds for scale and isolation.
Faint early-morning forest ambience: distant birds, soft wind, light gear movement, all slightly muffled by fog.

[dim gull]

Try to break things up, you're asking him to get out of the jeep, walk to the back and open the hatch, grab and don a bulletproof vest, and put a backpack on and put on a baseball hat, all in 8s.

[sinful spoke]

actually....fair point yea indeed i can do it in 8s! really! 😂

[bleak badger]

I just bought the subscription but why won’t it let me create any vids? When I hit the button on the top right with the video tape it takes me back to this.

[dim gull]

https://labs.google/fx/tools/flow

[sinful spoke]

i've seen some people do things like
subject: xxx
action:xxx
composition:xxx
in their prompts, i thought just the order is important, do we have to add those section names?

[dim gull]

I've found it helpful. That's how I have my prompting tool set up.

[dusty bluff]

Why i cant able generate videos in hindi ? Whenever i type it in veo it says google dont support this language for prompt

[plush oracle]

Create a hyper-realistic close-up video of a sharp stainless steel kitchen knife with a matte black handle slicing through a glass blueberry. The blueberry is entirely made of translucent glass, deep blue and glossy, with delicate highlights and realistic surface dimples. Under soft diffused lighting, it glistens like a precious ornament. The knife blade is midway through the cut, cracking cleanly through the glassy surface. A small, pristine slice peels away, revealing a glowing, detailed core with inner glass structure and fine light refractions. The scene is set on a neutral-toned marble countertop with ambient reflections and no background distractions. No music, no cinematic black bars, no watermarks, no captions. The camera is fixed in a frontal macro close-up and slowly pushes in toward the cutting action.

[plush oracle]

Can you translate?

[plain shuttle]

Why are there water marks on the videos generated on the pro accounts?

[tidal forge]

Jeep

#flow-tips-n-tricks・✏️ Jeep in sunny day

Create a hyper-realistic close-up video of a sharp stainless steel kitchen knife with a matte black handle slicing through a glass blueberry. The blueberry is entirely made of translucent glass, deep blue and glossy, with delicate highlights and realistic surface dimples. Under soft diffused lighting, it glistens like a precious ornament. The knife blade is midway through the cut, cracking cleanly through the glassy surface. A small, pristine slice peels away, revealing a glowing, detailed core with inner glass structure and fine light refractions. The scene is set on a neutral-toned marble countertop with ambient reflections and no background distractions. No music, no cinematic black bars, no watermarks, no captions. The camera is fixed in a frontal macro close-up and slowly pushes in toward the cutting action.

[sinful spoke]

So you add those to your prompts?
"Subject: an old man does xyz, action: he drinks a coffee" kinda thing? comma seperated?

[dim gull]

Yeah, this is the structure for mine.

[sage cipher]

eu tenho precisamente o problema inverso, quero que falem Português de Portugal, e os videos, especialmente os de desenhos animados produzem Português do Brasil. No prompt deves incluir "speaking in Portuguese language from Brasil" no caso de Portugal "speaking in European Portuguese language (as vezes funciona e as vezes não) igualmente acontece com a geração não desejada de sub titulos, é impossivel de evitar por enquanto.

I have precisely the opposite problem, I want them to speak Portuguese from Portugal, and the videos, especially cartoons, produce Portuguese from Brazil. In the prompt you should include "speaking in Portuguese language from Brazil" in the case of Portugal "speaking in European Portuguese language (sometimes it works and sometimes it doesn't) the same thing happens with the unwanted generation of subtitles, it is impossible to avoid for now.

[sage cipher]

I think it refers to the prompt that has to be written in English

[dusty bluff]

Yes but how come he generates in hindi language

[sage cipher]

I think they disapear only in Ultra

[sage cipher]

yes the characters can speak Portuguese also, but the prompt can´t be in Portuguese

[timber inlet]

I am trying to make more slices but unable to perfect it can you guide me how to make more slices as video ends with 2 and half slices

[sage cipher]

only if the character is The Flash 😛

[dusty bluff]

So how to make them speak in different languages without prompts

[reef sorrel]

@dusty bluff i prompted it without writing down the dialogue which is in korean language and it still generates

[dusty bluff]

Damn bro , i will try lol

[reef sorrel]

If anything, it’s just the Flow site that doesn’t handle non-English prompts yet (they're working on it, no worries). Veo 3 in Gemini and Google Vids already support non-English prompts just fine (i think)

[dusty bluff]

Unfortunately im in Australia and we still using veo 2 in gemini 😦

We dont have Veo 3 yet

[reef sorrel]

Have you tried Google Vids

https://docs.google.com/videos/

[dusty bluff]

I think i can but only on computer it says

If i use vpn for Us region do u think i will be able to get veo 3 Gemini

[heady shadow]

you did it in Flow or Gemini?

