Native synchronized audio is what sets Meta Muse Video apart — from Meta Superintelligence Labs, announced July 7, 2026 and ranked #3 on the LMArena human-preference leaderboard for text-to-video at preview. Sound is generated in the same pass as the picture: dialogue, ambience, sound effects, and music all come out of one generation. The trick to using it is simple — describe the sound world in your prose prompt and name which visible object or person makes each sound, so the model has a frame to sync to.

Every prompt below is one flowing paragraph written the way Muse rewards — subject, one action, an explicit camera move, setting and light, then the audio line, ending with the aspect ratio stated in words. No comma-tags, no --flags, no settings line. Meta notes sync is still improving, so spoken lines stay short. Paste any of them into the Meta AI app, meta.ai, Instagram, or WhatsApp. For the full formula see the best Meta Muse Video prompts roundup and the guide to prompting Muse Video for realistic video.

Advertisement

Dialogue scenes

Keep spoken lines short and put the exact words in quotes — sync is still improving, so name the one speaker and give them a single visible action while they talk. Seven talking-head and two-person beats.

1. Cafe confession

A young woman in a rust-colored sweater sits across a small cafe table, wrapping both hands around a warm mug as she leans in and says softly, "I should have told you sooner." The camera holds a steady medium close-up on her face. Warm afternoon light falls from a window on the left, milk steamer hissing faintly in the background. Portrait 4:5 format, one continuous shot.

Why it works: one short quoted line plus a steady close-up keeps her mouth readable while the background steamer supplies gentle ambience.

2. Rooftop goodbye

A man in a denim jacket stands at a city rooftop railing at dusk, turning to face someone off-camera and saying quietly, "Take care of yourself." He gives a small nod as he speaks. The camera pushes in slowly on his face. Cool blue evening light with the warm glow of streetlights below, a soft breeze and distant traffic humming under his voice. Widescreen 16:9, one smooth shot.

Best for: emotional short-film beats — the slow push-in and one clean line carry the moment.

3. Kitchen argument, defused

Two friends stand in a bright kitchen, one drying a plate; she sets it down and says, "Okay. You were right." The other exhales and half-smiles. The camera holds a steady two-shot from the counter's height. Clean daylight through a window, the clink of the plate on the rack and a faucet dripping softly under the exchange. Widescreen 16:9, a single unbroken take.

Why it works: a two-line exchange stays within safe sync range, and the plate and faucet give the audio real on-screen sources.

4. Detective at the door

A weathered detective in a long coat stands under a porch light at night, knocking twice on a wooden door and saying flatly, "We need to talk." Rain drips from the eaves behind him. The camera stays locked at eye level in a tight medium shot. Moody amber porch light against deep shadow, the two knuckle knocks and steady rain audible under his line. Portrait 4:5 format, one continuous shot.

Best for: noir and thriller cold opens. The knock is a visible action tied directly to its sound.

5. Product founder to camera

A founder in a plain tee sits in a bright modern studio, looking straight into the lens and saying warmly, "We built this for you." She lifts the product into frame as she finishes. The camera stays static in a clean medium shot. Soft even key light, a faint room-tone hush behind her voice so the line sits forward and clear. Vertical 9:16 format, one continuous shot.

Why it works: a static frame and a short line keep lip-sync clean for a direct-to-camera brand clip.

6. Grandmother's advice

An elderly woman sits in a sunlit armchair, knitting needles resting in her lap, looking up to say gently, "Slow down, dear." She pats the arm of the chair once. The camera holds a soft medium close-up. Warm golden window light and dust motes drifting, a clock ticking quietly on the mantel beneath her words. Portrait 4:5 format, a single unbroken take.

Best for: heartfelt vignettes — the ticking clock names a source and grounds the room.

7. Two friends, one joke

Two teenagers lean against a row of lockers; one nudges the other and says, "You actually did it?" The second grins and shrugs. The camera holds a loose two-shot at the hallway's level. Bright fluorescent school light, a bell ringing faintly down the corridor and distant chatter under their quick exchange. Vertical 9:16 format, one continuous shot.

