AwakePix · AI video prompts

Video prompts, built around your scene.

Explore 37 editable prompts across 6 categories. Choose a scene, adapt the subject and inputs, then use the model that fits your idea.

Product motion from one reference image: real exampleJump to real exampleA cat by the window: real exampleJump to real example

How to use these prompts

  1. 01 · Start with the photo
    Use a clear image you have the right to use, with the subject fully visible. Describe what is already in the frame.
  2. 02 · Plan the rhythm
    Set the scene goal first, then use the selected model's supported duration to pace it. Let one movement breathe or pause; choose a suitable model and duration for staged action.
  3. 03 · Review, then generate
    The button loads a prompt; it does not generate or spend credits. Upload your photo, then review the available model, settings and price.

Workflow: define the scene goal, then choose an available model that supports the duration, reference-image count, camera approach and audio options you need. Let one movement breathe or pause; do not force staged action into one short clip. The entry only loads a prompt; it does not select a paid model or generate automatically.

Prepare the images or references described on the card, then choose a model. Use the generator for supported durations, reference counts and audio options. Use an external editing tool for assembly, voiceover or precise beat alignment, following the card notes. Read the model guide

01

Structure & references

Treat a timeline as a plan; fit each beat to the selected model.

Generate a clip

Timeline beats

Plan visible actions in ordered beats when timing matters.

A woman places a handwritten postcard on a kitchen table, looks toward the rainy window, writes one final word, folds the card, and slides it beside a warm cup of tea. Start with a wide view, move to her hands, then hold on the postcard as the room settles into soft window light. Keep her clothing, table, cup, handwriting, and rainy background consistent; no extra hands, skipped actions, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a reference with a clear subject, location, and intended opening state.
What to change
The generator can receive the resulting prompt, but making a storyboard sheet or assembling separate clips happens outside this flow.
Watch for
Exact beat timing and transitions can drift; a single generation is not an edit timeline.
Use this prompt

Prepare ordered references

Identity lock

Separate what each reference supplies from what it must not change.

Keep the person from the supplied references consistent while she turns from a window, picks up a blue notebook, speaks one short line, and walks three steps into the room. Preserve her face, hair, body proportions, clothing, accessories, and distinctive marks from the matching reference only. Keep the background, pose, lighting, text, and composition from the scene reference; no cloned, merged, duplicated, or swapped attributes.

Inputs and workflow
Photo to use
Give each reference a distinct role and show the identity details you need to preserve.
What to change
Choose a model that accepts multiple reference images before using this workflow; for a single-image model, merge the identity details into one reference or keep only the clearest image.
Watch for
Unseen angles, heavy occlusion, and fast motion can still change facial or clothing details.
Use this prompt

Prepare a storyboard image

Storyboard bridge

Connect planned panels into a coherent video prompt.

Follow the supplied storyboard panels in order: a child finds a red kite, runs across a grassy hill, releases it into the wind, and watches it rise above the trees. Carry the same child, kite, hill, and late-afternoon light from panel to panel. Make each visible action lead naturally into the next framing; do not add panels, skip panels, redraw annotations, or introduce extra characters.

Inputs and workflow
Photo to use
Use a clean storyboard or panel reference with numbered order and readable subject placement.
What to change
Creating the storyboard image and combining multiple generated clips require separate external steps.
Watch for
A storyboard reference does not guarantee exact panel timing, continuity, or readable text.
Use this prompt

02

Realism & UGC

Use plausible camera movement and give reactions time to settle.

Generate a clip

Curious head tilt

A small expressive movement for a relaxed pet portrait.

One continuous shot: the pet blinks once, tilts its head slightly with gentle curiosity, then returns to a calm pose. Keep the pet's identity, markings, eyes, and position consistent; keep the camera still.

Inputs and workflow
Photo to use
Use a sharp, front-facing pet photo with the eyes visible and a stable background.
What to change
For a side profile, replace the head tilt with one slow ear twitch and a blink.
Watch for
Fur edges, whiskers, eyes, and collars can warp during the facial movement.
Use this prompt

Generate a clip

Gentle tail wag

A contained tail motion for a standing or seated pet.

One continuous shot: the pet stays seated and gives its tail two gentle side-to-side wags, with a small natural blink. Keep the body, face, paws, and background still; use a locked camera.

Inputs and workflow
Photo to use
Choose a full-body image with the tail separated from the body and a little space behind it.
What to change
If the tail is hidden, use a slow ear movement or a small chest rise instead.
Watch for
A hidden or cropped tail cannot be reconstructed reliably from the source image.
Use this prompt

Generate a clip

Sunlit curious sniff

A tiny nose movement for a calm outdoor pet photo.

