FLUX 3 Max Camera Movement: How to Phrase Every Move
How to write FLUX 3 Max camera movement: push-in, orbit, dolly, crane, handheld, locked-off and whip pan, with framing words, copy-paste prompts and which ad each move suits.

FLUX 3 Max camera movement is written in plain words inside the prompt, after the subject and action: one framing and one movement per shot, such as "medium close-up, slow push-in". This guide covers how to phrase the push-in, orbit, dolly, crane, handheld, locked-off and whip pan, with copy-paste prompts and which ad each move suits.
The camera is the second of the four parts in every prompt (subject and action, camera, light and look, sound), as set out in the FLUX 3 Max prompt guide. Its developers describe FLUX 3 Max as tuned for stronger prompt adherence, and the camera sentence is one of the easiest places to check that claim in your own drafts: did the move happen, in the direction and at the speed you wrote?
How do you prompt FLUX 3 Max camera movement?
Give the camera its own short sentence, straight after the action. Four habits keep it readable:
- One framing, one movement. "Close-up, slow half-orbit." Not three moves chained together.
- A speed word. Slow, steady or quick. Without one, the model picks a pace for you.
- A direction and a target. "Pushes in toward the label", "tracks left to right alongside her", "cranes up to reveal the coastline". A move with a destination is easier to follow than a bare verb.
- A move that fits the length. A 5 second clip holds a short push-in or a single whip pan. A slow crane or a half-orbit wants 10 seconds or more.
If you leave the camera out, FLUX 3 Max expands the prompt and chooses a shot itself. That is fine for exploring, but write the camera whenever framing matters to the ad.
Which framing words should you pair with a move?
The framing tells the model where the move starts. Pair one of these with one movement.
| Framing | What it shows | Typical use |
|---|---|---|
| Extreme close-up | A detail: a drop, a stitch, an eye | Texture and product detail |
| Close-up | A face, or a product filling the frame | Hero shots, reactions |
| Medium close-up | Head and shoulders | Talking to camera, UGC |
| Medium shot | Waist up, with some setting | Demos, presenters |
| Wide shot | The whole person and the place | Lifestyle, establishing |
| Top-down | The surface from directly above | Unboxing, food, flat lays |
| Low angle | Looking up at the subject | Footwear, power, scale |
| Eye level | The viewer's own height | Anything that should feel natural |
How do you phrase each camera move?
Each move below has what it does, the phrasing that reads cleanly, and a full prompt. Costs use the standard FLUX 3 Max rates of 17 credits per second at 720p and 29 credits per second at 1080p.
Push-in and pull-out
A push-in moves the camera toward the subject, so whatever it approaches grows in importance: a face as the line lands, a label, a product detail. A pull-out does the reverse and reveals context. Phrase it as "slow push-in toward her face" or "slow pull-out from the bottle to reveal the kitchen".
A woman in her thirties looks straight into the camera across a kitchen
table, a half-finished coffee in front of her. Medium close-up, slow push-in
toward her face as she speaks. Soft window light from the right, warm natural
grade. She says, "I almost cancelled. Then I read the second page." Quiet
kitchen tone, a chair creak, no music.
9:16, 8 seconds, 720p: 136 credits.
Orbit and half-orbit
An orbit circles the subject and shows its shape, which is why it is the default move for product hero shots. A half-orbit is usually enough and keeps the move calm. Phrase it as "slow half-orbit from left to right around the bottle".
A matte black smartwatch rests on a round slate plinth as a thin line of
light travels across its face. Close-up, slow half-orbit from left to right
around the plinth. Hard top light on a black background, deep blacks, crisp
highlights. A low electronic hum, a soft metallic tick, no voice.
1:1, 10 seconds, 1080p: 290 credits.
Dolly and tracking
A dolly moves the whole camera forward, back or sideways, and a tracking shot follows a moving subject. Both feel smooth and deliberate. Phrase it as "steady tracking shot alongside her, moving left to right" or "slow dolly forward through the doorway".
A chef in a white apron walks the length of a busy restaurant kitchen
carrying a plated dish, passing cooks at their stations. Medium shot, steady
