flow-tips-n-tricks・✏️

[severe aspen]

first

[alpine sorrel]

Second post. First request.

"The people rejoice as the prompts channel emerges from the spiralling void in the sky."

[unkempt swallow]

Samurai slash :p

[grand spade]

Lol I know most will request for me. I might as well be a Veo sponsor or something. Change my username color to purple gold

[clear valley]

https://tenor.com/view/its-been84years-titanic-rose-dawson-gif-5372593

[normal coyote]

You're in the heart, always.

[grand spade]

@unkempt swallow in the heart of the sea. .good movie too

Didn't mean to tag him too lol

[reef sorrel]

video prompt : Camera focuses on the crowded street market scene, then zooms out as a massive, towering tsunami wave forms in the distance. The wave rapidly approaches, casting shadows over the buildings and market stalls. The water crashes down with force, flooding the street while people run in panic. Lighting dims slightly as the wave blocks out the sun, with realistic water reflections on the market stalls.

[alpine sorrel]

If the contributors here get it, I'll join you.

[alpine sorrel]

Pikmin hurriedly drag gold bars and bags of cash during a bank heist while Captain Olimar throws the tellers and the national deficit into the emptied vault.

And the Imagefx output

[hexed berry]

“A cozy pixel art of night scene in the arctic” can anyone try this

[supple lynx]

Hello could you please try this prompts for me?

A tense Latin American scene on a city street. Several members of a police tactical unit, wearing b. lk uniforms. and bulleproof vests labeled "CSP" in whyte, stand guard in front of van with flashing red and blue lights on its roof. The officers are positioned on a crosswalk, their backs facing the camera, creating a sense of anticipation. Cars are passing. Green foliage and trees line the street in the background, adding to the atmosphere of urban tension. The overall mood is serious and watchful"

[supple lynx]

Please run my prompts too. Thanks

[alpine sorrel]

@hazy jay coming through, awesome, thanks so much

[ebon sigil]

@grand spade thanks for the prompt it was very funny

[raw talon]

A long exposure time-lapse video, capturing the motion blur of the tiny, diverse 'festi-fam' figurines, each with unique tie-dye outfits, as they dance and celebrate. The miniature city inside the mason jar is entirely dedicated to a tie-dye music festival. Buildings and pathways swirl with vibrant rainbow spiral tie-dye patterns. Multiple miniature stages dot the landscape: one shaped like a giant peace sign, another a geodesic dome covered in tie-dye fabric. A central monument, a swirling tower of tie-dyed glass, reflects the festival lights. The scene is lit by a mix of colorful stage lights and the soft glow of the mason jar, creating a dreamlike atmosphere. The glass of the mason jar distorts the miniature city slightly, like a snow globe, adding to the whimsical and majestic feel

Alternate:

Close-up on a miniature tie-dye festival scene inside a jar. Macro lens, focus on the intricate patterns. Pulsing lights, a giant peace sign stage gleaming with kaleidoscopic patterns. Right beside it, a miniature Ferris wheel spins slowly, its colorful lights reflecting in the glass. Sense the joy and childlike wonder. Time is of the essence, capture the lapse of light as the figures dance and the music swells. Scale the heights of the Ferris wheel, then peer down at the tiny, bustling world. Watch as the colors shift and the textures morph. Feel the energy of the crowd. Subtle camera shake adds a touch of realism. Evolving textures. Bokeh background. Ring in the good times with the sounds of laughter and music. Raise the energy with a quick zoom in on a specific stage, perhaps one where tiny musicians play their hearts out. Lead the viewer on a journey through this miniature wonderland

[supple lynx]

Isn't anybody going to help with my prompts?

[clear valley]

patience is key

[supple lynx]

Just a reminder. Some prompt dropped earlier have already been generated while mine is still there.

[clear valley]

ah i see

[supple lynx]

Ion know why they are being super selective.

[proper wasp]

Prompt: Low quality dashcam footage of driving on a empty road in a dense forest at night

New prompt: Grainy, low-resolution dashcam footage capturing a solitary drive along an empty, winding road cutting through a dense, shadowy forest at night. The video emulates vintage analog recording with natural camera shakes, subtle lens flares, and muted dashboard lights, all contributing to an eerie, isolated atmosphere under a dim, moonlit sky

[supple lynx]

Check your DM. I've generated it for with another tool. Can't post it here cuz this is strictly for veo 2.

[vestal salmon]

I want prompt to generate vedio inemichant

[normal coyote]

https://chatgpt.com/g/g-67796a981ac4819189a7cfa435c96d6f-veo2-prompter

[ebon sigil]

A cinematic, slow-motion close-up captures an elderly man with silver hair and a neatly trimmed beard, wearing a faded flannel shirt and suspenders, meticulously rubbing a deodorant stick across the hood of his vintage blue sedan. His expression is one of quiet determination, completely absorbed in his unconventional cleaning method. Sunlight filters through the trees, casting a warm, golden glow on the car’s polished surface. The camera tracks the smooth, deliberate motions of his hand, emphasizing the contrast between the waxy white streaks and the gleaming paint. The moment is surreal yet oddly satisfying.

[stark sphinx]

veo 2 blocks these prompts

[normal coyote]

What do you mean? It’s a prompt generator.

[stark sphinx]

its seems to not know the google politics

[normal coyote]

It’s safe to generate with it.

[stark sphinx]

i didnt say that, im saying that this generator uses words that veo 2 blocks

[alpine sorrel]

That's why you imagen test it

wide-angle shot captures a surreal cult ceremony in a moonlit forest clearing, blending eerie reverence with whimsical charm. A group of anthropomorphic animals, dressed in patchwork robes, encircle a tiny lamb adorned with a delicate floral crown. The lighting is soft and diffused, casting a dreamy glow over the misty ground. Fireflies flicker like floating embers as the animals sway in unison, their movements both ritualistic and oddly endearing. Ornate banners flutter gently in the night breeze. The atmosphere is a peculiar mix of innocence and the uncanny, evoking the storybook wonder of Babe: Pig in the City with an unsettling undercurrent of devotion.

[crimson thistle]

#flow-tips-n-tricks・✏️ stick figure doing jumping jacks

[humble scaffold]

If possible, this image and then "Chasing after a mail truck as bags of mail fly out the back of the van and letters fly out the window" or "A dynamic tracking shot captures the surreal hybrid creature—a pug with a rhino horn—bounding after a speeding mail truck. The camera follows the chaotic scene as bags of mail tumble from the open back of the van, bursting on impact and scattering envelopes into the air. The hybrid’s stubby legs pump furiously as it charges forward, its horn glinting in the sunlight. Letters whirl through the air in slow motion, creating a flurry of paper against the motion-blurred background of a suburban street. Diffused natural light adds a comedic, cinematic tone to the chaos."

If you cannot upload images, this prompt for the image:
"An extreme close-up shot of a bizarre, fantastical hybrid creature—its body a plump, cream-colored pug, but with a weathered, earthy rhino horn jutting prominently from its forehead. The camera captures the playful yet unsettling detail of its wide grin and expressive dark eyes, creating a contrast between adorable and uncanny. The textured gray surface beneath it adds a raw, grounded aesthetic. Soft, diffused lighting highlights the horn’s rugged texture and the pug’s smooth fur, lending a surreal, cinematic quality to the scene, evoking both humor and intrigue."

Man I love Imagen 3's funny abilities, I wish I could reproduce that image, but I made it with Gemini, and Gemini is my voice assistant so that chat is very very buried.

[calm kiln]

@hazy jay A busy road intersection in Isla Verde is blocked by police officers wearing grey uniforms with bulletproof vests. They have set up bright yellow restriction barriers, stopping a long line of civilian cars from proceeding. The drivers, some frustrated and others curious, wait in their vehicles. Police officers stand in formation, some directing traffic while others speak into their radios, maintaining order. The flashing red and blue lights of police vehicles illuminate the scene, reflecting off the windshields of the blocked cars. In the background, Isla Verde’s modern skyline and palm-lined streets stretch into the distance. The motorcade that prompted the roadblock is not visible, but the tense anticipation lingers in the air.

[hazy jay]

think we had something similar the other day…

[calm kiln]

That's very very good.

[calm kiln]

Yes, maybe.

[ebon sigil]

@grand spade this might be funny xd: A stick figure dashes across a chaotic landscape, avoiding falling pieces. Bright flashes explode behind him, slowed for effect. He vaults over obstacles, then unexpectedly stumbles! Everything slows as he falls, landing hard. A super close-up reveals him holding a tube of lotion, presenting it with a determined look. Text: 'Hand Cream - for life's toughest moments.' Dust settles as the scene fades.

[ebon sigil]

Thanks that crazy at 2d, can I modify the prompt? cause I kinda wanted it to be in 3d here: A cinematic movie shot of a 3d rendered stick figure dashes across a chaotic landscape, avoiding falling pieces. Bright flashes explode behind him, slowed for effect. He vaults over obstacles, then unexpectedly stumbles! Everything slows as he falls, landing hard. A super close-up reveals him holding a tube of lotion, presenting it with a determined look. Text: 'Hand Cream - for life's toughest moments.' Dust settles as the scene fades.

[ebon sigil]

yeah thats fire man thanks!

[alpine sorrel]

Prompt:
"Create a dynamic, high-energy video scene set at a Prodigy concert in the late 90s. A mixed group of 80s punk rockers and 90s hip-hop enthusiasts, dressed in their iconic styles, rush down a flight of stairs with rebellious energy. The punks have spiked hair, leather jackets, and ripped jeans, while the hip-hoppers wear baggy pants, oversized hoodies, and snapback hats. They push open double doors with force, disappearing into a massive, pulsating crowd of concert-goers. The crowd is illuminated by strobe lights and lasers, with the Prodigy's intense electronic music blasting in the background. Four overwhelmed security guards in uniform stand by, watching helplessly as the group merges into the chaos. The scene is gritty, fast-paced, and captures the raw energy of the era."

[alpine sorrel]

---

Ok so I'm going to overwhelm it a bit with options. This is intentional. Gabriel was experiment in the general videofx chat if you want context for this.

Some word is banned in the discord.

The prompt is here:
https://pastebin.com/g1iQPqXf

[dapper thistle]

Crazy good!

[modern geode]

The first one is stunning!

[hexed berry]

Can anyone try this “A pixel art, nighttime in the southern streets with a tavern on a windy night”

[spark atlas]

Here are your four options.

[barren mist]

Create a video ad for Google AI in an abstract visual style using pointillism animation. Animate with distinct dots of vibrant color to form clean, vector-like lines and smooth motion, with a sense of instantaneousness and flashing visuals. Show quick, abstract scenes of Google AI being helpful in everyday situations, like using Maps, Translate, and Search, presented in a fast, flashing manner. The ad should convey a sense of "Helpful Intelligence" and universal access to information, with visuals that flash and change quickly. Conclude with the Google AI tagline: "Making AI helpful for everyone.

[hexed berry]

Thanks

[hexed berry]

“A pixel art, nighttime in a grassland with stars in the sky, wolf howling in the distance”

[alpine sorrel]

I've wanted to see if the sources of prompted sound are added. Excited to see if a wolf pops up.

[grand spade]

I have a prompt for someone lol...

[alpine sorrel]

What's that ?

[modern geode]

The man who does requests wants a request 🤯🤯?

[grand spade]

"A man looks at his feet, low angle shot, to notice a band of fire ants walking over them"

[modern geode]

It has to be necessarily a video made with veo2, right?

[grand spade]

@modern geode yeah but I was just kidding. You can try the prompt elsewhere if you choose to

[alpine sorrel]

I have insane reactions to red ant bites . Why are you calling nightmares

[proper tree]

A description of a cat pursuing a small mouse, illustrating the natural behavior of hunting.