Why it works: a short back-and-forth plus the nudge gives Muse two clear actions and hallway ambience to sync against.

Ambience & atmosphere

No dialogue here — the audio is the environment. Name the sources: rain on a surface, wind in leaves, waves, café clatter. These make effortless silent-companion loops that arrive already scored by their own world.

8. Rain on a tin roof

A covered wooden porch during a steady evening downpour, water streaming off the corrugated tin roof and pooling on the steps below. The camera stays locked off, framing the rain curtain and a single glowing lantern. Cool blue twilight with the warm lantern glow, the audio carried entirely by rain drumming on the tin, drips off the eaves, and a low rumble of distant thunder. Widescreen 16:9, one continuous shot.

Best for: sleep and focus loops. Naming the tin roof and the eaves gives the rain a specific texture.

9. Forest at first light

A misty pine forest at dawn, shafts of light cutting through the trunks as fog drifts low over the ferns. The camera makes one slow push forward along the trail. Soft cool morning light warming toward gold, the soundscape built from birdsong overhead, a light breeze rustling the pine needles, and a stream trickling somewhere off to the right. Widescreen 16:9, a single unbroken take.

Why it works: three named sources — birds, needles, stream — layer into a full forest bed synced to the drifting fog.

10. Rush-hour crossing

A busy downtown intersection at golden hour, crowds streaming across a wide crosswalk as buses and taxis wait. The camera holds a static elevated wide shot. Warm low sun flaring between the buildings, the audio a dense city bed of footsteps on pavement, honking horns, a bus engine idling, and the chirp of the pedestrian signal. Widescreen 16:9, one continuous shot.

Best for: title backgrounds and city establishers. Each sound is pinned to a thing you can see moving.

11. Corner cafe hum

The interior of a cozy corner cafe on a slow morning, steam rising from the espresso machine and a few patrons at scattered tables. The camera drifts in a gentle slow dolly across the room. Warm bulb light and soft window daylight, the ambience made of the espresso machine hissing, cups clinking on saucers, low murmured conversation, and a spoon stirring in a mug nearby. Portrait 4:5 format, a single unbroken take.

Why it works: the machine, cups, and spoon are all visible, so the café hum feels anchored rather than generic.

12. Ocean at dusk

A wide empty beach at dusk, waves rolling in and sliding back over dark wet sand as the tide turns. The camera stays locked low near the waterline. Deep violet-and-orange afterglow on the horizon, the audio carried by waves breaking and receding, foam fizzing on the sand, and gulls calling somewhere overhead. Vertical 9:16 format, one continuous shot.

Best for: calming vertical loops. The receding foam gives the wave sound a natural rhythm.

13. Thunderstorm on the porch

A wooden porch swing sways gently as a summer thunderstorm rolls in over the fields beyond the railing. The camera holds a static wide shot. Grey stormy light with occasional lightning flashes brightening the scene, the soundscape built from heavy rain on the porch roof, the creak of the swing's chains, rolling thunder, and wind gusting through the trees. Widescreen 16:9, one continuous shot.

Why it works: the creaking swing gives a foreground foley source that lifts the storm above a flat ambience bed.

Advertisement

Sound effects & ASMR

Foley reads best when it is tied to a visible action. Frame the hands or the object tightly, name the exact sound each motion makes, and keep the camera close so the picture and the crunch or pour land together.

14. Chopping on the board

Close on a pair of hands rhythmically slicing a red bell pepper on a wooden cutting board, the knife rocking through in even strokes and the pieces sliding aside. The camera stays in a tight overhead macro. Bright soft kitchen light, the audio built entirely from the crisp knife-through-pepper cuts, the blade tapping the board on each stroke, and the pieces brushing together. Portrait 4:5 format, one continuous shot.

Best for: cooking ASMR and recipe intros. The tap of blade on board syncs to every visible stroke.