One continuous shot: the pet leans its nose forward slightly as if catching a scent, pauses, and gives one soft blink. Keep the pose grounded, preserve the markings, and let only the nearby grass move faintly in the breeze.

Inputs and workflow
Photo to use
Use an outdoor image with the face in focus and visible foreground space near the nose.
What to change
For an indoor image, replace the grass movement with a soft shift of window light.
Watch for
The generated sniff is an interpreted gesture and may affect the muzzle or mouth shape.
Use this prompt

Generate a clip

Clouds drifting overhead

A calm sky movement for a wide landscape still.

One continuous shot: soft clouds drift slowly across the sky while the landscape stays still and the camera remains locked. Keep the horizon, buildings, trees, and colors stable; use gentle natural light.

Inputs and workflow
Photo to use
Use a wide landscape with clear sky area and a distinct horizon.
What to change
If there is little sky, use a slight movement in distant haze or treetops instead.
Watch for
The model may alter cloud edges or distant details while inventing depth.
Use this prompt

Generate a clip

Quiet water ripples

A localized ripple for lakeside and coastal scenery.

One continuous shot: small concentric ripples travel once across the visible water surface while the shoreline and camera remain still. Preserve reflections, horizon, and overall color; keep the motion calm and minimal.

Inputs and workflow
Photo to use
Choose a scene with a visible water surface, stable shoreline, and readable reflections.
What to change
For a waterfall or ocean image, use a single slow flow direction instead of concentric ripples.
Watch for
Reflections and water boundaries may shimmer or shift beyond the intended ripple.
Use this prompt

Generate a clip

Slow cinematic push-in

A restrained camera move that adds depth to a still scene.

One continuous shot: the camera makes a very slow, smooth push-in toward the main landmark with no pan or tilt. Keep the landmark's shape stable, preserve the composition and lighting, and avoid introducing new objects.

Inputs and workflow
Photo to use
Use a sharp image with one clear focal landmark and foreground depth cues.
What to change
For a close-up, reduce the push-in to a subtle parallax shift around the subject.
Watch for
The generated depth is inferred from one view and may cause edges or repeated patterns to warp.
Use this prompt

Generate a clip

Handheld UGC vlog

A candid phone-recorded moment with believable camera flaws.

Make this feel like an authentic handheld phone vlog: imperfect framing, mild arm drift, natural shake, brief autofocus hunting, ordinary available light, and unpolished exposure. Keep the person and setting recognizable. Let the action develop through a simple glance, gesture, or reaction. Use diegetic room or street sound only; no cinematic crane moves, beauty filter, subtitles, or watermark.

Inputs and workflow
Photo to use
Choose a casual photo with natural light and enough background context to support a phone-recorded moment.
What to change
If the source is polished, reduce the cinematic language and describe only believable phone imperfections.
Watch for
Generated camera flaws may become distracting or change fine facial and background details.
Use this prompt

Generate a clip

First-person POV

Derive camera motion from the wearer's body and point of view.

Use a first-person body-mounted view. Derive the camera movement from the wearer's head, chest, bicycle, or handlebar: small body sway, natural forward motion, and brief readjustments. Keep the hands, tools, and immediate path coherent as the subject performs one clear task. Preserve real scale and contact; no floating camera, third-person operator, teleportation, or artificial text.

Inputs and workflow
Photo to use
Use a reference showing the hands, equipment, or path that should define the point of view.
What to change
A static portrait cannot supply a full POV rig; keep the movement modest and grounded in visible anatomy.
Watch for
Occluded hands, tools, and unseen surroundings may deform during rapid movement.
Use this prompt

Generate a clip

Early-2000s DV

A small everyday incident framed as old camcorder footage.

Render the scene as early-2000s home-video footage: handheld camcorder shake, autofocus hunting, slight exposure pumping, modest interlacing-like softness, consumer microphone ambience, and imperfect white balance. Keep the incident small and ordinary, with one clear action and a natural reaction. Avoid modern cinematic grading, smooth gimbal motion, added captions, and watermarks.

Inputs and workflow
Photo to use
Use a candid image with an ordinary setting and enough texture for consumer-camera artifacts.
What to change
Add the period camera treatment in the prompt; do not rely on post-processing or a separate filter step here.
Watch for
The model may imitate the look without reproducing consistent frame-level analog artifacts.
Use this prompt

Generate a clip

Pet takes the lead

Let one animal drive a phone-shot moment with natural behavior.