[molten parcel]

yall someone should do like a glass shifting through different shapes on a white background

by glass i mean drinking glass

[modern geode]

A mysterious planet floating in the vast abyss of space, surrounded by swirling cosmic clouds in shades of light blue and light purple. Suddenly, the planet explodes in a burst of glowing energy, sending shockwaves through the colorful nebula. As the explosion unfolds, four letters—P, E, I, S—come flying in from different directions, spinning and glowing with a soft neon effect. The letters collide and seamlessly join together at the center of the screen, forming the word 'peis' in bold, futuristic typography, completely covering the screen.

[alpine sorrel]

Is of be and all really

That's the prompt.

Whisk gallery has the image and musicfx output of it, animation #44.

[modern geode]

What?

[alpine sorrel]

Matches the ephemeral uneasiness of the musicfx output.

I've never known why it likes back facing the camera shots.

[alpine sorrel]

"Snail Eating Ritz Cracker"
https://youtu.be/UoYMK1olf4s?si=wABWKqmwNAPuGDMi

[reef bobcat]

An African-American man smiling and talking, looking into the camera

[modern geode]

A bright light blue Volkswagen beetle jumps off a ramp and through a giant block of hay, there is a slow motion close-up of the impact

[alpine sorrel]

Ghostbusters ecto1-1 classic cadillac races a time travelling delorean. The delorean activates time travrel and disappears. [ Don't forget the intricate lighting, massive landscape, awe-inspiring, beautiful, great composition, golden ratio, very intricate. ]

[modern geode]

A breakdancer in modern urban clothing performs an impressive headspin, spinning rapidly on his head with arms extended for balance. The scene takes place on a neon-lit dance floor, surrounded by cheering spectators. The reflective floor and graffiti-covered walls enhance the vibrant atmosphere.

[ebon sigil]

@grand spade or Anyone can you try this? : photorealistic cinematic first person view of someone driving down a steep hill and suddenly stopping due to a large text appearing that says: Kor & Minecraft Productions, gloomy day, photorealistic and cinematic

[ebon sigil]

You have compleated your side quest

[alpine sorrel]

Russian, Hungarian and Georgian folk dances. Moscow Operetta Theatre (1967)

[alpine sorrel]

The flat Earth inside of the hollow Earth.

[alpine sorrel]

That'd be a curious prompt.

[alpine sorrel]

Inspired by @formal jacinth

POV drone fisheye view, fast zoom through a canyon carved by giant sandworms on a desolate planet, strange rock formations, dust devils swirling, distant alien sun setting, feeling of isolation and exploration."

Drone fisheye footage, low-altitude flight, accelerating through a sprawling sci-fi metropolis with chrome arcologies, flying vehicles zipping past, holographic advertisements flickering, alien sky with two moons, sense of vertigo and wonder."

Drone fisheye perspective, dynamic zoom through a vibrant, underwater sci-fi city with glowing coral buildings, aquatic creatures swimming by, sunlight filtering through the water, sense of immersion and discovery."

Kind video maker, choose whichever world appeals most to you.

Drone fisheye view, zooming through a vast, alien megacity at dawn Sci-fi world, epic zoom, neon lbillboards, bioluminescent flora., and Isense of speed, awe-inspiring scale

ImageFX provided to 1. Demonstrate Videofx is likely to accept the prompt
2. Break up the monotony of text .
3. Show I'm willing to put in as much work as I can and as a gesture of respect for our generous volunteers.

[ebon sigil]

@grand spade or anyone please try this might be epic: --------- In a dark and gloomy DIY forest nestled in the heart of Central Europe, an older man with a weathered face and a leather jacket tears through the misty trails on his vintage 1980s motorcycle, its engine roaring like a beast unleashed. The scene is captured in the grainy, nostalgic texture of 35mm film, with cinematic shadows dancing across the handmade claymation landscape. Suddenly, hordes of grotesque clay zombies emerge from the twisted trees, their jagged limbs and hollow eyes lunging toward him. With a steely gaze, the man pulls out a revolver, firing at the undead with precision, each shot sending chunks of clay flying in slow motion. The tension is palpable, the atmosphere thick with dread, as the man races against time, the forest, and the relentless clay horrors, all while the haunting echoes of an 80s synth score pulse in the background, blending action, nostalgia, and eerie beauty into a surreal claymation masterpiece.

[alpine sorrel]

@grand spade can you try one of the drone prompts I posted above?

[grand spade]

Good stuff. I've been doing everything via mobile

[alpine sorrel]

Flying shots, whether Minecraft or anything else, all maintain that same speed and exact chase distance

Phone, tablet, two screen PC, laptop. That's too many screens but here I am.

[grand spade]

Ok cool

[alpine sorrel]

Fascinating. Concise with better adverbs makes a big difference

Have you tried going the opposite direction with just a phrase like "act of whimsy" ?

Just putting in like three words to see where it takes them.

And then building from there.

That's how I got the juggling bear I put into Imagefx yesterday and then the weird art that followed built from that

Interesting mirror to what people consider Whimsical.

That's an afternoon of stop motion work.

Gnome has a process.

Nope

You can screen shot it and explain what it's doing in the video.

Google Cloud has video intelligence ?

[hexed berry]

Prompt “A drone footage of a strange alien world filled with iridescent crystals, creatures made out of obsidian, comets in the night sky”

[severe aspen]

VideoFX UXR opportunity!

if you have access to VideoFX, we want to hear from you! check out this opportunity for a paid interview with the Labs squad!

#flow-announcements・📣 message

[alpine sorrel]

Prompt: a zeppelin labelled Google rains coins onto a giant 50 foot tall man in a business suit with a movie camera for a head. Ant-sized peasants with paintbrush heads scurry at the feet of the camera-man.

[hexed berry]

@rough wing Prompt “A drone footage of a strange alien world filled with iridescent crystals, creatures made out of obsidian, comets in the night sky”

[knotty tendon]

@SK prompt "an ultra realistic pov video, of running in the night on a dark forest, bodycam style, realistic motion"

[hexed berry]

Thanks

Can you try this too “A giant honeybee leaking honey is flying through the streets of New York filled with honey”

[stoic oak]

@rough wing could you use this prompt? "Ultra realistic pov style video, of running on the streets, people are running, and cars are crashing themselves, apocalypse style"

[hexed berry]

Thanks

Wow

[stoic oak]

Yoooo veo 2 is crazy!

[hexed berry]

Yeah insane imagine an entire movie like that

[hexed berry]

Can you try a giant godzilla?

[alpine sorrel]

Kaiju

[stoic oak]

Sure it is

[stoic oak]

They look good, but they arent pov

[hexed berry]

Wow these are crazy good

[stoic oak]

THATS what im talking about

Incredible!

[hexed berry]

@rough wing try this too “A giant honeybee leaking honey is flying through the streets of New York, The streets are filled with honey””

[stoic oak]

Ok i will let you rest

Nah dont do it,

[hexed berry]

This one’s good

[stoic oak]

@rough wing hello, can you make "ultra realistic found footage video, showing on pov style, how an oven opens on an apartament, and water comes out from the oven and floods the room" this one is complicated, so i understant if it messes up

Yeah you're right

Exactly!

Ok

[stoic oak]

Dammit, its because the prompt is difficult or what?

Maybe i can think something different

[stoic oak]

Oh

Damn

[alpine sorrel]

I keep a separate default Chromium install for when I need a site with nothing but browser loading. Which happens more than I'd like.

No idea.

[stoic oak]

I have added audio to this video, it sounds really scary lol

I literally use another ai to add audio, its called mmaudio

some fix? or is the error still there? Just saying

Yeah also

[stoic oak]

This one is the strange prompt, basically an arm reaches out the oven and water comes out of it, and floods the room

[wooden current]

I workshopped a short story about my cats going to Las Vegas to test the text to voice when converting it to a script and then built out some Image FX shots trying to capture the absurdity. Anyway I can get some video footage like a movie preview if I provide a few images and prompt / audio to add?

[alpine sorrel]

Prompt: roving dolly shot of a Swarm of battle-worn robots dismantling themselves, fusing parts together to form a towering, colossal megarobot. Camera pulls back dramatically. Sparks fly as metal limbs snap into place to form a massive limb, glowing cores merging in a dramatic transformation. A million smaller dots light up, tracing piping lines along the megarobot's swarm based body. "

[stoic oak]

How strange how this is happening to you right now, this ever happened to you?

[stoic oak]

For some reason this message was sent around 6 am.

[stoic oak]

Oh ok

Aight

[stoic oak]

I've been trying other ai, so you are not the only one experiencing this errors

This happened to you isnt it?

[wooden current]

If your able to fix the issue, Id love to get some footage of a pho short of an animated teaser trailer video of a test script about my forced adoption of two siamese cats that plan their escape from my house jumping head or whiskers first into a blue electric doge viper complete kitted out for an adventure in Las Vegas. I

[wooden current]

I'm working on a 15-20 minute animated script in my spare time. I've voiced the first 10 minutes and think having an AI-generated visual reference could help me sync the voices better. Eventually, I’d love to render the full animation if I find a cost-effective video prompting tool. So far, this project has been a fun learning experience, even with all the clunky bits, and I feel like I’ve created something really entertaining as a one-person team.

[visual thorn]

I just made an animated movie with VEO2, and VEO2 does animated movies super well too!

https://x.com/Eric520CC/status/1892431311656407484

Metaverse | Eric (@Eric520CC) on X

🎬Trailer: Forest Adventure
Videos: VEO2
Voices: Hailuo
Edited in: CapCut
All clips are generated by VEO2 Text To Video.
Let's start a forest adventure together!
#Veo2 #hailuo_ai @Hailuo_AI @GoogleDeepMind @labsdotgoogle

X · Feb 20, 2025, 4:29 AM

[hexed berry]

@rough wing can you try this Prompt " A spaceship in outerspace flying near a supermassive blackhole"

[alpine sorrel]

"the reality simulator breaks down. Coffee shop starts glitching and then everything melts to goo."

[hexed berry]

Pretty good

[stoic oak]

@rough wing could you try this prompt? "A realistic recorded video displayed on a television screen, showing a breaking news broadcast. The news channel is reporting on a global invasion of dark, shadowy creatures emerging from the unknown. The screen displays chaotic footage of cities being overtaken, people running in terror, and military forces attempting to fight back. The broadcast includes a red "LIVE" indicator, urgent headlines, and a distressed news anchor struggling to maintain composure. The creatures appear as eerie, shifting silhouettes with glowing eyes, spreading rapidly across the world. The scene is dimly lit, with the television being the primary light source in a dark room, creating an ominous and suspenseful atmosphere."

[visual thorn]

Debugging through text descriptions, then doing a lot of clips and then picking some of the recognizable consistency ones to clip

My VEO2 won't generate videos after the upgrade, is yours still working?

[chilly fox]

Has anyone got the hang of prompting specific camera moves? Often times it just ignores what i prompt it to do, for example, camera orbiting/circling/arcing around a subject, i haven't found a way to get it to consistently follow the directions, anyone have better luck with this?

[chilly fox]

probably 80% of my failed/wasted gens are repeat attempts to get a specific composition/camera move, adjusting the prompt to add ever increasingly detailed and hyper specific verbiage to please, pretty please, just put the camera where i want and move how i want it to, then it just does the same thing again. Veo is amazing but that's prob biggest frustration with it atm!

[hexed berry]

@rough wing can you try this prompt “A cat floating in outer space”

[quasi hatch]

“A giant, hyper-realistic brown owl with golden eyes stands in a lush, green forest, its massive form towering over a tiny human figure. The owl tilts its head slightly, then leans forward and gently pokes the man with its large, curved beak. The man, dressed in dark casual clothing, stirs, slowly gets up, and looks at the owl with amazement. As he reaches out cautiously, he places his hand on the owl’s soft feathers, gently stroking its plumage. The owl blinks slowly, enjoying the moment, its feathers ruffling slightly in the soft breeze. Sunlight filters through the dense trees, casting warm highlights on the owl’s textured feathers, while subtle dust particles float in the air, adding a magical touch. The camera captures smooth, cinematic close-ups of the owl’s expressive face and the man’s cautious yet amazed reaction, before pulling back to show their full interaction in the serene forest setting.”