15. Coffee pour close-up

A tight macro of dark coffee streaming from a gooseneck kettle into a clear glass carafe, the level rising and a swirl of crema forming on top. The camera holds steady on the pour. Warm side light catching the steam, the sound made of the steady pour splashing into the glass, the hiss of steam, and a faint drip as the stream cuts off at the end. Vertical 9:16 format, a single unbroken take.

Why it works: the pour is one continuous visible action, so the splashing audio has an obvious frame to follow.

16. Gravel footsteps

A low tracking shot following a pair of leather boots walking steadily along a gravel path in a garden, each step pressing and shifting the stones. The camera glides just behind the heels at ankle height. Soft overcast daylight, the audio driven by the crunch and scatter of gravel under each boot, with a light breeze and distant birdsong behind it. Widescreen 16:9, one continuous shot.

Best for: footstep foley and transitions. Every crunch lands on a footfall you can see.

17. Mechanical keyboard ASMR

A close overhead shot of fingers typing quickly across a mechanical keyboard on a dark desk, keycaps depressing in fast succession under a warm desk lamp. The camera stays locked in tight. Cozy low light with a soft glow on the keys, the audio carried by the crisp clack of each keystroke, the occasional deeper thock of the spacebar, and a soft click of the mouse at the end. Vertical 9:16 format, one continuous shot.

Why it works: naming the spacebar thock and mouse click gives the model distinct, visible-source sounds to place.

18. Crackling campfire

A close view of a small campfire at night, orange flames licking around split logs as embers drift upward and a log settles with a shower of sparks. The camera holds a steady close shot. Deep darkness beyond the firelight, the soundscape built from the crackle and pop of the burning wood, the low roar of the flames, and the hiss of a settling log. Portrait 4:5 format, a single unbroken take.

Best for: ambience loops and cozy backdrops. The settling log gives a punctuating foley moment.

19. Page-turn and pencil ASMR

A tight overhead shot of a hand turning the pages of an old book and then sketching soft lines with a pencil in the margin. The camera stays locked close on the paper. Warm lamp light and gentle shadows, the audio made of the crisp rustle of each page turning, the scratch of the pencil tip on paper, and a quiet tap as the pencil is set down. Portrait 4:5 format, one continuous shot.

Why it works: both sounds — page rustle and pencil scratch — are pinned to hand motions in frame, exactly what foley sync needs.

Music-synced moments

Describe the musical cue and the mood, and describe the motion hitting it — a swell on a reveal, a beat on a footfall. Do not name copyrighted tracks; name the feel and the timing the picture should match.

20. Piano swell on a reveal

An aerial shot rising slowly over a misty mountain valley, the fog parting to reveal a lake and distant peaks as the frame widens. Soft dawn light breaking gold across the ridges. A gentle solo piano begins quietly and builds to a single warm swell exactly as the lake comes into view, so the music peaks on the reveal. Widescreen 16:9, one smooth continuous shot.

Best for: cinematic openers. Tying the swell to the moment the lake appears locks music and motion together.

21. Beat-drop city run

A runner sprints across a neon-lit night crosswalk, the camera tracking alongside at a fast steady pace as light streaks past. Rain-slicked pavement reflecting magenta and cyan signs. An upbeat electronic track pulses under the run and hits a hard beat drop the instant the runner leaps a puddle, the motion timed to land on the drop. Vertical 9:16 format, one continuous shot.

Why it works: naming the exact moment the drop hits — the puddle leap — gives Muse a visible beat to sync to.

22. Waltz in an empty ballroom

A lone dancer in a flowing gown turns through slow sweeping steps across a grand empty ballroom, the camera orbiting gently around her. Chandeliers casting warm pools of light on the parquet floor. A soft three-four waltz plays, strings and piano, and her turns fall on the downbeat of each bar so the choreography rides the rhythm. Widescreen 16:9, a single unbroken take.

Best for: elegant brand and music moments. The waltz meter and her turns share one clear pulse.