Keep exactly one animal as the lead throughout the shot. It behaves like a real animal: noticing the camera, approaching with curiosity, pausing, and briefly filling the lens without speaking or walking like a person. Use handheld phone framing, natural autofocus, and diegetic household or outdoor sound. Keep fur pattern, eyes, ears, paws, and count consistent; no duplicate reflections, costumes, subtitles, or watermark.

Inputs and workflow
Photo to use
Use one clear animal reference with its full markings and a visible path toward the camera.
What to change
For a calm pet portrait, replace the approach with a small natural gesture and keep the phone camera steady.
Watch for
Paws, whiskers, eye direction, and the final close-up can drift when the animal fills the frame.
Use this prompt

03

Commercial & portraits

Plan a reveal, a detail and a closing view around the product.

Generate a clip

Slow turntable reveal

A restrained rotation for a centered product hero shot.

One continuous shot: the product makes one slow, smooth quarter-turn on an invisible turntable while the camera stays locked. Keep the product centered, preserve its shape and branding, and use steady studio light with a soft natural shadow.

Inputs and workflow
Photo to use
For a handle or an asymmetric outline, use a sharp three-quarter product photo with the full object visible. For an unmarked, symmetric product, a centered front view works well.
What to change
If the object has a handle or an asymmetric outline, keep the quarter-turn small so its identity stays visible. For an unmarked, symmetric product, replace the rotation with a very slow push-in and keep the layout fixed.
Watch for
Fine logos, text, thin edges, and unseen sides may change during motion.
Use this prompt

Generate a clip

Packaging lid lift

A compact opening gesture for a packaging still.

One continuous shot: the box lid lifts a few centimeters and settles, revealing a small glimpse of the product inside. Keep the box in place, keep the camera locked, preserve the printed design, and use soft directional light.

Inputs and workflow
Photo to use
Choose a front-facing box photo with a clear lid seam and enough empty space around it.
What to change
If the lid is not visible, use a gentle label shimmer or a short camera push-in instead.
Watch for
The hidden interior and exact hinge mechanics are inferred from the single image.
Use this prompt

Generate a clip

Fabric caught by a breeze

A subtle textile movement for apparel or home goods.

One continuous shot: a light breeze passes once across the fabric, creating a gentle ripple that fades while the product remains in place. Keep the camera still, preserve the color and weave, and avoid adding hands or extra objects.

Inputs and workflow
Photo to use
Use a well-lit image where the fabric surface and its edges are clearly visible.
What to change
For rigid products, use a small light sweep across the surface rather than fabric motion.
Watch for
The direction and strength of the breeze are visual approximations, especially on layered fabric.
Use this prompt

Generate a clip

Natural breath and blink

A restrained portrait loop with almost no pose change.

One continuous shot: the person takes one subtle breath, blinks once, and makes a barely perceptible relaxed shift of the shoulders. Keep identity, hairstyle, clothing, framing, and background stable; use a locked camera.

Inputs and workflow
Photo to use
Use a sharp portrait with the face unobstructed, relaxed shoulders, and even lighting.
What to change
For a close crop, omit the shoulder shift and keep the movement to breathing and blinking.
Watch for
Small facial details and teeth may change during the blink or breath.
Use this prompt

Generate a clip

Look toward the light

A gentle eye-line change for a window-lit portrait.

One continuous shot: the person slowly shifts their gaze toward the window light, turns their head only a few degrees, and settles. Keep the expression calm, preserve identity and clothing, and keep the camera fixed.

Inputs and workflow
Photo to use
Choose a three-quarter portrait with a clear face orientation and visible directional light.
What to change
If the face is straight-on, use only an eye-line shift and skip the head turn.
Watch for
A large head turn may reveal unseen features and reduce identity consistency.
Use this prompt

Generate a clip

Small candid smile

A modest expression change for a friendly everyday portrait.

One continuous shot: the person relaxes into a small closed-mouth smile, blinks once, and holds the expression. Keep the head position, face shape, hair, clothing, and background steady; use a locked camera.

Inputs and workflow
Photo to use
Use a well-lit portrait with a neutral or already-friendly expression and the mouth clearly visible.
What to change
For an existing smile, use a tiny cheek lift and blink instead of a new expression.
Watch for
Teeth, lips, and asymmetrical expressions are especially prone to visual drift.
Use this prompt

Audio editing may be needed

Creator product review

A creator unboxes, handles, and explains a product on camera.