[wooden current]

I just made an animated movie with VEO2

[severe aspen]

please see our latest VideoFX announcement, and mark your calendars: #flow-announcements・📣 message

[hexed berry]

Can anyone try this. It’s not letting me do it in freepik

“A 4k HD live action cinematic movie, Ryuk and Light Yagami on a park, overcast day, dark fantasy theme”

[icy pasture]

YOOO cant wait!!

[icy pasture]

https://tenor.com/view/spongebob-squarepants-begging-pretty-please-beg-on-your-knees-pray-for-mercy-gif-10678931350545522063

[hexed berry]

Who is omar little lol

[honest plaza]

I want to see something along the lines of “Charizard from Pokemon with heavily defined musculature, wings, upper body shot. Minimalist 2D digital art drawing style, flat colors, minimal shading” animated with Veo 2

Ahem

[grand spade]

I tried your prompt. I have to rerun it didn't animate fully

[honest plaza]

Oh lol

[grand spade]

Look lol

[honest plaza]

Dang 😭

Just like old Luma video

Why is that blocked 😭

[grand spade]

Idk

[honest plaza]

Maybe they don’t like people talking about it???

Also I don’t think the image made Charizard muscular

[grand spade]

Nah probably not trained enough on Pokémon

[honest plaza]

But it still has Pokemon footage like legends arceus

And also TOTK

[grand spade]

Hmm I will try other characters

[honest plaza]

Ok

[urban wedge]

Is there a way to negative prompt?

[chilly fox]

no

[visual thorn]

https://x.com/Eric520CC/status/1894208871821935030

Metaverse | Eric (@Eric520CC) on X

Advertisement: The Bright Star
Videos: Veo2
Sounds：Pixverse V4
Edited in: CapCut
All clips are generated by Veo2 Text To Video.
#Veo2 #PixVerse @PixVerse_ @labsdotgoogle @GoogleDeepMind

X · Feb 25, 2025, 2:12 AM

[alpine sorrel]

You didn't need to spam it across all the channels

[calm kiln]

Please someone here help me run this prompt. Thanks
"A Hausa woman dressed in an elegant blue Hausa lace outfit with intricate embroidery and a matching flowing hijab is performing a daring wheelie on a sleek blue power bike. The scene takes place on a lively road in Kano, with the sun casting a golden glow on the surroundings. Crowds of people, men and women in traditional attire, line both sides of the road, cheering and clapping in excitement as she skillfully lifts the front wheel of the bike into the air. The bike's engine roars as she maintains perfect balance, her hijab flowing dramatically behind her. Some onlookers capture the moment on their phones, while others wave and shout in admiration. The background features bustling Kano architecture, market stalls, and banners in Hausa script, adding to the vibrant energy of the scene. The video captures her confidence, the thrill of the stunt, and the celebration of empowerment in a dynamic, cinematic style."

[wooden canopy]

yes... wait 2 min till its done

[calm kiln]

Thanks.

[wooden canopy]

🫡

[alpine sorrel]

Asked how it's doing and it, uh, it broke the screen.

[candid rapids]

Hey everyone, I've been using Veo 2 all day today. I noticed I had some issues generating just one prompt, so I wanted to share it here in case anyone wants to try it to see if it works on their end.

Film quality, professional quality, rich details. The video begins to show the surface of a pond, and the camera slowly zooms in to a close-up. The water surface begins to bubble, and then a blonde woman is seen coming out of the lotus pond soaked all over, showing the subtle changes in her facial expression, creating a dreamy atmosphere.

[alpine sorrel]

@keen urchin can i ask your dancing prompt phrases or words

[keen urchin]

"Vinyl record grooves magnified to show tiny dancers moving within them, their choreography creating the music as the needle passes. Macro photography with warm analog colors."

[calm kiln]

@waxen tree "cinematic, a sweeping aerial view of a massive, modern government building in a fictional African country. The architecture is sleek yet infused with traditional African design patterns, showcasing grandeur and cultural heritage. Below, a vast crowd of thousands gathers before a grand stage, draped in the national colors of Angabolia. The atmosphere is charged—some people cheer with anticipation, while others whisper with uncertainty. On stage, a row of high profile people sit, dressed in formal suits and traditional garments, their faces serious. Armed security personnel stand strategically, scanning the crowd. The golden glow of the late afternoon sun casts dramatic shadows, adding depth to the scene.

[waxen tree]

got u

[calm kiln]

I'll drop the rest

A sleek dark presidential limousine glides smoothly along a red-carpeted path leading to the grand inauguration stage. The vehicle's polished exterior reflects the golden hues of the setting sun. Armed security personnel flank the sides, their expressions tense and alert. The crowd watches in anticipation—some cheering, others murmuring with uncertainty.. The limousine comes to a slow halt. " Cinematic, movie

[calm kiln]

In slow motion, the rear door of the sleek dark limousine swings open, revealing an Africa woman. She steps out with quiet, her tall frame exuding confidence. She wears a dark, impeccably tailored suit. Her hair plaited, catch the sunlight as she lifts her chin, scanning the massive crowd. The camera lingers on her strong. She waves at the crowd. The atmosphere is thick with tension, anticipation crackling in the air."

[waxen tree]

A sleek dark presidential limousine glides smoothly along a red-carpeted path leading to the grand inauguration stage. The vehicle's polished exterior reflects the golden hues of the setting sun. Armed security personnel flank the sides, their expressions tense and alert. The crowd watches in anticipation—some cheering, others murmuring with uncertainty.. The limousine comes to a slow halt. " Cinematic, movie

[calm kiln]

"An African woman strides forward with slow, deliberate steps on the red carpet leading to the grand inauguration stage. The golden sunlight casts long shadows behind her, accentuating her commanding presence. The crowd watches in silence, some holding their breath. Her suit, adorned with subtle African textile patterns, moves elegantly with each step. Armed security personnel flank her at a distance, their eyes scanning the surroundings. On stage, the Prime Minister and military officials observe her closely, their expressions unreadable. The rhythmic beat of ceremonial drums echoes in the background, heightening the gravity of the moment.

[waxen tree]

this one didnt generate, it got flagged

[calm kiln]

I think we need to include "African" here. This isn't Africa in the video. The people too are white.

[clear valley]

training data im guessing

[calm kiln]

Ooh,

[calm kiln]

Yeah but it may work if I include "African"

[clear valley]

try using dark woman

[waxen tree]

yes it might

[waxen tree]

this one works

[calm kiln]

Ok

"The tall, commanding African woman with short locs streaked with silver stands behind the grand wooden podium, gripping it firmly. She wears a dark, impeccably tailored suit adorned with intricate African textile patterns, blending tradition with authority. The golden glow of the setting sun casts a warm hue over her face, highlighting her piercing gaze as she looks over the massive crowd. The atmosphere is thick with tension—some in the audience hold their breath, others exchange uncertain glances. On stage, high-ranking officials watch intently,.

The camera captures the crowd's reaction—some erupt in cheers, others remain motionless, eyes filled with skepticism. The national flags behind her ripple in the wind as history unfolds in real time."

[calm kiln]

Can I rewrite this prompt to get African output?

[waxen tree]

yeah do it

[calm kiln]

A sleek dark presidential limousine glides smoothly along a red-carpeted path leading to the grand inauguration stage in an African setting. The vehicle's polished exterior reflects the golden hues of the setting sun. Armed security personnel flank the sides, their expressions tense and alert. The crowd watches in anticipation—some cheering, others murmuring with uncertainty.. The limousine comes to a slow halt. " Cinematic, movie

[waxen tree]

The tall, commanding African woman with short locs streaked with silver stands behind the grand wooden podium, gripping it firmly. She wears a dark, impeccably tailored suit adorned with intricate African textile patterns, blending tradition with authority. The golden glow of the setting sun casts a warm hue over her face, highlighting her piercing gaze as she looks over the massive crowd. The atmosphere is thick with tension—some in the audience hold their breath, others exchange uncertain glances. On stage, high-ranking officials watch intently,.

The camera captures the crowd's reaction—some erupt in cheers, others remain motionless, eyes filled with skepticism. The national flags behind her ripple in the wind as history unfolds in real time.

[calm kiln]

Videos Are very good but i think It's goin to be difficult due to lack of consistency. Also, the Vids don't look very cinematic but I really appreciate your effort.

[waxen tree]

yeah they def need revisions. all comes down to the structure of the prompt, but even then it will still be undeterminalistic

[calm kiln]

I think so far Veo 2 is better when creating documentaries the vids look real.

[waxen tree]

A sleek dark presidential limousine glides smoothly along a red-carpeted path leading to the grand inauguration stage in an African setting. The vehicle's polished exterior reflects the golden hues of the setting sun. Armed security personnel flank the sides, their expressions tense and alert. The crowd watches in anticipation—some cheering, others murmuring with uncertainty.. The limousine comes to a slow halt. " Cinematic, movie

[calm kiln]

Has this work?

[waxen tree]

yes it worked

[calm kiln]

Good. I appreciate so much. I'll see if I can edit the vids to make them look a bit cinematic.there are more prompts I would love to generate but I don't to be a bother

[waxen tree]

just send them anyways. even other people wont mind doing it

[calm kiln]

Ok. Thanks

[waxen tree]

lmk if u have anymore requests later. cheers

[calm kiln]

Sure. Thanks

[calm kiln]

I'm not the one writing the prompts

[calm kiln]

Thanks for the advice. However when it comes to character consistency, it's difficult. Especially when generating Africans. I've noticed this.

I'm not saying it doesn't Generate a certain group of people rather there is lack of consistency in African character unlike Latino etc.

[calm kiln]

Wow. Lots of amazing people on this server.

[pallid delta]

VideoFX seems to understand "solarpunk" directly, even a very simple prompt like "drone footage of a happy solarpunk town" seems to work. Obviously you can decribe what you want in more detail.

[dense temple]

take a screenshot and let chatGPT or any other LLM with vision describe the image for you. You can also ask it to make a detailed Text to Video prompt for it.

[candid rapids]

Yes! There's also a CustomGPT pretty well made by https://x.com/allarhaltsonen/status/1895430212029968637

Allar Haltsonen (@AllarHaltsonen) on X

Mini tutorial for @GoogleDeepMind 's VEO 2 Prompt Buddy. (Works well with other AI video tools as well).

I’ve created my VEO 2 Prompt Buddy Custom GPT to help me go beyond my own brain and prompts. It can describe your images, enhance your prompts, be creative and more.

It will

X · Feb 28, 2025, 11:05 AM

[dense temple]

Nice, thanks for sharing! Working on something similar for myself, trying to find the best cinematic approach

[alpine sorrel]

https://www.youtube.com/watch?v=JAhkhIp_7Zs Inverted request - who can recreate this?

[calm kiln]

A small, thin desert lizard with light brown scales and large yellow eyes curiously pokes at a discarded red soda can lying in the golden desert sand. The can suddenly tips forward and slips over his head, trapping him inside. Now blind, the lizard stumbles and rolls uncontrollably across the rocky terrain, kicking up dust. The shiny soda can reflects the sun as it bounces off cacti, skids over dunes, and startles other desert animals. A scrappy fennec fox and a clumsy vulture watch in amusement as the lizard desperately tries to free himself, leading to a chaotic, slapstick chase through the hot desert.

[reef sorrel]

Let me try

There was an attempt ahaha

@calm kiln

[molten parcel]

just a random idea

a slow dolly in shot of a typewriter typing itself on an empty library

[calm kiln]

Close but not close enough

[oblique current]

motion video about kindergarten

#flow-tips-n-tricks・✏️ motion video about kindergarten

[reef sorrel]

I think last time I saw someone did that one video of clown and bunch of kids in background

But yeah sam what kind of ..kindergarten

[molten parcel]

to those who doesnt have access to veo 2, i could definitely do your prompts if ya want!

[molten parcel]

missed an r, but i’ll let it slide :)

[hidden belfry]

Aww dang it it almost got the text right. Just added an extra "i" in aquired

[hidden belfry]

OH MY GOSH IT GOT THE TEXT RIGHT!!