tracking shot alongside her at shoulder height, moving left to right. Warm
overhead kitchen lights, steam in the air, rich warm grade. Pans sizzling on
the left, a ticket printer chattering, indistinct voices calling orders, no
music.
16:9, 12 seconds, 1080p: 348 credits.
Crane up and crane down
A crane lifts or lowers the camera. Craning up over a subject reveals the world around it; craning down lands on a detail. Phrase it as "slow crane up from road level to a high view of the coastline".
A lone cyclist rides along an empty coastal road at sunrise, small in the
frame. Wide shot, slow crane up from road level to a high view of the
coastline as the cyclist rides on. Low golden sun on the left, long shadows,
soft film grade. Wind across open ground, waves breaking far below, the faint
whir of the bike chain, a slow swelling string score.
21:9, 15 seconds, 1080p: 435 credits. Slow, wide moves like this suit 21:9, the widest ratio in the editor.
Handheld
Handheld is the creator look: small natural shake, eye level, often selfie framing. Polish breaks UGC, so this is the move for testimonials, reviews and street content. Phrase it as "handheld selfie framing, eye level, natural shake as he moves".
A man in his twenties in a denim jacket holds a small portable speaker up to
the phone camera on a crowded festival field, then turns it toward the stage
behind him. Handheld selfie framing, eye level, natural shake as he moves.
Late afternoon sun, hazy warm grade. He says, "Survived three days of this.
Still going." Crowd noise, distant bass from the stage, wind on the mic.
9:16, 10 seconds, 720p: 170 credits. Fill-in templates built on this move are in FLUX 3 Max UGC prompt templates.
Locked-off
A locked-off camera does not move at all. Use it when the action is the point: a pour, a demo, a transformation on a surface. It is also the cleanest base for adding captions in the editor. Phrase it as "top-down, locked-off" or "locked-off medium shot".
Two hands pour thick honey from a glass jar onto a stack of pancakes on a
white plate, and the honey slowly spreads over the edges. Top-down,
locked-off. Soft morning light from a window above, warm clean grade. A slow
sticky drip, a fork set down beside the plate, quiet kitchen tone, no music.
4:3, 8 seconds, 1080p: 232 credits.
Whip pan
A whip pan swings the camera so fast the frame blurs. Inside one clip it snaps attention from one subject to another in the same place. Phrase it as "quick whip pan from the barista to the customer".
A barista sets a finished latte on the counter, then the camera whips right
to a customer reaching for it with a grin. Handheld, quick whip pan from the
barista to the customer. Bright cafe light, warm natural grade. The clink of
the cup on the counter, a whoosh on the pan, cafe chatter, no music.
9:16, 5 seconds, 720p: 85 credits.
A whip pan also works as a transition between two clips. Finish clip A on a whip pan to the right, open clip B moving out of a whip pan in the same direction, and join them in the editor. The cut hides inside the blur.
Why does stacking camera moves fail?
Every extra move competes for the same few seconds, and moves can contradict each other: a camera cannot be locked-off and pushing in, or handheld and gliding on a smooth track. Pick the one move that serves the beat. If the idea genuinely needs two moves, it is usually two clips.
| Stacked | Cleaner |
|---|---|
| Handheld crane orbit push-in | Slow half-orbit |
| Locked-off with a slow zoom and pan | Locked-off, or a slow push-in |
| Fast whip pan, then slow dolly, then crane up | Three clips, one move each |
| Smooth cinematic handheld gimbal | Steady tracking shot, or handheld if it should feel raw |
How does FLUX 3 Max camera movement change in image to video?
With one reference image attached, your image becomes the first frame. The move therefore starts from the photo's framing, and the prompt should describe only motion and sound, because the image already carries the look:
The bottle slowly rotates on the marble counter as condensation runs down the
glass. The camera makes a slow half-orbit. A crisp cap twist, a soft fizz,
light lo-fi music.
Three things to check in the photo before you choose a move:
- Room for the move. A push-in needs a sharp subject to push toward. A crane up needs headroom above it.
- What the photo cannot show. An orbit reveals sides the image does not contain, which have to be generated. Prefer a half-orbit or a slight arc, and check the product's back and label before the clip runs.