[drowsy canyon]

Where did you write those instructions at? Chatgpt or gemini?

[odd birch]

Thread

[slate minnow]

Movimiento

[timber inlet]

i am trying to make more slices can any one guide me a little as I am unable to make more slices

[zinc spear]

i tried so many times i could only make one slice 🤣 how to create two slices?

[zinc spear]

what tool are you using to create this prompt?

[dim gull]

Evidently I can't tell you or ill get a warning for advertising. 😄

[prisma zinc]

@dim gull @dim gull The videos are only 8 sec it's limited time it can only probably fit about two slices

[prisma zinc]

Dumb ?

[wind panther]

How is requesting a single language to be used racist? What does language have to do with "race" Race is an artificial construct best defined as Race - who ever, where ever, and when ever you want it to be

[modern totem]

whats the prompt for this ?

[ivory magnet]

FYI-I tested the creating the prompt in both ChatGPT vs Gemini the ChatGPT came out much better

[cinder sluice]

Love this, How did your create this?

[zinc spear]

oh.. okay hahahahah sorry sorry sorry

[zinc spear]

love it! how did you prompt two slices? i tried on my prompt to command one slice and then another instruction for that but it could only accept one.

i used this on my channel. i generated a lot of glass slicing asmr. i couldn't make it to two slices 🤣

[deft thistle]

my personal expierience is add prompt before anything. it seems to understand our needs better. seems to act different if it cant tell if its a script or promt

[timber inlet]

I will share the prompt once i will be from Desktop

[zinc spear]

thanks 😍 please tag me so i'll be notified.

[north jacinth]

https://www.youtube.com/watch?v=6ha6GJwGWsE&ab_channel=AIFutureSociety

Street interview selfie style vlog with Bigfoot. I spent a lot of time and credits just to get the right prompt structure. Most of the dialogue is random but related to the setting. The difficult part is that there are recurring characters that appear in almost all of the video generations even the prompt didn’t specify for a particular character. Or Bigfoot will same the intro line for some reason even without direct input for dialogue.

[timber inlet]

@zinc spear

Each slow, deliberate cut produces a crisp “ting”—the sound of glass shattering—followed by the soft clinking of tiny shards and the deep, resonant scrape of steel against wood. These sounds layer together to create a rich ASMR soundscape in an otherwise silent room. The overall pacing is slow and soothing, crafting a tranquil ASMR experience.```

[jaunty saddle]

Highly realistic ultra 8K ASMR video of a human hand slicing a hyper-detailed, glass sculpture of a (lemon) on a wooden cutting board. The object resembles glossy blown glass — reflective, slightly tinted, and sparkling with internal highlights. The camera is in close-up with shallow depth of field, capturing cinematic lighting7AKand detailed glass textures. The knife moves smoothly and deliberately, making three separate satisfying cuts - one after another - creating 3 to 4 clean, evenly spaced slices. Each stroke feels precise and crisp, with realistic slicing sound and natura • o motion. The sliced pieces gently separate, glimmering like cut crystal. Within the slices you see crystal glass like prism in the lemon slice. background distractions, same iconic camera angle as professional food ASMR videos.

[jaunty saddle]

Highly realistic ultra 8K ASMR video of a human hand slicing a hyper-detailed, glass sculpture of a (lemon) on a wooden cutting board. The object resembles glossy blown glass — reflective, slightly tinted, and sparkling with internal highlights. The camera is in close-up with shallow depth of field, capturing cinematic lighting7AKand detailed glass textures. The knife moves smoothly and deliberately, making three separate satisfying cuts - one after another - creating 3 to 4 clean, evenly spaced slices. Each stroke feels precise and crisp, with realistic slicing sound and natura • o motion. The sliced pieces gently separate, glimmering like cut crystal. Within the slices you see crystal glass like prism in the lemon slice. background distractions, same iconic camera angle as professional food ASMR videos.

[sage cipher]

Yes

[sage cipher]

You have to use prompts otherwise how do Veo know what language you need the to speak?

[sage cipher]

One thing is that we use English to comunicate with each other and a different one is to avoid that a person post a video where the characters speak in other language. The prompts are still written in English as well as the comments.

You have to use your google account and create a Google AI subscription. After that go to flow ai and you can start using flow.