No I was talking about the "Stone Orb Acquired" thingy

But isnt Soltzle a word from another language?

Nope nevermind I put it through translator and it did not detect one single word

ChatGPT - Soltzle — a word from the ancient tongue of Varyndrial, a language once spoken by wandering sages and elemental seekers.

gasp

Yesss

[hidden belfry]

The percentage thing is disappearing again

Im gonna try to clear my cache and maybe that will work

Nope didn't work

It works now 😄

[alpine sorrel]

@green breach Hey, let me make my earlier message even clearer. I'm looking for you to put your chat in general, and then here for prompt talk,

I said if you do that, then I'd be willing to help you.

[green breach]

I wasn't even aware we had this channel

my current prompt is "scifi movie scene, selfie footage of a british punk rock man. he is manly. beard stubble. unkempt short dark hair. piercing blue eyes. manly, virile, masculine. wearing a top hat with a screen on it. punk jacket with wires, slightly futuristic. he shows that his own head is cut open in a straight line, wires inside. skull cut open in a straight line, wires inside. scifi movie. he looks sarcastic."

had to remove the word skull or I get a literal skull

[alpine sorrel]

It's not a secret.

Oh well.

[green breach]

not only for ppl who don't have access?

I'm sure there's some magical words that make the ai understand the concept

[alpine sorrel]

Here is the pinned message in general from Georgia the mod and google employee.

[green breach]

as an english teacher, I'll consider it a challenge

mmmh. scalp may be interesting

[green breach]

perhaps the prompts that work in imagen dont work on veo

what was the prompt here?

and he takes his hat off to reveal he.... mostly has a normal forehead

on the plus side I got character consistency on my dude

[alpine sorrel]

you'll get there.

i layered some surgical terms in there

[green breach]

do tell, that's a clever idea

my next idea is to make a movie trailer. and I know the hardest part will be showing the Mad Hatter's lack of brain so I must get it done first

[alpine sorrel]

what I normally do is pick a surgery or medical procedure. ask a reliable LLM for related terms connected to that procedure. Sometimes it teaches me about medical interventions I had no idea existed.

[green breach]

yes but prompt? 😭

maybe I'll make an easier movie trailer on a different idea

[alpine sorrel]

got my rough proof of concept. looks a little medical mannequin

tried to add your lid idea and it left the skin behind. didn't specifically request it of course. you have to go additive then subtractive.

[green breach]

it's a very difficult concept for the ai

[alpine sorrel]

5% man i hate when it's brand new

it's a difficult concept to communicate to the ai. there's a distinction.

[green breach]

and why did I write "liddle" instead of "lid"? perhaps it's time to rest

[alpine sorrel]

things like that help actually

[green breach]

why?

[alpine sorrel]

So... AI models often interpret words phonetically or contextually, meaning a slight misspelling can lead to alternative associations that a strictly correct term might not evoke. For example, a prompt like "cybernetic hateter" instead of "cybernetic hatter" might push the AI toward a more sinister or enigmatic version of the character,

I didn't test those because i'm about to head out the door and disappear for a few hours at least

But you get my idea

Typos can act as soft randomness injectors, forcing the AI to infer intent rather than rigidly adhering to a predefined concept, sometimes yielding more novel, unexpected, or serendipitous results... Everything in moderation though, too much and it'll throw it all away.

haha i deleted eyes by mistake

[green breach]

Mmh I gotta try it

An idea: instead of wires, a flat lid

[ivory mortar]

"lightspeed warp space-travel, bluestreaking stars" worked for me on MJ --

"lightspeed warp space-travel blueshifting stars" was a little better but not by much

[frozen escarp]

Where should I ask for a prompt to recreate this video? https://youtu.be/0Gj1QaqDk2Q?si=Hkaep73mEZb-DKU4

I only want the collision of the car and the utility post.

[frozen escarp]

@rough wing Please try to show us an example

[alpine sorrel]

Sorry, what exactly are you trying to do? Just an over the shoulder on a guy pointing with a closed mouth?

Otter Detective working the case

Otter Detective is the serious minded cousin of the Hawaiin shirt Otter of Snail and Otter fame.

[alpine sorrel]

Well, at any rate, I'm enjoying following along with the development of your submission.

Yeah I agree, that feels sub-thread.

[alpine sorrel]

The galactic heralds

[alpine sorrel]

@rough wing so do you have any best practices ypu use for keeping consistent between prompts? Another reason I dont close tabs is I don't want to lose the seed I'm on.

Still sometimes I copy paste the same "setting the table" text and it'll jsut throw it all away.

[alpine sorrel]

well, hot streak, whatever it is. The same prompt can get on a run of good generations 10 long but I pop the same prompt into a fresh tab and stella loses her groove.

[green breach]

anyone wants to try the whole "cyberpunk dude with no brain" concept I was trying?

[alpine sorrel]

That case was ages ago. The otter detective agency is focused on the amazing collosal horse case brought forth by one Neary Far.. [get it, like biggie smalls? I'll be in my detectatorium.]

[green breach]

I haven't understood a single word, but I'll nod in agreement

[green breach]

Fine! I will, and I will not share the prompt when I manage it

[calm kiln]

Very good. Thank you

[alpine sorrel]

There should be a button where it scans my neurons to make my imagination directly and also run an IV drip of dopamine.

[muted trench]

Anime prompt
Anime still frame inspired by Kyoto Animation, realistic anime art style. Detailed and soft shading. Intricate textures, subtle gradients. Semi-realistic proportions with stylized facial features. Expressive character emotions and detailed hair rendering. Modern anime aesthetic. [Prompt] High resolution, 8k, realistic skin rendering, ambient occlusion, volumetric lighting. Subtle chromatic aberration and film grain for realism.

[alpine sorrel]

Trying out different studios is huge. I also like to look up creative staff names that don't have pop culturue recognition to see if the AI made a link.

[waxen tree]

😁

[frozen escarp]

Can someone please try to write me a prompt for a video just like this? https://youtu.be/MVIJcm8_pRo?si=kAzX1iEsIaaCzXLx

[alpine sorrel]

Have you tired Gemini, chatgpt, claude or msitral maybe? Any of the llms?

you could also screenshot it and give the images to just Whisk if talking to chatbots isn't your thing

[muted trench]

I found out the picture in some llm. Then ask how to recreate for a prompt. If using Veo2 try putting the prompt in text to image to video first to a picture.

You'll get a sense of how close it's going to be

That's what I did using Claude or Gemini

If it doesnt work show the image it gave and say why it's wrong and how rewrite the prompt or add to

[frozen escarp]

I'll try that, thanks.

[muted trench]

Welcome

[alpine sorrel]

I copy-pasting full prompts including motion controls into both boxes. I also start and end the prompt with "Seamless transition." I also reinforce character names and descriptions twice over.

I've also been experimenting with using the motion box to intentionally make a whole different shorter clip. It feels different thatn straight text to video, probably becaus eof hte linkage it tries to do.

[craggy widget]

I tried every method which would help, but can't get one full video result

[green breach]

have you noticed how some prompts always want to stubbornly use a specific style? I want to make this real-life, but EVERY time it tries to make it into bad cgi. In this particular video, it decided to make the -mutant- realistic and the -normal woman- a cgi production...

[green breach]

how do you make a morningstar with a flail chain? I can't do it properly

[green breach]

the physics here needs some adjustements......

[pallid delta]

Generally, my experience is that you are more likely to get continuous shots with text to video, and that you are more likely to run into issues with the text to image to video tool that you report using.

More generally, note that this is a early prototype--one is just going to be frustrated if you expect a finished tool that "just works" If you hate the way it cuts, then thumbs down those videos, or give more detailed feedback, or mention this if you happen to participate in surveys and interviews.

But you are just setting yourself up for frustration if you have the expectation that this is a finished and complete product. Instead it's something you are helping shape. Sure, those of in a beta a bit longer can give you tips that seem to help help us avoid sequences with lots of cuts, which might help in a limited way. But fundamentally, engineers at Google are the ones who can make improvements to this prototype, and your feedback can help guide what they will learn about what needs to change.

And, perhaps, for some people, perhaps their main use cases are "low effort memes" and having a simple prompt come up with a simple short clip with a bunch of cuts meets their needs. Which is a very different product than one built for people who want to assemble a bunch of shots into longer video.

[warped fog]

In my experience, it helps a lot to re-describe the entire scene in the motion prompt, and then describe the motion prompt. It's almost always more consistent, at least in my testing

[pallid delta]

It seems like you're genuinely

[haughty river]

Definitely hit the "dislike" button on videos with cuts guy! (at least if it was not intended). I'm sure with time and updates, if you do that, it'll happen less and less!

[haughty river]

It should be not too hard for whatever system they are using to "learn" that cuts are bad if the majority of the time it generates a video with it, it gets hit with the dislike i think. particularly when the 4 videos are very similar, the ones with cuts gets disliked, I think it definitely should be able to "understand" the pattern

Let's see how it evolves with the updates

[heavy axle]

Can we get this a try:

A dim bleek pixel art scene of an industrial complex barely visible through a dense fog. The environment has a deep, liminal atmosphere with a desolate landscape in the foreground, textured with minimal details to enhance the eerie emptiness. The industrial structures are dimly illuminated with scattered, small yellow lights, flickering in the distance. The sky is an overwhelming gradient of grays, blending seamlessly into the fog, obscuring most of the details. The pixel art style should have a moody, atmospheric aesthetic with a limited color palette dominated by empty blues, grays, and midnight abysses. Soft dithering should be used to create the illusion of depth and mist. The overall tone should feel isolated and surreal. smoothly the camera glides along a horizontal axis to the screen right revealing the endless misty landscape

[alpine sorrel]

Running it. It's a bit long so we'll see.

[heavy axle]

can you start off a base image?

@alpine sorrel thank you

[alpine sorrel]

Well we can generate an image from text and then run that. No option to upload your own. Do you have the prompt that made that ?

You could also upload it ask Gemini or one of the others to do it.

I share too much. Really need to figure out how to monetize this.

[alpine sorrel]

I should remake the otter and snail video game one with Veo now

[alpine sorrel]

So I shouldn't just be selling the dream of waluigi cheese jacuzzi water.

[muted trench]

I need help I'm trying to do a scene with only showing the massive base of this colossal tree. Like this bad paint drawing.

Extreme close-up of the colossal base of an ancient tree. The base fills the entire frame, towering above the viewer. Low angle shot, emphasizing its immense size and overwhelming presence.

That prompt didn't work and show the branches and leaves at the top and not cover the screen

[muted trench]

Thanks but instead of anime I just went with my 3D UE prompt

[frozen escarp]

Im still hoping that someone can please write me a prompt for a real life video like this : https://youtu.be/MVIJcm8_pRo?si=XcvIucwRrDWq1-ko

[muted trench]

Use gemini

Tell it what happens and what it looks like so it can give a prompt

[frozen escarp]

I've tried to use the Google Gemini app many times. That's why I had to ask here.

[humble scaffold]

Try using Gemini via the developer API console

I prefer the UI there anyways

[humble scaffold]

I request that we test Veo 2 VS Veo 2 Turbo

[frozen escarp]

@humble scaffold Why would that be different than using the Gemini app? The app writes prompts that creates videos that are kind of close. But, Veo 2 doesn't make the video that I want.

[humble scaffold]

If Gemini is making the prompts, then that's the issue. The developer API console allows you to use any model, and any system prompt, and any temperature. Usually I ask it to make me a system prompt for a specific thing using 2.0 flash thinking, input that as the system prompt in a new chat, and tell it what I want.

[frozen escarp]

@humble scaffold Sorry man, that is all "over my head" Even though I have an API key and I have made a personal Gemini app. I have never used the developer API console. Because, 12 years ago, I was the guy on the motorcycle.

I have a random question about the Videofx. Does it even matter if I delete the ones I don't want? I have created over 40 videos. But, I don't know if I deleted them all.

[frozen escarp]

Why give us the option to delete them?