23. Lo-fi study loop

A cozy desk by a rainy window at night, a cup of tea steaming beside an open notebook and a small plant, the camera drifting in an almost imperceptible slow push-in. Warm lamp glow against the cool blue rain outside. A relaxed lo-fi hip-hop beat plays softly, mellow and looping, in easy sync with the drifting steam and the rain streaking the glass. Vertical 9:16 format, one continuous shot.

Why it works: a calm loopable beat matched to gentle motion is the classic study-video feel, generated in one pass.

24. Trailer riser to hit

A lone figure in a long coat walks slowly toward the camera across a windswept dune as sand streams past, the camera pushing in low and steady. Harsh backlit sun flaring behind the silhouette. A deep orchestral riser builds tension throughout and lands on a single booming drum hit the moment the figure stops and lifts their head, the picture freezing on the hit. Widescreen 16:9, one continuous shot.

Best for: teaser and trailer beats. The riser resolving on the stop-and-look is a reliable music-sync structure.

Advertisement

Foley & voiceover / narration

Here a narrator's voice rides over b-roll — no lip-sync to worry about, so lines can breathe. Name the tone of voice and let ambience or foley sit softly beneath it. Six narrated openers. See more in the cinematic b-roll pack.

25. Nature documentary open

A slow aerial glide over golden savanna at dawn as a herd of elephants moves across the plain below, dust rising behind them. Warm low light stretching long shadows. A calm, resonant male narrator speaks over the shot in a measured documentary tone: "At first light, the herd begins to move." Soft wind and distant birdcalls sit beneath the voice. Widescreen 16:9, one smooth continuous shot.

Best for: documentary cold opens. Voiceover over b-roll sidesteps lip-sync entirely.

26. Product explainer voiceover

A clean tabletop shot as hands slide a sleek smartwatch out of its box and set it upright, the camera easing in on the face of the watch. Bright soft studio light and a minimal white surface. A friendly, upbeat female narrator says clearly over the b-roll: "Meet the watch that keeps up with you." A gentle click of the box and a subtle whoosh of motion under the line. Vertical 9:16 format, one continuous shot.

Why it works: naming the tone plus a light box-click foley keeps the ad voice forward and the visuals crisp.

27. Recipe how-to narration

An overhead shot of hands folding blueberries into pale batter with a spatula, the fruit streaking purple through the mix. Warm even kitchen light. A warm, encouraging narrator explains over the action: "Fold gently, just until it comes together." The scrape of the spatula against the bowl and the soft plop of batter sit quietly beneath the voice. Portrait 4:5 format, a single unbroken take.

Best for: recipe reels. The spatula scrape gives the narrated b-roll a real foley layer.

28. Travel vlog cold open

A handheld shot stepping out of a narrow alley into a sunlit Mediterranean plaza, the camera tilting up to the bright facades and fluttering laundry lines. Warm midday sun and deep blue sky. A relaxed, curious narrator says over the reveal: "Some mornings, you just follow the light." Footsteps on cobblestone, distant chatter, and a scooter buzzing past sit under the voice. Vertical 9:16 format, one continuous shot.

Why it works: the cobblestone footsteps and passing scooter give the narrated open a lively, sourced ambience.

29. Meditation voiceover

A still shot of dawn mist drifting over a calm lake, the water barely rippling as the sky slowly brightens, the camera in an almost imperceptible slow push-in. Soft cool light warming to pale gold. A slow, gentle narrator speaks over the scene in a hushed calming tone: "Breathe in. And let it go." Faint water lapping and distant birdsong sit far beneath the voice. Widescreen 16:9, one continuous shot.

Best for: wellness and meditation content. A hushed tone over barely-moving b-roll needs no lip-sync at all.

30. Brand manifesto b-roll

A series of quick b-roll beats — a welder's sparks, a designer's pencil on paper, a runner lacing a shoe — cut together with steady confident camera moves. Rich contrasty light across each shot. A determined, warm narrator delivers over the montage: "We build for the ones who keep going." The spark hiss, pencil scratch, and lace pull sit softly under the voice, each tied to its shot. Widescreen 16:9, a short continuous sequence.