[surreal torrent]

The VEO 3 frame to video starts out great but then looks terrible by the end. I have come up with a solution. Can you please provide this fix to the google engineers and see if they can figure out what is happening during Frame to Video to cause this problem and resolve it so we don't have to do this fix? https://www.youtube.com/watch?v=ldTI3zRf2PM

[reef sorrel]

'remixed' your prompt kinda

In an extreme close-up, macro videography style shot with a very shallow depth of field focused on texture, a human hand gently and deliberately uses a sharp knife to slice a sparkling, hyper-detailed glass sculpture of a lemon on a simple wooden cutting board. The glass object is exquisitely glossy and reflective, with stunning internal highlights and beautiful prism-like facets visible in the sliced pieces, resembling cut crystal. The scene is bathed in soft, calming lighting, creating a soothing ASMR atmosphere. The knife makes three smooth, distinct, and precise cuts, accompanied by realistic, tingle-inducing crisp slicing sounds. Clean, evenly spaced slices separate softly from the main body, a gentle sound accompanying their movement, each piece glimmering under the light. The camera maintains a professional food ASMR angle throughout, focusing purely on the visual and auditory ASMR triggers with no distracting background elements.

[pulsar briar]

"Only English prompts are supported at this time", does anyone know the fix?

[exotic jasper]

https://youtu.be/nvnKllABdM8

[zinc spear]

thanks mate!

[zinc spear]

thanks guys! i love this group

[next compass]

https://youtu.be/4BxXWQ9xMTs?si=fX4Nr5nsunVp1CqJ venom reminds mary jane she does not need weapons to protect herself

[worthy ether]

Can anyone recommend best way to prompt to get this kind of first person POV effect? Through eyes of the vedeo maker https://youtube.com/shorts/4oALgZh82Fc?si=YsEx-fxSo_xiAg4m

[fossil stream]

On the pockmarked lunar surface, under the serene glow of distant nebulae and the illuminating embrace of stars, hovers a mesmerizing solar storm cloud—a miniature tempest with all the ferocity of its cosmic origins, crackling with threads of electric fire and dancing particles. Suspended just above a galaxy-colored log stump, resembling the spiraling arms of our Milky Way, the solar storm cloud shimmers with a dynamic energy, each tendril pulsating with stellar radiation. Cosmic hands, ethereal and translucent, carefully grasp an electric blue galaxy colored tool, its edge aglow with a spectrum of celestial hues. As the tool descends, it meets the turbulent cloud with an ethereal ease, parting the storm in full, clean slices. Each division yields distinct segments that, despite the moon's weak gravity, descend gracefully as if weighted by an unseen celestial pull, their internal clocks of cosmic time manifesting in hypnotic slow motion. The slices reveal intricate layers of glowing plasma and energy patterns within, like illuminated veins of stardust, threading through a luminous fabric of internal structure. The plasma threads and particle streams within the exposed sections pulse like a living tapestry, rich in cosmic stories untold, exhaling the soft whispers of creation and legacy. As each piece gently settles into place, they maintain their mesmerizing celestial appearance, the luminous glow flickering with the rhythmic pulse of space-time vibrations. The spectacle is framed with the subtle interplay of deep space lighting effects, casting an ethereal glow that enhances the delicate dance of swirling energy patterns, while particles collide and merge in silent cosmic dialogues. Audio: deep space resonance, cosmic energy discharge, no music, no talking.

post your results of this one if you wanna

Beneath a sky of silent, silver constellations on the tranquil lunar surface, an obsidian platform hewn from a fallen asteroid hosts a palm-sized compact black hole, its warped event horizon shimmering with authentic gravitational lensing and faint violet Hawking-radiation halos that ripple like liquid starlight; extreme macro focus drifts slowly around its curved gravity well, every photon bent into swirling crescents as nebular light spills across crater walls and dust motes gyrate in micro-orbits. Two translucent cosmic hands—skin a tapestry of faint galaxies—emerge, steady and deliberate, guiding a plasma-purple galaxy-patterned pairing knife whose blade glows with ionized filaments and the unmistakable text “VeoSatisfy” etched in white aurora font; soft stellar radiation gleams along its edge. As the blade meets the black hole, spacetime hums, the knife encountering velvety resistance before cleaving cleanly through the singularity’s dense gravity gradient, carving a luminous rift that blossoms with dark matter filaments and cobalt particle streams. The halves part in slow, majestic arcs, obeying realistic mass: miniature gravitational wells tug at drifting regolith grains, twisting them into tiny accretion spirals before they tumble back to the gray dust. Internal strata reveal concentric rings of compressed spacetime—molten midnight blues threaded with pink neutrino traces—while energy patterns pulse like heartbeats, releasing sparkling tachyon mist that dances in low-g lunar gravity. Each separated piece remains a functional micro-black-hole slice, its edges lensing the scene and emitting soft jets of x-ray light that curve into the void, drifting outward before settling onto the asteroid platform with a muted thud of displaced lunar soil. Cosmic particles shimmer in the vacuum, arcs of magnetized plasma flicker along the cut surfaces, and the camera lingers on the knife’s radiant tip as residual gravity waves gently ripple across the dust. Audio: deep space resonance, cosmic energy discharge, no music, no talking.