[alpine sorrel]

@stoic orbit good news I was able to get my test subject to quiet the mouth. Bad news is that it turned them into a low budget "did egyptians worship shelled aliens" documentary cgi.

Are you doing anythign special to get the Turlte's Tale looking guy ? In theory if I add your character consistency check, his mouth should stay shut.

[stoic orbit]

Let me grab the exact prompt

"turtle with mouth closed American CG Adventure - Computer-generated 3D animation with hyper-detailed textures, realistic physics, and cinematic camera movements. Known for family-friendly characters. location: great plains of africa, yellow grass, against baby blue morning sky"

Nice work regardless, thanks! How is your seed research going? Re-using the same seed, even in ImageFX for example gives me pretty different generations each time for a short prompt, with longer prompts (like above) I seem to get consistency regardless of seed

[alpine sorrel]

Yeah my ten word intentonally vague phrases are all over the places, but when I force the linkages downa particular rabbithole it seems to be a little nicer. I got a couple weeks of experiments in me

[alpine sorrel]

Front loaded too many silent phrases and it lost the turtle

The simulations are inconclusive.

[viral spade]

Wondering where the tiny horses came from 😂

[alpine sorrel]

tiny horses in my heeeeeeead

hand. wait. dang. elton. why

[viral spade]

Right??

It's very fun to watch

[alpine sorrel]

She experiences life very differently than we do, Gaius.

[viral spade]

I've heard that before....

[alpine sorrel]

Did alien turtle hitler make deals with reptilian's outside tut's tomb? Stay tuned through these ads for hotels and lotto tickets to find out more.

I am doing things to turtle jawlines that have the fast model crawling. and then it defies me even more.

It also drive sme crazy that I can edit video on my phone faster than my laptop. Why did we not standardize this

[stoic orbit]

Oh no, what have I started

Didn’t get around to substituting other animals in place of turtle

Wonder if this is talking turtle specific , or covers all animals

Makes sense if the training data is cg animations with character dialog but still

[alpine sorrel]

otter detective is also chatty

[stoic orbit]

Feels like our very own how many r’s in a strawberry problem

[alpine sorrel]

Literally cannot hold his mouth shut.

i've tried every invisible glue, stirngs, paralyisis, 700 words linked to silent contemplation from gemini that keeps going back to the turtle. it's a blabbermouth

I even made it a tiny person inside a mascot outfit. and NOPE

[stoic orbit]

Haha I really do appreciate the effort, guess we’ll have to leave this in the hands of the team

That explains it!

[viral spade]

Made my night

[alpine sorrel]

how do we feel about paper mache origami

NEVERMIND

[alpine sorrel]

2d no budget no talent 1950s north korean animation won't even quiet

[viral spade]

Depends on what you're trying to make?

[green breach]

trying to make a movie scene with a woman and a little robot similar to astrobot, any prompt idea?

I seem to not get it right

still not quite what I had in mind.....

[calm kiln]

Generous People Here Please;

A middle-aged Hausa woman sits confidently behind the wheel of a large SUV car, aggressively doing donuts in a dusty market square in Kano. The shot is captured from the side window, showing her intense focus as she grips the steering wheel, effortlessly controlling the spinning vehicle. Dust billows outside, partially obscuring the background of market stalls and old buildings, while sunlight glows through the haze. She wears casual Hausa attire, with a traditional headscarf (gele) wrapped securely, slightly shifting from the motion. The interior of the car is dimly lit, contrasting with the bright, dusty chaos outside, as the tires screech and the ground vibrates with raw power.

A middle-aged Hausa woman skillfully skateboards through a dusty market square in Kano, balancing a large sack on her head with effortless grace. Captured from the side in a passing shot, she glides smoothly over the uneven ground, her body steady and composed. She wears casual Hausa attire, with a wrapper tied securely around her waist and a traditional headscarf (gele) wrapped tightly, keeping her look dignified despite the unconventional ride. Market vendors and shoppers watch in surprise as she maneuvers past stalls and pedestrians, dust kicking up from her wheels. The golden sunlight filters through the lively market, casting a warm glow on the scene as she passes by with perfect balance.

[green breach]

on it, give it a couple minutes

very real looking

[calm kiln]

Ok thanks

[calm kiln]

This is amazing and i like the fact that you sent all four outputs.

[green breach]

you're welcome

[waxen tree]

was that fast mode or regular

[green breach]

regular

[calm kiln]

Thanks. Were you able to generate the car drift one?

[green breach]

oh, sorry, I didn't notice. I thought you had repeated the same prompt twice accidentally

aight, I'mma do it and then work on something. I FINALLY GOT THE INSPIRATION

[green breach]

@calm kiln

[calm kiln]

Thanks so much. I appreciate.

[calm kiln]

Notifications are on, but I missed your video alerts! Is that why I sometimes miss important messages here? 🤔

[pallid delta]

Yes, using mix of prompts you write by yourself, and prompts (from range of different chat models, if possible) is a good way to stretch and grow and learn, and get better at tweaking things.

[alpine sorrel]

Also bots do a lot of purple prose that have low impact on final output.

[warm zenith]

someone needs to do backrooms stuff w this

[clear valley]

Liminal spaces he means

it’s gotten popular over the years ever since this image resurfaced on a board on 4chan (an old website)

Youtuber Kane Pixels also popularised the idea of liminals spaces and how you noclip into them randomly

[green breach]

I'm trying to make pigeons landing on a man and completely covering him, but it appears the ai struggles

[calm kiln]

If you don't mind, show us the output.

[clear valley]

^^

[calm kiln]

That's your channel? Subscribed.

[clear valley]

Nope, not mine

lunsei posted this in #flow-videos・📹

[calm kiln]

Got it. The pigeon stuff

[frozen escarp]

Is a prompt like this too vulgar to make a video of it?

Subject: Motorcycle Accident Recreation - Flip onto Hood and Rider Ejection

Scenario: A rider on an all-black GSXR750 motorcycle is traveling on the road. The motorcycle collides with the front right fender of a car, causing it to flip forward and land on the hood of the car. As the back wheel of the motorcycle lifts off the ground during the flip, the rider is ejected, flying 30 feet high and 60 feet forward. He then strikes a utility pole head-first and falls to the ground.

Camera Instructions:

• Positioning: Start behind the motorcycle, following its movement.
• Key Moments: Capture the following key moments:
  • The motorcycle colliding with the car's front right fender.
  • The motorcycle flipping forward and landing on the car's hood.
  • The rider being ejected as the back wheel lifts off the ground.
  • The rider's flight through the air.
  • The rider hitting the utility pole head-first.
  • The rider falling to the ground.
• Camera Movement: Track the motorcycle and then the rider smoothly throughout the sequence.
• Angle: Use a wide-angle lens to capture the full scene, including the road, the car, and the utility pole.
• Stability: Maintain a stable and level shot.
• Speed: Match the speed of the motorcycle and then the rider's trajectory.

Specific Actions:

• Motorcycle: Travel on the road, collide with the car's front right fender, flip onto the car's hood.
• Rider: Be ejected from the motorcycle as the back wheel lifts off, fly through the air, hit the utility pole, fall to the ground.
• Car: Be impacted by the motorcycle.

Environment:

• Location: A road with a utility pole visible.
• Time of Day: Your preference.
• Weather: Clear conditions.

I'm just trying to get this video re-created, for real :
https://youtu.be/MVIJcm8_pRo?si=E3c-pegrnuHr48NV

But,, it won't do it

[frozen escarp]

I'll try that, thanks.

But, wait... The clip I gave you is only like 5-6 seconds long.

Good call. That I will try. 👍

[reef sorrel]

its kinda hard to make all of this for an 8 seconds video. You could try using other ai video generation. but, i tried.\

My personal favorite, the rider no clipped himself (blooper generated)

[frozen escarp]

@reef sorrel Thanks for trying. I understand that I might be asking a bit too much.

[reef sorrel]

Don't worry at all, you're not asking too much! I'm actually more fascinated by the challenge of seeing how Veo will interpret your prompt. It's a complex sequence of events, and I'm particularly interested in how it will handle the physics of the motorcycle and rider's movements without relying on explicitly graphic descriptions.

That's part of the unpredictable and exciting nature of working with Veo 2.0

[frozen escarp]

@reef sorrel Good tries. Do you happen to work for Google?

[reef sorrel]

Oh no i'm not haha. I'm just a regular Google fanboy

But it would be a dream to be part of the Google team 👁️ ✨

[frozen escarp]

Oh, okay. But, I have to admit that I am the biggest fanboy of the company called Google. 👍

I actually own hundreds of shares of Alphabet stock. 👍

[frozen escarp]

Sure thing.

[frozen escarp]

The last thing I gotta say before I crash for the night :
This is why I might be asking too much. Because, I am probably confusing Veo. I am not giving the angle of the first impact, so how is the utility post supposed to be in the middle of the road?

[heavy wraith]

hello! tell me, what is the maximum length of a hint? at what volume does the context start to get lost?

[reef sorrel]

I’m still figuring that out too, I even took offficial google veo prompt examples from their deep mind page and analyses the word count, it ranges average 98-120 words

[heavy wraith]

Henry Daubrez I wrote that it is optimal to use 200 words to describe all the details of the frame. This is the one who created KITSUNE

[reef sorrel]

With the release of Veo Fast Mode, I noticed that it adheres to prompts differently than the existing Normal Mode.

[heavy wraith]

the seed phrase affects the style. I think if we want a sequence in the character when we first generated it, we need to ask gpt if we can describe it in detail, including all the small details, face shapes, details on clothes and accessories. And then use this description. And so every time something new appears in the story, what needs to be reproduced in the following scenes

[reef sorrel]

I mean, yeah, if you want the subject/character to stay consistent in every generation, then giving a detailed description is the way to go

[green breach]

the problem is that a detailed description takes a lot of the prompt, so you can't describe the rest of the scene

[frozen escarp]

@rough wing want to try your prompt improver bot?

[green breach]

I'm incurring into this issue where, whenever I ask for something magical or fantastic, the AI intentionally makes it look not real

anybody has a solution?

[heavy wraith]

why, I write a 500-word hint and everything works

[reef sorrel]

It’s like the model just picks whatever word it wants from the prompt to generate, instead of following the full intent 🙂‍↔️

Meanwhile I’m still confused as to why VEO 2 Fast mode is generally more consistent in following my prompt (though not always)

Whereas VEO 2 Normal mode tends to latch onto the first few words and then generate something completely unrelated. At this point, I’ve tried every other possible workaround to make the model actually generate without skipping any words in the prompt.

Earlier, I was trying every possible way to prompt and get the result I had in mind. It got to the point where I was about to rage quit, so I just typed a six-word prompt out of frustration.

To my surprise, one out of the four generated videos was exactly what I wanted. Before that, I kept rewriting the prompt, trying my best, but it was just frustrating

At this point, VEO felt totally RNG, like it was just rolling the dice.

And honestly, the seed setting wasn’t all that useful for me..

When the toggle seed lock was on, this was not what i had in my mind

i thought seed lock would means it stays consistent with generating subject style etc

I sure hope those four same videos didnt consume the quota lmfao

also since you were an earlier access user way before most of us, was the prompt adherence much better back then? or is it the same

Hmm, now that I think about it, what if we use reverse psychology on the model? Like, take any previously generated video from the library and prompt it exactly as it appears to see if it can recreate the same video

I'm gonna create a thread in VideoFX Bug that says: Can y’all try resetting the VEO model? Just turn it off and on again, see if that helps. 😆

/jk i wont actually do it

WHY IT TRANSITION INTO DESTROYED CITY LMFAO

[reef sorrel]

@rough wing Have you tried running your prompt through this? https://fal.ai/models/fal-ai/video-prompt-generator

I tried it earlier, and surprisingly, the output prompt actually made VEO adhere better, at least from what I can tell. I was using the Claude Sonnet model.