- The ratio. On auto the clip follows your image; a named ratio overrides that.
Product-specific motions to pair with these moves are in FLUX 3 Max product video prompts.
Which FLUX 3 Max camera movement fits which ad?
| Ad type | Move | Why |
|---|---|---|
| UGC testimonial or review | Handheld, eye level | Reads as filmed by a person, not a studio |
| Product hero shot | Slow half-orbit or push-in | Shows shape, or makes one detail matter |
| Demo or pour | Locked-off | The action is the point |
| Lifestyle or walk-through | Tracking or dolly | Moves with the subject through a place |
| Brand film opener | Crane up, wide, often 21:9 | Reveals scale and setting |
| Fast-cut social edit | Whip pan | Energy, and a built-in transition |
In vertical placements, keep the subject inside the safe zone through the whole move. TikTok's in-feed ad specifications tell advertisers to check the safe zone, and Meta's Reels ads page advises keeping the main creative elements inside it. A push-in that ends with the product under the caption bar has wasted the move. More wide-format ideas are in FLUX 3 Max cinematic brand videos.
How do you test a camera move without wasting credits?
Test the move short and cheap, then generate the real length. A 5 second draft at 720p is 85 credits, enough to see whether the push-in pushes or the orbit orbits. Once the move behaves, generate the delivery clip: 15 seconds at 1080p is 435 credits.
Change only the camera sentence between drafts, so you learn what the wording did. Plan allowances are on the pricing page.
Frequently Asked Questions
How do you prompt camera movement in FLUX 3 Max?
Write the camera as its own short sentence after the subject and action: one framing, one movement, a speed word and a direction, such as close-up, slow half-orbit from left to right. If you leave it out, the model chooses a shot for you.
How many camera movements should one prompt have?
One. Extra moves compete for the same seconds and often contradict each other. If an idea needs two moves, generate two clips with one move each and join them in the editor.
What is the difference between a dolly and a zoom?
A dolly physically moves the camera toward or away from the subject, so the perspective and background shift. A zoom changes the lens focal length from a fixed position, which flattens the image. In prompts, push-in or dolly forward describes the first.
How do I get a handheld UGC look in AI video?
Write handheld selfie framing, eye level and natural shake, set the clip to 9:16, use a real room with window light, and keep the sound to room tone and the line with no music. Avoid orbits and smooth tracking, which read as studio work.
Why does the AI ignore my camera direction?
The usual causes are stacked or contradictory moves, a move too long for the clip duration, and a camera sentence buried at the end of a long prompt. Keep one move, give it a speed and a direction, and place it right after the action.
Can FLUX 3 Max shoot in 21:9 for cinematic moves?
Yes. The VIDEO AI ME editor offers 21:9 alongside auto, 9:16, 16:9, 1:1, 4:3 and 3:4, at 720p or 1080p. Slow, wide moves such as a crane up or a lateral tracking shot suit the format.
Frequently Asked Questions
Share
AI Summary

Paul Grisel
Paul Grisel is the founder of VIDEOAI.ME, dedicated to empowering creators and entrepreneurs with innovative AI-powered video solutions.
@grsl_frReady to Create Professional AI Videos?
Join thousands of entrepreneurs and creators who use VIDEO AI ME to produce stunning videos in minutes, not hours.
- Create professional videos in under 5 minutes
- No video skills experience required, No camera needed
- Hyper-realistic actors that look and sound like real people
Get your first video in minutes
Related Articles

FLUX 3 Max Audio: Sound and Dialogue Prompt Guide
FLUX 3 Max generates sound on every clip. How to write dialogue it will speak, direct ambience and effects, decide on music, and replace the track when you need to.

FLUX 3 Max Aspect Ratios and Durations Explained
Every FLUX 3 Max aspect ratio and clip length on VIDEO AI ME, how auto behaves in each mode, which placement wants which shape, and what each duration costs.

FLUX 3 Max 1080p vs 720p: When 1080p Is Worth It
FLUX 3 Max costs 17 credits per second at 720p and 29 at 1080p. When the higher tier earns its rate, when 720p is enough, and how to test cheap and deliver sharp.