[static moth]

Use flow on google labs

[north jacinth]

Can anyone help with getting the characters to speak their lines? My prompt is setup for random dialogue but only Bigfoot actually speaks most of the dialogue 😔

[dim gull]

What's the prompt you're actually using?

[north jacinth]

Create a chaotic cinematic, unfiltered live studio dating drama 1990s videography set on a stage styled like The Jerry Springer Show, with Bigfoot as the host. Bigfoot has shaggy brown fur, oversized diamond chain reading “BIG FEET,” and carries a mic. The camera is filming with 1990s style video quality, shaky, spontaneous movements, mixed with fast zooms. The visuals are crisp, vivid, and lit with flashy studio lighting. Bigfoot is loud, rapper-style host yelling dramatic intros and jumping into emotional chaos. Guests are reactive, yelling back or confessing while the audience claps, boos, and chants “BIG-FOOT!” Bigfoot’s reactions are over-the-top, mixing surprise with hype-man swagger. Bigfoot has separate dialogue from the guests. You will add ALL the dialogue for Bigfoot and the 2 guests in the video. Bigfoot Interviews ONLY 2 guests at a time. The guests being interviewed MUST have their own separate dialogue. The guests MUST complete their own separate sentences. You will create different relationship drama scenarios. Increase the dramatics, outrageousness of the guests and their relationship problems. You can choose from the following topics or add your own. OLTP – ambient live audience noise dominates: claps, gasps, booing, laughter, plus occasional stage bell sounds and hype DJ beats between scenes. The set features studio lights, jungle decor, a bold signage: The Bigfoot Show. The audience is sitting in the background seating rows.

I don’t have specific dialogue for the characters. I let google ai come up with dialogue for randomness

[molten parcel]

Hello everyone, Can someone explain me how to put the langage in french for my prompt because, French is not allowed for VEO3. Thx !

[dim gull]

Letting it come up with the dialogue is going to make it so you have zero control over who says what.

[north jacinth]

True, I’m testing on getting the right prompt for the characters to speak by themselves.

But the randomness is just so good that I can’t even come up with it

[dim gull]

You can use a prompting tool to come up with dialogue then, but what gets fed directly into Veo 3 should be concrete.

[north jacinth]

I see

[fossil stream]

Create a chaotic cinematic, unfiltered live studio dating drama with 1990s videography set on a stage styled like The Jerry Springer Show, with Bigfoot as the host. Bigfoot has shaggy brown fur, oversized diamond chain reading "BIG FEET," and carries a mic. The camera films with 1990s style video quality, shaky, spontaneous movements, mixed with fast zooms. The visuals are crisp, vivid, and lit with flashy studio lighting.

DIALOGUE STRUCTURE - THREE SEPARATE SPEAKERS:

1. BIGFOOT (HOST): Loud, rapper-style host yelling dramatic intros, jumping into emotional chaos, over-the-top reactions mixing surprise with hype-man swagger
2. GUEST 1: Reactive person with relationship drama, yelling confessions and emotional outbursts
3. GUEST 2: Second person involved in the drama, arguing back with their own emotional responses

CRITICAL REQUIREMENTS:

• Bigfoot speaks FIRST as host introducing the drama
• Guest 1 speaks SECOND with their side of the story
• Guest 2 speaks THIRD with their counter-argument or confession
• Each speaker must complete their own full sentences
• All three characters have EQUAL speaking time
• Dialogue alternates between all three speakers throughout the scene
• Each character's dialogue is clearly distinct in tone and content

SCENE DYNAMICS:

• Audience reacts with claps, gasps, booing, laughter, "BIG-FOOT!" chants
• Studio lighting flashes dramatically during emotional peaks
• Camera zooms between speakers during their dialogue
• Bigfoot mediates and hypes up the drama between guest responses
• Guests interrupt each other and argue directly
• Background features jungle decor and "The Bigfoot Show" signage

AUDIO: Ambient live audience noise dominates with claps, gasps, booing, laughter, plus occasional stage bell sounds and hype DJ beats between dialogue exchanges.