The preset actually helped me understand to build prompts since the prompt generator isn’t tied to any specific AI video model

[green breach]

-everything-? I'll try

[green breach]

why does

"a 100kg african man, wearing jeans, sitting on a street lamp. the camera zooms on him eating a hamburger, and a pigeon lands on his head." ......get blocked

trying to understand what triggers the filter....

uhu

Even this gets blocked: "a dark skinned man, wearing jeans. eating a hamburger, and a pigeon lands on his head."

uh. If I remove all references to his skin colour or nationality, it suddenly works

"d4rk colour3d"

(also the fact discord blocks those words is another level of strange)

now to try to make him overweight.... then have him sit in a surreal place, like a street lamp...

step by step prompt fixing

ooooooh, at long last! man that took some rewording

[half wolf]

hello guys

who has access to veo 2 make video with this prompt The Mercedes-Benz Maybach GLS roars forward, leaving behind a sunlit trail of golden grass and dust. The shot is taken from behind, capturing the taillights glowing subtly against the evening sky. The emblem on the wheels spins dynamically, emphasizing motion. The camera’s slight motion blur adds realism, making the viewer feel as if they’re watching a moment of unstoppable power and elegance.

[reef sorrel]

On it.

[reef sorrel]

Idk if this is the result you expected .

[barren geode]

I also just tried this. Glad to see it's already been reported!

[zealous mason]

$50 steam gift - steamcommunity.com/gift/id=70324882

[half wolf]

thank you

[barren geode]

Agreed.

[vestal seal]

How would you expect “lock” to work. 1 result back?

[calm kiln]

I would like someone to help me generate a military parade video. I'm just curious, will Veo really allow it?

[reef sorrel]

drop your prompt , i'll give it a try

[calm kiln]

Ok thanks. Lemme drop em

[calm kiln]

A military parade takes place in a vast, paved square in a coastal city. The scene opens with a sweeping aerial shot, showcasing the grand scale of the event. The camera then smoothly transitions to a ground-level view, capturing rows of soldiers clad in beige ceremonial uniforms as they march in perfect synchrony. On both sides of the road, spectators line the streets, watching intently. In the background, Mexican white buildings with Spanish colonial-style architecture stand proudly, their brown terracotta roofs glowing under the coastal sunlight

We were advised to stop relying entirely on bots when writing our prompts so I started practicing it.

[reef sorrel]

this was on veo fast mode so it adhere prompt differently, now im waiting for normal mode to finish

[reef sorrel]

@calm kiln

[calm kiln]

Cool. Gracias.

[calm kiln]

Could you try one more? Just one more.

[reef sorrel]

same prompt?

or

[calm kiln]

No another one

Lemme write

[reef sorrel]

sure

[calm kiln]

A group of police academy cadets is jogging through a busy market square, all in grey t-shirts and shorts. Their heads are completely shaved, so you can tell most of them are young. They look super motivated, moving in perfect sync. People on both sides of the street stop what they're doing to clap for them, and some even pull out their phones to take pictures. The market is full of life—colorful stalls, fresh produce everywhere. The white buildings with terracotta roofs in the background just add to the whole vibe.

[calm kiln]

It didn't do a great job with the prompt adherence on the vids though. Camera didn't transition to ground level.

[calm kiln]

But I'm very grateful.

[calm kiln]

Nice one!

[reef sorrel]

genersting

[calm kiln]

Ok..

[reef sorrel]

fast mode

i

i think it ignored shorts

[calm kiln]

Yeah. That's why sometimes when writing prompt on image FX, I write some words 2-3 three times. Sometimes when you put emphasis it hardly ignores them.

[calm kiln]

No one gives a crap about them here. Everyone's just minding their business 🤣😂🤣

Thank you so much for vids. I really appreciate.

[reef sorrel]

im generating another one

with modified prom,pt

[calm kiln]

Ok that's great

Sound on

[reef sorrel]

Finally

people clapping

[calm kiln]

@reef sorrel you're so generous. Tnx

[reef sorrel]

have a blooper one lol

failed gemeration vid

anothrt one

[calm kiln]

One dude literally disappeared here😂🤣

[reef sorrel]

he fell and shy to get up

i can relate

[calm kiln]

He melted😂

[reef sorrel]

another one i missed, it was same batch of fastmode idk why it generated like this

[calm kiln]

Just seeing this now. This is nice, too.

[calm kiln]

But Wrigley pardon my ignorance. What does this fast mode means? I'm trying to understand which of two is it; Generating videos quickly or videos motion is fast?

Forgive my English it's not my first language

[reef sorrel]

basically, Fast Mode is a new mode that generates video quickly, alongside Veo 2

[calm kiln]

Okay, so it's a different model with a different quality, etc., separate from Veo 2, since you said 'alongside.' (You know, where I live, words tend to have different meanings.) Or is it the same thing?

Just that on fast mode, videos generate quickly. Is that the only difference?

[reef sorrel]

Fast Mode is a version of Veo 2 that generates videos faster

i might've misphrase that

[calm kiln]

I get it now.

[reef sorrel]

Yeah, it’s fast at generating, but it handles prompts differently. I’m still experimenting with it

[calm kiln]

That's great.

I'm just waiting for the day when I will finally put my hands on Veo 2. I'm kinda jealous of you guys. You can just sit down and imagine something in your minds, and then just turn it into video anytime you want.

I want to be the first African living in Africa to have access to it😂🤣

[reef sorrel]

Don’t worry, you’re not missing out on much ! VideoFX is still a work in progress, but it’ll be available for you eventually soon

I can’t say or know when it will be available, but for now, the Google Labs team is still working on improving it. It might take some time before a public release.

[calm kiln]

Yeah and I hope this delay is for the good of the users.

[reef sorrel]

Rest assured, if you need a video generated, just ask anyone in this channel.

Early access users here are quite generous. 😉

[calm kiln]

I can attest to that. I'm grateful to y'all.

[formal marsh]

Something like: seasonal transformation of the same landscape scene from spring to summer to fall to winter.

• timelapse of the above

Any!

Results from Prompt 1, Fast Mode

Prompt 2

Prompt 3

[formal marsh]

Definitely challenging. I found that when I put a circular fountain in as a focal point it seemed to look a bit better. I can try more later.

[molten parcel]

i hope this means that image to video is also coming to the actual videofx website🙏

[reef sorrel]

Oh my god

Really had to release that feature on the day my fal credit hit -200 balance lmfao

Wait does that mean a lot of other inference api website got access to this now

Let me check Poe

Oh which one?

Poe still hasn’t got access to image to video yet

the forth one is so good

i need to try this out lmao

its the moast natural movement that came out of image to video generation

[reef sorrel]

these are so cool

I wonder if image to video hav any issues with prompt adherence (jump cut to scene after 2 sec of showing original image etc etc)

MY BODY IS READY GOOGLE LABS TEAM

I'm excited to see what they have in store for us

[reef sorrel]

No doubt, they’re getting it ready for us 👀

[modest agate]

So maybe they gonna make a sign up limited seats for this feature or smthn ?

[reef sorrel]

I think it’s gonna roll out to early access users as usual.

[frozen escarp]

I prompted :Bruce Lee swinging his nunchucks around him .
https://youtu.be/m3Yn-UhcYWs?si=eAHT6PKq7UtRKoiR

Am I asking too much? It won't make the nunchucks correctly.

[frozen escarp]

Good call. Do you happen to be a Google employee who is working on this?

I mean, really, is anyone from Google ever in here?

[reef sorrel]

Would this help clear up

Oh what the heck

If this ain’t me

Oh my god

What was the prompt even?

But that’s a cool way to summon horse midair 🙈

[molten parcel]

https://tenor.com/view/horse-flying-flying-horse-dream-dreams-gif-19273718

[reef sorrel]

That’s wizard-level horse magic right there 🧙‍♀️ 🐎

[frozen escarp]

@rough wing Thank you for telling me about the "Flying through the air" Hallucination troubles that every model has. 👍 I'm done talking about this. Peace ✌️

[reef sorrel]

Oh

Wait speaking of that, mine have that problem where some of my generated video were deleted in my library even though I didn’t

[frozen escarp]

Not that I know of. I was just trying to show you some appreciation. But, I hope you can stop tagging me in your messages. Go back to normal.

[reef sorrel]

Hmmm now let’s not quarrel

[frozen escarp]

^Good one. I just saved it. 👍

[reef sorrel]

@rough wing i think those blank video u see in ur library are failed generation

if u cant see any prompt when u hover ur mouse on it , then it might be fail generation

at least from my experience

can u try click swipe card

see if text prompt exist for the blank vid

yeah those are indeed fail generation-

but weird why u have that many

i do notice one of my success generation video was gone, and i didnt delete it

ok thats not relevant

sorry lmao

but

tldr library buggy as hell

[reef sorrel]

Wha

[barren mist]

Monochrome animation, minimalist style, halftone technique, dot matrix imagery, quick cuts, bold geometric shapes evolving, Google AI ad, highlighting AlphaFold protein structures, AlphaZero chess algorithm, Imagen image generation, Google AI video generation technology. The animation should visually represent the evolution of these AI systems, starting from abstract dot patterns and culminating in clear, simplified depictions of protein structures, chessboards, generated images, and video snippets. Modern, dark and light contrast, ending with the Google AI logo.

[barren mist]

Can someone try this?

[reef sorrel]

Sure

Ok let me try again

[reef sorrel]

I’m pissed cuz I couldn’t obtain the generated videos from the library page, it’s laggy on my iPhone and I’m away from my laptop unfortunately

[reef sorrel]

Surprisingly Fast Mode seems to produce better results when it comes to this style of video

[reef sorrel]

This was normal mode and , it was 1 out of 4 of the video that actually showed “google ai” text

So earlier I was experimenting something, applied the prompt to the Step 2 "Motion" section of Text-to-Image-to-Video, where apparently the "Prompt Enhancer" is enabled.
—

Original : Monochrome animation, minimalist style, halftone technique, dot matrix imagery, quick cuts, bold geometric shapes evolving, Google AI ad, highlighting AlphaFold protein structures, AlphaZero chess algorithm, Imagen image generation, Google AI video generation technology. The animation should visually represent the evolution of these AI systems, starting from abstract dot patterns and culminating in clear, simplified depictions of protein structures, chessboards, generated images, and video snippets. Modern, dark and light contrast, ending with the Google AI logo.

Enhanced prompt : Monochrome animation in minimalist style using halftone and dot matrix imagery with quick cuts showcasing bold geometric shapes evolving dynamically. The animation visually represents the evolution of Google AI systems—AlphaFold, AlphaZero, Imagen, and video generation—starting from abstract dot patterns. These patterns morph into simplified protein structures, chessboards, generated images, and short video clips, all rendered in a modern style with dark and light contrast. The animation's motion is fast-paced, with shapes shifting and reforming rapidly, reflecting the speed and complexity of AI processes. The video concludes with the Google AI logo appearing.

[reef sorrel]

They actually updated that now haha

the improvements to text to image to video

[reef sorrel]

Another try with “prompt enhancer” from Text to Image to Video Motion section

Original prompt (originally by @formal marsh ) : seasonal transformation of the same landscape scene from spring to summer to fall to winter. + timelapse of the above realistic

Enhanced prompt : Timelapse of a realistic seasonal transformation of the same landscape scene from spring to summer to fall to winter. The camera pans slowly across the scene, capturing the subtle changes in the foliage and the overall atmosphere. In spring, vibrant green shoots emerge, flowers bloom, and the landscape is alive with color. As summer arrives, the greens deepen, and the scene is filled with lush vegetation. The camera smoothly tracks the transition to autumn, where the leaves turn golden, crimson, and russet before falling to the ground. Finally, winter sets in, blanketing the landscape in a pristine layer of snow. The camera's steady movement and the gradual change of seasons create a mesmerizing visual journey through the year. The timelapse effect compresses the passage of time, allowing the viewer to witness the complete seasonal cycle in a short, captivating sequence

To my surprise, it still won’t generate video of actual Timelapse from spring to summer etc even with using their enhanced prompt

[reef sorrel]

This

[reef sorrel]

Check the library

And look at the prompt

[reef sorrel]

It applies only to the text-to-image-to-video feature. The prompt enhancer uses the generated image as a reference. If the user adds motion-related inputs in step 2 (e.g., “move” or “camera track”), it automatically combines the image-generated prompt with the user’s motion input.