Why it works: each foley sound is pinned to the beat that makes it, so the montage stays synced even under a single voiceover.

For the full method behind these — the prose formula, camera language, and audio line — read how to prompt Muse Video for realistic video, keep the cheat sheet open, and browse the whole best Meta Muse Video prompts roundup.

Frequently Asked Questions

How does native audio work in Meta Muse Video?

Meta Muse Video generates sound in the same pass as the picture, so dialogue, ambience, sound effects, and music are produced together rather than added afterward. To use it, describe the sound world inside your prompt and name which visible object or person makes each sound — the barista's steam wand hissing, rain drumming on the tin roof, a narrator's calm voice. Because the audio is tied to the frame, saying what makes the noise gives the model something on screen to sync to.

How do I keep lip-sync clean for dialogue?

Meta notes audio-to-video sync is still improving, so keep spoken lines short — a single sentence or two of a few words. Put the exact words in quotation marks, name the one speaker, and give them one visible action while they talk, like leaning forward or setting down a cup. Shoot the speaker in a steady framing rather than a fast camera move so the mouth stays readable. Do not promise perfect lip-sync in the prompt; write for a short, clear line and expect a light touch-up in an editor if you need frame-accurate timing.

How long should a spoken line be?

Short. One line of dialogue, or a two-line exchange at most, keeps the mouth movement and the audio aligned while sync is still maturing. Long monologues drift, so break a longer script into separate shots and stitch them, or move the words to a voiceover over b-roll where lip-sync is not a factor. As a rule of thumb, if a line takes more than a couple of seconds to say out loud, split it.

How do I describe sound effects versus music?

For sound effects, tie each sound to a visible action and its source: the knife tapping the board as it chops, water pouring into the glass, boots crunching on gravel. That gives Muse Video a frame to sync the foley to. For music, describe the cue and mood the motion should match rather than naming a copyrighted track — for example a slow warm piano building to a single swell, or an upbeat pop beat with a clear drop — and describe the motion hitting the beat so the picture and music feel locked together.

Can I generate a clip silent and add audio later?

Yes. Native audio is a strength, not a requirement. If you want full control over timing, dialogue, or licensed music, prompt for the visuals only, leave the sound world out, and add your own voiceover, foley, and score in an editor. But when you want a quick, self-contained clip — an ambience loop, an ASMR moment, a talking-head line — letting Muse generate the audio in the same pass saves a whole editing step.

What is Meta Muse Video and who makes it?

Meta Muse Video is Meta Superintelligence Labs' text-to-video model, announced on July 7, 2026 alongside Muse Image and built on the same pretraining base. It ranked third on the LMArena human-preference leaderboard for text-to-video at preview. Its standout edge is native synchronized audio — sound generated with the picture. It is rolling out to creators through the Meta AI app, meta.ai, Instagram Reels and Stories, and WhatsApp, and supports text-to-video, image-to-video, and video-to-video modes.

How do I write a good Muse Video prompt?

Write one flowing prose paragraph, not comma-tags or --flags and with no numeric settings line. Name the subject and appearance, give one clear action, add an explicit camera move, set the scene and lighting, then write the audio line — dialogue in quotes, ambience, or a music cue — naming which visible thing makes each sound. Finish by stating the aspect ratio in words, such as vertical 9:16 format in one continuous shot. Muse reasons over the whole brief before drawing, so detailed, ordered prose lands closer.

How is native audio different from Muse Image?

Muse Image is a still-image model with no sound; Muse Video adds motion and generates the audio track in the same pass as the frames. Both share a pretraining base and both reward long, structured prose prompts, so the writing style carries over — full sentences, aspect ratio in words, no parameter flags. The new part is the audio line: in Muse Video you also describe the sound world and its on-screen sources, which has no equivalent in the image model. Compare the still-image side in the best Meta Muse Image prompts.

Advertisement