Create an outrageous relationship drama scenario with equal dialogue distribution between all three characters.

try that

im about to send it lol i think i naied it @north jacinth

[north jacinth]

Ok thank you 😊

Got the characters to speak their separate line and now I just gotta tweak it.

[fossil stream]

hows this

ahh lol see i was not sure exactly what ya were going for but nice

thats much better than the one it outputted for me

you use the quality output?

[north jacinth]

Veo fast

[north jacinth]

Veo quality is not much different from Veo fast

[fossil stream]

yeah lol thats unfortunate fr as its 5x costlier

[north jacinth]

Yea there is barely in difference in quality and the outputs are still randoms. Veo 3 fast is quicker and cheaper since I can buy more credits to spend.

[sinful spoke]

I want a video with a person standing in front of a sign, everytime i try that the sign only reads weird gliberish and no real text, is there any prompt that actually does proper text?

[sinful spoke]

Nevermind, worked...just doesnt work always 😦

[carmine basin]

Cearte an ASMR video

(The video opens with a slow pan across the vibrant, dewy leaves of the Amazon. The ambient sounds of the jungle are soothing and immersive. The camera slowly focuses on KAEL, the monkey, who has his eyes closed as if listening to the forest. He slowly opens them, looking directly into the camera lens.)

KAEL:
(Voice is a gentle, breathy whisper, very close to the microphone)
Hello… hello there. Welcome. Welcome to my little corner of the world.

(He offers a small, gentle smile. He reaches out a long, delicate finger and softly taps the microphone twice. Tsk-tsk. The sound is crisp and clear.)

KAEL:
Shhhh… listen. Can you hear that? That is the sound of home. The breath of the jungle… tch-tch-tch… the heartbeat. I wanted to share it with you today.

how the ASMR or a monkey v log video we can make here? can anyone guide me plz?

[surreal torrent]

@carmine basin let me try

[carmine basin]

https://tenor.com/buXLr.gif

[surreal torrent]

@carmine basin trying another one. one sec

[midnight rampart]

https://youtu.be/G927Akf_MME

[surreal torrent]

This is the best I've got so far, I've tried a few and the camera direction is getting confused.

Here is the prompt:

The video opens with a slow pan across vibrant, dewy leaves in the Amazon jungle. The camera then slowly focuses on a close-up of KAEL, a gentle capuchin monkey. KAEL starts with his eyes closed, as if listening to the forest. He slowly opens his eyes and looks directly into the camera lens with a small, gentle smile. He reaches out a long, delicate finger and softly taps the microphone twice. The style is hyper-realistic, with a shallow depth of field. Full-frame 16:9 aspect ratio.

Audio: The ambient sounds of the jungle are soothing and immersive. KAEL speaks in a gentle, breathy whisper very close to the microphone. He says, "Hello… hello there. Welcome. Welcome to my little corner of the world." The sound of the two microphone taps are crisp and clear. KAEL continues, whispering "Shhhh… listen. Can you hear that? That is the sound of home. The breath of the jungle…" He then makes a soft "tch-tch-tch" sound. He finishes by whispering, "...the heartbeat. I wanted to share it with you today."

I was trying to not have it do the 'fade'

[carmine basin]

dear how did you do this, this looks amazing!

can you teach me? how to make this video? what are the requirements? etc

[surreal torrent]

@carmine basin Download this document and copy and paste it into chatgpt without saying anything. And then after just describe what you want and it will create prompts for you.

You can use ChatGPT or Google Gemini , some people make Gems if you want to do that.

[carmine basin]

daer I have prompt for a whole video and I got that from gemini but I don't know how to get video shots for this whole script. Where I can use this magical tool veo 3 and how I can get videos according to my video script with the text and character actions scenes.

[carmine basin]

@surreal torrent daer I have prompt for a whole video and I got that from gemini but I don't know how to get video shots for this whole script. Where I can use this magical tool veo 3 and how I can get videos according to my video script with the text and character actions scenes.

*dear

[surreal torrent]

@carmine basin you don't know how to use flow to generate VEO 3 videos? https://www.youtube.com/watch?v=OaU7vYmw0TU

Also watch these:

https://www.youtube.com/watch?v=fvV95J0LiOE&t=854s

https://www.youtube.com/watch?v=1lktT4dVAT4&t=7s

[carmine basin]

Did you paid subscription or can we do this with free account?

[surreal torrent]

@carmine basin VEO 3 is one of the most complex and sophisticated things that requires an incredible amount of processing power, data mining, power and electricity to consume to make it happen. Imagine, it's generating VIDEO out of NOTHING. So yah, it costs money. If you're playing with this in any capacity you need the google one ultra subscription which is like $250 a month ($129 first 3 months)