If the generated image is an empty room or anything that doesn’t have any obvious visible strong subject, and the user adds a motion prompt in step 2 that isn’t relevant (e.g., a strongly worded text-to-video prompt unrelated to the empty room), the system will enhance and prioritize the user’s input in step 2, effectively ignoring the original image.

I only find out about this because of this particular post.

[frozen escarp]

Why don't I ever hear sound in the videos?

It might be because, I haven't added instructions for the sound in the prompt.
For videos like these, what should I put in the prompt?

https://youtu.be/Faz3sNwL4rw?si=z5tmhWK5Yxoex3uO

https://youtu.be/G3uo32MNkZE?si=hjbaKHUd8Wq7m-yH

https://youtu.be/mcAJE7xLbxA?si=fy7RFUgDf_Ywaolp

[reef sorrel]

VideoFX doesn’t generate sound, only video

[heavy spade]

Ultraman Tiga and Godzilla fighting.

[half stream]

Is anyone good at prompting able to help me?

[normal coyote]

We’re here to help, just let us know what you need, and we’ll do our best to assist!

[calm kiln]

Hey is anyone going to help me run some prompts? I'd appreciate it greatly. Thanks in advance

[fair venture]

A clean slice of tomato getting made in a salad

[low berry]

Sure what is your prompt

[fair venture]

Thank you

[calm kiln]

"A bright, sunny day. A confident African man stands in front of the UMTK office, wearing a blue company t-shirt with "UMTK" printed in white across the chest. A matching blue baseball cap with the same white "UMTK" logo shades his face. His mouth moves as if explaining something important. His hands gesture naturally, emphasizing key points.
The camera smoothly pans, revealing a neat row of gleaming solar panels arranged in front of the office. The panels reflect the sunlight, showcasing their efficiency. The company building stands in the background, modern and professional."

[calm kiln]

Thank you so much.

[plush jungle]

Cat

[wise tinsel]

That's really good. Never expected these from Veo 2
It's not available in my country yet, so I'll have to wait to use myself 🥲

[clever cipher]

"a luxurious stone fireplace in a warm, inviting living room. Show realistic flames dancing and crackling, with occasional embers rising. Include subtle smoke wisps above orange-red flames. Capture the gentle flickering light casting warm shadows on surrounding walls and furniture. Add a soft, ambient glow illuminating plush armchairs, wooden details, and decorative elements nearby. Style should be photorealistic with cinematic lighting. Frame the shot to leave space for text overlays."

[clever cipher]

shanks my friend ✨

[clever cipher]

"cheese pizza. Start with a close-up shot of a slice being pulled away from the whole pizza, showing long, stretchy cheese pulls glistening in warm lighting. The cheese should be perfectly melted, creating those irresistible stringy pulls that extend dramatically. Include slow-motion footage of the golden, bubbling cheese surface with some herbs visible. End with the slice being lifted higher, creating even longer, more dramatic cheese pulls against a blurred restaurant background. Use warm colors and subtle steam to enhance the appetizing quality."

[normal coyote]

Ok, I order pizza now.

[formal marsh]

Same.

[clever cipher]

xd

thanks @normal coyote

thanks brother @formal marsh

[clever cipher]

"a macro photograph or digital artwork showcasing the mesmerizing details of a vibrant pink flower, possibly a tulip, with its layered petals and subtle color gradations; use soft, diffused lighting and a shallow depth of field to emphasize the delicate textures and intricate forms, evoking a sense of tranquility and inviting the viewer to immerse themselves in the beauty of the natural world."

[summer lagoon]

Cinematic wide shot of a dark, overgrown garden at night. The garden is filled with ancient, twisted trees, glowing flowers, and dense vines illuminated by soft moonlight. A leafless, gnarled tree stands at the center, with fog drifting low across the ground. Scattered throughout the garden are aged stone statues of skeletons, partially covered in moss and cracked from time. The air shimmers with faint particles of light, like dust or spores. A faint breeze causes the glowing flowers and vines to sway gently. A vibrant blue butterfly flutters slowly through the scene, occasionally resting on a branch before taking flight again. The lighting is cinematic and moody, with deep shadows and silvery highlights. Ambient sounds of wind, distant whispers, and chimes echo softly, creating an eerie yet beautiful atmosphere. The scene feels timeless, surreal, and photo-realistic — like a moment from a forgotten myth.

• Style: Cinematic realism, photorealistic textures, dark fantasy
• Lighting: Soft moonlight, high contrast, fog-enhanced ambiance
• Lens: 35mm with shallow depth of field
• Mood Reference: Similar to an A24 horror or dark fantasy film — atmospheric and elegant, not horror-jumpy

[summer lagoon]

Oh nice, that doesn't look too bad actually.

[clever cipher]

a dynamic visual representation of data transmission through fiber optic cables. A single line of light bursts into multiple strands of vibrant blue and purple, with glowing particles traveling along them, mimicking the flow of information. The cables extend and connect to a surface reflecting the light, creating patterns of illuminated data points that shift and change, symbolizing the complex processing of digital information in a network or computing system. The overall effect is a futuristic and abstract portrayal of highspeed data transfer and digital connectivity.

[hidden belfry]

moiré effect

@reef sorrel An enormous salt titan with jagged, crystalline limbs forms a colossal hammer from its own body. Towering over a cracked desert landscape, the titan lifts the gleaming weapon high above its head. The camera captures a wide-angle shot, slowly dollying back to emphasize the scale. With a thunderous roar, the titan slams the hammer into the ground, sending a shockwave of crystalline dust and shards into the air. The camera shifts to a low-angle close-up of the debris scattering, dust clouds rising, and the titan’s stoic face partially obscured by the settling mist. Cinematic, epic lighting with sun rays filtering through the haze.

[reef sorrel]

Damn it really adhere to your prompt well

[hidden belfry]

yeah

[reef sorrel]

Was the result something you had in your mind in the first place

[hidden belfry]

Kind of

[hidden belfry]

I was trying to represent the hammer slam thing in Pls Donate roblox

[reef sorrel]

Oh

[hidden belfry]

Wait that sounds good Ima try to generate it

[molten tulip]

We're aware and working on a fix 🙂

[summer lagoon]

Documentary style footage of a rainforest

[summer lagoon]

Oooh

[plain heath]

First-person perspective, ultra-detailed and cinematic, walking down a bustling city street in 1960s Iran. Traditional Iranian architecture blends with modernist 60s buildings—flat concrete facades with decorative brickwork. Classic cars like a Paykan, Chevrolet Impala, and Citroën DS drive slowly through the street, honking occasionally. You pass a row of shops with Persian signs in Nastaliq script, fabric awnings casting striped shadows on the pavement. A boy sells newspapers from a wooden crate; headlines in Farsi flutter in the breeze. A woman in a floral chador walks past holding a basket of vegetables. Posters of Mohammad Reza Shah are pasted on the side of a building. The sun reflects off the chrome bumpers and windshields.

[hidden belfry]

Anyone know how to stop the whole random words thing when I try to type a prompt in on mobile?

[summer lagoon]

Can someone generate documentary grade footage of a rainforest pls? 🙏

[summer lagoon]

Oooh

[reef sorrel]

Wait sorry I missed yours

[summer lagoon]

No problemo friend, it was a short message compared to the usual

[reef sorrel]

I generated this on the go from my phone , I was struggling to load library page

[summer lagoon]

It’s beautiful, thank you Wrigley!

[molten parcel]

A ‘Newest to Oldest / Oldest to Newest’ sorting feature would also be a really cool addition to the VideoFX library.

[plain heath]

thanks a lot!

[molten parcel]

i wanna do a calm anime scene with this character, but in a cafe drinking coffee while raining outside

p.s. his name’s haruki :)

[waxen tree]

whisk animate ya'll we finally here

[livid scarab]

Please create a video:

Flying camera movement through a dense forest, ultra-wide angle lens perspective (10mm), smooth, aerial-like motion weaving between trees. Various forest animals visible in the distance, several foxes moving naturally through the undergrowth, a large, luminous sign reading "llacker" glowing among the trees, dappled sunlight filtering through the canopy. Cinematic color grading with rich greens and golden highlights, slight fog or mist for atmospheric depth, warm backlighting creating silhouettes of animals, subtle particle effects (dust, pollen) in sunbeams. Ambient forest sounds (birds, rustling leaves), soft, atmospheric musical underscore, minimal sound design for the glowing sign (gentle hum or shimmer). Slow, deliberate camera movement. The video should evoke a sense of mystery and discovery as the viewer is taken on a journey through this enchanted forest where technology and nature converge.

[neon zodiac]

Someone can generate a video for me too?

[molten parcel]

sure, send me your prompt

[neon zodiac]

for some reason, I can't send it here

[hidden belfry]

It's because of the first word in the prompt

[neon zodiac]

😂 😂 😂 omg

[hidden belfry]

Yep

[neon zodiac]

dark background with a transparent red sphere in the center full of various activities and tasks, such as learning a language, playing the guitar, playing soccer, writing a book, studying... and in the other scene they are subdivided into small transparent red spheres with individual activities and tasks in each sphere. image style surrealism, anime and cartoon

[hidden belfry]

There we go

[molten parcel]

ok!! generating right now :)

[neon zodiac]

this is incredible! Thank you so much

[molten parcel]

np!

[neon zodiac]

can you make this too? : the character sitting at the table in his room studying. his expression is concentration and determination. surrealism anime and cartoon style

[neon zodiac]

Thats funny haha, thanks

[reef sorrel]

One thing I learned so far from using VideoFx, if you want Veo 2 to generate four actual different variations based on your prompt, keep your prompt short, rough, like an initial idea.

This is a great example

“the character sitting at the table in his room studying. his expression is concentration and determination. surrealism anime and cartoon style”

[neon zodiac]

Can you generate this for me:
character walking confidently but restrainedly. simple clothes. Anime and cartoon style. neutral colors

[hidden belfry]

👍

[neon zodiac]

Thank you 😊

[hidden belfry]

No problem 😄

[wild relic]

Hi! Could someone try a prompt for me? I2V

For this image:

[mellow girder]

@wild relic

[wild relic]

Thank you so much, I really like the TV effect on them 🙌

[molten parcel]

why is nobody doing my request? :(

[wooden canopy]

Write me your prompt ill do it

[summer lagoon]

I think it would be funny to see how it interprets what a cult of artists would do/look like

[reef sorrel]

Let me try

[summer lagoon]

Thank you kind sir

[reef sorrel]

@summer lagoon

[reef sorrel]

this is my fav one lmao

[summer lagoon]

💀

[reef sorrel]

im generating a new one , was refining the prompt so that it can do lfound footage style

[summer lagoon]

😂 I love it

[reef sorrel]

@summer lagoon

[wild relic]

Hi guys, could someone try this image on I2V for me please?

[reef sorrel]

im not sure if this would pass but ill try,

[wild relic]

Thank you wrigley

[summer lagoon]

💀

https://tenor.com/view/me-atrapaste-es-cine-its-cinema-cinema-esto-es-cine-gif-17729711691959966457

[reef sorrel]

@wild relic

same girl

this is actually amazing ngl

[wild relic]

omg thats amazing

look at what sora generated for me 😂

[reef sorrel]

oh my 👁️

[wild relic]

😭

scary

[reef sorrel]

not surprised as expected from sora lmao

idk if they intentionally make sora turbo image to video work badly

[wild relic]

I thought so at first but now with the new imagegen I think its just bad (sora video)

[reef sorrel]

the gap between how good the image generation is and how underwhelming the Sora video model turned out is just way too obvious.

[summer lagoon]

💀

[reef sorrel]

but they might be cookin somethin behind the scene, so we dont know 👁️

[wild relic]

exactly

[wild relic]

They definitely have a non-turbo version of Sora that is really good. are they ever going to release it? who knows

[reef sorrel]

What’s funny is Veo also has a 'turbo mode', they just call it Fast Mode, and it actually performs way better than Sora.

But anyway, let’s not get too into it! 👀

[wild relic]

True True

[molten parcel]

feel free to edit if u like

bruh why can’t they allow anime prompt (T_T)

ugh fine

[reef sorrel]

I wanna say , it’s that one word

[reef sorrel]

let me try

Can you cheer me up

jk

one thing to know about me is

i mean , i love me some bird and some men and some women and some they them-

also thats not mine

LOL

@molten parcel I did it

[reef sorrel]

this is like the closest thing we get for "Reference" feature except its img2video

The level of fidelity it produces is impressively high

[wild relic]

Because it’s too expensive, just like Veo2 still isn’t released (I guess that’s the reason, I’m not sure)

[reef sorrel]

technically veo is released through api

has been *

[wild relic]

Through third party only right?

[reef sorrel]

yeah

i mean theres google api vertex veo , but its only for selected users

just like early access user for videofx- oh..

https://tenor.com/view/spongebob-spongebobsquarepants-gif-9146238

[wild relic]

I’m ngl i would definitely pay 200 dollars per month for Veo2

I did for sora and it wasn’t worth it

[reef sorrel]

oof

theyre probably waiting until they can combine it into all in one tools for gemini (dont quote me on that) 🙂

[wild relic]

I saw a tweet saying that the built in code from Gemini had sources of Veo2 generation inside the chat bot

It would be a great addition

[reef sorrel]

yeah codename "robin" , which is codename for generating video using veo 2

In all fairness, the VideoFX tools as of right now, are still not quite ready for release to the public

Yeah, even the team and staff are still struggling with a few things, like right now, the mobile UI isn’t fully ready or optimized for user experience yet. So they’re really counting on early access feedback to help shape things up

[molten parcel]

i like the 3rd & 4th one :)

[reef sorrel]

well they lurk and they heard 👀

For now, I’ve been sending feedback directly through the website instead

[molten parcel]

i wonder with veo, how close can we get to recreate “Up&Up” by coldplay

the MV, i mean

[reef sorrel]

im more surprised it does 2d+3d+realistic, i wonder what happened if i specifically prompted it "anime"

[formal marsh]

Is there anything about how to best prompt for I2V? Like this: https://cloud.google.com/vertex-ai/generative-ai/docs/video/video-gen-prompt-guide

Are you using the same prompting strategy for I2V as for T2V?

[reef sorrel]

What I did was give it a rough short initial idea as a prompt, since when you send one, it automatically enhances or expands it based on both the image input and what you wrote.

[wild relic]

That’s an AMAZING idea

[reef sorrel]

like for example , this is what i wrote

[reef sorrel]

And the outcome was exactly what I wrote in the initial prompt.

Not 100% sure, but you could try use any prompt assistant for image-to-video. It usually gives a medium-length or long prompt. If your prompt’s too short, the model will kinda freestyle based on your input and the image

🙂

[formal marsh]

I tried shorter and longer ones, but my results haven't been great so far. I'll keep trying.

[reef sorrel]

wait can i give your image a try?

[formal marsh]

I agree with that.

It's finding the balance of specific + vague.

[molten parcel]

is that one of the scene where a woman jumps off of a turtle like a cliff?

or something else…

[wild relic]

Something else, just made it in the same style

[molten parcel]

ah i see

[reef sorrel]

i do that all the time when i dont have any idea what i want it to do with the image i uploaded

i simply just put "."

from my experience when you blank prompting, the model will rely on the image input 100%

[formal marsh]

For this image: #flow-videos・📹 message

When I made a complicated prompt trying to get the mural to animate, it didn't animate the mural at all. It just did a normal panning video around the mural. So I simplified the prompt down to "Birds fly" and the bird looked great, but it put a random boy in.

The second video was "Birds fly. People dance gracefully." and it was a bit better.

[reef sorrel]

im gonna give this a try 😂

maybe if i say please it will do it

its generating

[formal marsh]

ok

[reef sorrel]

yep nothing it generated makes the mural comes alive,

[formal marsh]

So it seems like anything painted is challenging.

[reef sorrel]

@formal marsh

instead of animating the mural on wall, it create its own drawing and animate it on the wall

https://tenor.com/view/spongebob-spongebobsquarepants-gif-9146238

slightly W I D E wolf

i mean at least its not as wide as this

@formal marsh

the closest thing i could get it generated

[formal marsh]

Wow! I love it!!!

What did you do?

[reef sorrel]

oh gosh i deleted the vid so i didnt get the prompt but

i basically use prompt assistant on chatgpt, uploaded the image and asked it to generate image to video prompt,

[formal marsh]

ok

[reef sorrel]

tho it was only 1 out of 4 video that was right

Static wide-angle shot, perfectly front-facing to maintain mural composition Subject Focus: A surreal, multi-limbed humanoid figure with interwoven arms extending outward, surrounded by birds perched and mid-flight Subject Movement: Gentle, slow rippling of the many arms as if flowing underwater, with fingers slightly curling and uncurling, eyes of the central face slowly blinking, while perched birds occasionally ruffle feathers or chirp and flying birds flap wings subtly in place Subject Details: The figure is formed of layered, semi-transparent textures with crosshatch patterns, earthy and metallic tones blending together, floral motifs blooming along the arms and base, with small figures emerging from the body, all gazing outward with solemn expressions Scene: The entire scene is fixed on a beige stucco wall, maintaining the wall texture in the background to keep the mural grounded in its real-world setting; minimal elements outside the mural frame Camera Language: Locked-off camera in a centered wide-angle view, absolutely no panning, tilting, or zooming—preserving the exact orientation of the original street art Lighting: Soft daylight with ambient diffusion, subtly shifting highlights on the arms and birds to simulate natural lighting without changing the angle, light refractions pulsing faintly on reflective textures Texture: Chalky wall surface with brush-stroke layering, painterly textures mixing crosshatched lines and organic floral swirls, matte stone-like finish for the background wall contrasted by semi-gloss on the character forms Atmosphere: Dreamlike and introspective, evoking a feeling of connection, fragility, and flow between humans and nature, poetic and contemplative with a surreal, mythological tone Environment: An urban exterior wall acts as a static canvas, no added elements outside the painted mural—animation breathes life into the existing shapes without distorting their original form or placement

yes its very long, i simply copy and paste it without modifying anything '

[formal marsh]

Wow that is so long and it turned out so well.

[reef sorrel]

u can give it a try! maybe differ users get different result

[formal marsh]

I2V adding a parrot to the reference car image.

[formal marsh]

I tried your prompt. I think your video turned out better.

[summer lagoon]

Triangular prism*

[reef sorrel]

for mine it was only one out four video that managed to get that result

heres another take

[formal marsh]

Looking good!

[reef sorrel]

I refer to the prompt,or the first step after uploading an image, as ‘instruct’ :
analyze the drawing of mural (subjects and everything) on the WALL, then animate fully the mural drawing on the wall to make it alive and fully animated, only the drawing on the wall should be animated

[formal marsh]

This is actually how I imagined it coming to life 😍

[reef sorrel]

also im on it 👀

[reef sorrel]

none of mine dispersed rainbow unfortunately

im trying again

[summer lagoon]

As always, thank you kind sir 😌

[summer lagoon]

That’s pretty cool, it seemed to treat the color as a fluid of sorts

[summer lagoon]

Sweeeet, that first one looks close to the dark side of the moon album cover

[golden lichen]

oh hey. I’d like to make a VideoFX video before the sun explodes.

[fallow frost]

Single Prompt "driving" and he's clearly having a great time. Love the detail in sunglasses and watch, looks like he is cruising on a Sunday. Very cool.

[full cove]

Really good 👍

[calm kiln]

Someone please help me run this (i2v)

A white Toyota Land Cruiser aggressively performs high-speed donuts on a wide, dusty tarmac road flanked by traditional Kano-style mud-brick houses. Thick white smoke billows from the spinning tires, mixing with the golden dust kicked up by the powerful motion. The animation features dynamic camera switches: a low-angle front shot capturing the rotation and smoke bursts, a high drone shot circling the vehicle, a rear tire close-up showing the intense friction and smoke, a fast side tracking shot keeping pace with the spin, and a slow-motion cinematic close-up as the car slides with sunlight flaring in the background. The scene is intense, energetic, and filled with raw power.

[calm kiln]

Wow. Cool. Except, no driver in the car.

[tulip valve]

Is it possible to make a looping video in Veo with prompting?

[calm kiln]

Someone please run this on I2V. Thanks

"Animate the man so that he is delivering a speech confidently at the podium. His mouth should move naturally as if speaking, with occasional hand gestures for emphasis. Add subtle head movements and blinking to enhance realism. The red screen behind him should glow slightly, and the logo should remain clearly visible. Keep the camera angle steady, focused on his upper body and podium. The atmosphere should feel formal."

[reef sorrel]

Let me try

[calm kiln]

Ok. Waiting.

[reef sorrel]

“Fast Mode”

@calm kiln

[calm kiln]

The fast mode ones are the best. Exactly how i want the Videos to be, no changing camera angles.

Thank you.

[icy pasture]

Im not even sure which is better, the normal one as you said keeps changing things for no reason, but fast mode (in my case) has really bad motion

[calm kiln]

I just love how it doesn't alter the look of the picture

[icy pasture]

Yeah you're right

[elfin rose]

shih tzu blanca en un cohete

/crea

[reef sorrel]

I am not a 🤖 but here you go

[elfin rose]

I don´t know how creat my videos

[reef sorrel]

Estaría genial, pero Discord no tiene esa opción jaja. Toca usar otro programa 😄

[stark sphinx]

whats the limit gen on aistudio? is it daily or monthly? and how much?

[calm kiln]

I think it's four to six.

[stark sphinx]

Monthly or daily?

[calm kiln]

Monthly? No. Lol!

The limit isn’t even daily—it applies within a single session. Once you're in a session, you can’t generate more than three videos. After that, it takes a few hours to reset. I’ve tested this multiple times and successfully generated three videos per session on several occasions, today. Get it?

[stark sphinx]

Session you mean after generating the first one? And the hours are random?

[steady mesa]

i heard someone mentioned that the rate limit resets each five minutes or so

he generate 3 videos then hit the limit then waited 5 minutes then he was able to generate 3 more videos

[calm kiln]

It's very possible. I'm not sure though but it's possible

[calm kiln]

I don't know how to put it. English isn't my First language but let me say it this way, you may understand; If you wake up tomorrow and generate three videos in the morning, you may not be able to generate it until maybe afternoon...and so on.

[steady mesa]

have you tried?

[calm kiln]

Not after 5 mins though. It was after like 1-2 hours.

But tomorrow. I'll try it. I'll Generate the vids when I hit the quota, I'll wait after 5 mins and try again

[calm kiln]

You don't have access?

[steady mesa]

not yet, i am still waiting

i am not seeing it on my ai studio

[stark sphinx]

let me test it

[steady mesa]

so?

[stark sphinx]

didnt worked, i waited exactly 5 minutes after generating the 3 videos

[steady mesa]

oh

[stark sphinx]

i waited 10 and didnt worked too

[stark sphinx]

i think i got it, youre probably right, its not 5 or 10 minutes, maybe couple of hours at least

[steady mesa]

well, I still dont have access to it, so I cant test it

[verbal nacelle]

does it counts the failed attempts as *a legit attempt like once you hit Run even if it fails it counts as one of 3? Cause thats what its doing

for me

[calm kiln]

Yes. That's something I observed this morning. Even a failed attempt counts.

And again, sometimes you can't even generate three vids. It's just one. You have to wait until after a few hours to generate another one.

[neon zodiac]

Someone can generate a video for me?

The prompt: dark ox dressed in a rock jacket, boots and sunglasses in a show dancing with kids 3D cartoon pixar style

[calm kiln]

Ah "kids"!! Forget it. That's something you'll never see on Veo 2.😂😂

[neon zodiac]

Thank you! 🙏