Show a creator reviewing the product naturally: introduce it, pick it up, demonstrate one visible feature, and finish with a measured reaction. Keep the creator's face, hands, clothing, packaging, logo placement, and product shape consistent. Write short spoken lines in the requested language and align each line with the matching gesture. Use room ambience; no invented claims, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a clear creator-and-product reference where face, hands, package, and label are visible.
What to change
Exact spoken audio or subtitle timing may need recording and editing after generation.
Watch for
Lip shape, small label text, and detailed product claims may drift; review every spoken line.
Use this prompt

Generate clips, then edit

Cinematic product commercial

Build a polished product reveal from a controlled set of shots.

Create a polished product commercial with a clear visual arc: establish the setting, reveal the product, show one material or use detail, then end on a clean hero frame. Keep product geometry, label placement, lighting direction, and color palette consistent. Use deliberate camera moves and restrained sound design. Avoid invented text, warped logos, extra products, and watermarks.

Inputs and workflow
Photo to use
Use a sharp product reference with the label and key silhouette unobstructed.
What to change
A multi-shot commercial may require separate generations and external editing; this prompt does not assemble them automatically.
Watch for
Continuity between separate shots and exact hero-frame typography require post-production review.
Use this prompt

Generate clips, then edit

Process transformation

Show a material, craft, recipe, or space changing in ordered stages.

Show the transformation in clear visible stages. Start from the supplied state, preserve the workspace and key materials, and let each stage change one physical property before the next. Keep scale, direction, and continuity believable; end with a readable finished state. Avoid instant unexplained swaps, extra hands, floating tools, duplicated objects, text, and watermarks.

Inputs and workflow
Photo to use
Use a reference with the starting material, workspace, and intended change visible.
What to change
For many stages, plan separate clips or a storyboard externally rather than forcing every step into one generation.
Watch for
Intermediate states, exact measurements, and hands-on continuity may be invented or skipped.
Use this prompt

Generate clips, then edit

Fashion lookbook

Keep one look coherent while locations, poses, and light change deliberately.

Present one model and one complete look with editorial restraint. Lock face, hair, body proportions, garments, footwear, accessories, and colors. Use a distinct location, gesture, and light direction for each planned moment while preserving the outfit. Keep the camera elegant and the pose readable; no garment mutations, extra models, unreadable branding, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a full-body or three-quarter reference where the outfit and key accessories are visible.
What to change
A lookbook across locations is better treated as multiple clips with external editorial assembly.
Watch for
Fine fabric patterns, jewelry, footwear, and face continuity can drift during pose changes.
Use this prompt

Making a product video: settings and practical limits

04

Story & performance

Give each beat a clear result; split longer stories into clips.

Audio editing may be needed

Dialogue performance beats

Tie short spoken lines to visible reactions and causal performance.

Stage a short exchange: Maya says, "The last train is leaving," and Leo looks at the clock before replying, "Then we should run." Show Maya pointing toward the station, Leo grabbing his coat, and both turning toward the door. Keep speaker identity, turn-taking, eye lines, and emotional cause clear with natural room tone and pauses; no overlapping lines, invented subtitles, mouth duplication, identity swaps, or watermark.

Inputs and workflow
Photo to use
Use references that make each speaker's face, position, and relationship legible.
What to change
Exact voice recording, lip-sync correction, and subtitle timing may need audio or editing work after generation.
Watch for
Long lines, turn-taking, and fine lip motion can be inconsistent; keep speech short and review it.
Use this prompt

Generate clips, then edit

Narrative short

Give a compact story a visible setup, turn, and ending state.

Follow a traveler through a quiet station: she checks a map, opens an empty envelope, notices a brass key taped inside the flap, and looks toward a locked blue door at the end of the platform. Show the setup, discovery, and final look in clear visual actions. Keep her coat, red bag, station geography, and evening light coherent; avoid gore, teleportation, extra characters, subtitles, and watermark.

Inputs and workflow
Photo to use
Use a reference that establishes the protagonist, setting, and the prop that carries the turn.
What to change
Multi-act stories usually need separate clips and external editing; this entry does not create a complete film timeline.
Watch for
Causal continuity, exact pacing, and complex reveals may be simplified or lost in one generation.
Use this prompt

Generate clips, then edit

Travel montage

Connect place, movement, and a small travel observation across scenes.

Build a travel sequence around one traveler and a clear sense of place. Move through distinct locations with a readable action in each: arriving, observing, walking, or sharing a small local detail. Keep the traveler, bag, wardrobe, and visual tone coherent. Use ambient location sound and restrained camera movement; no invented landmarks, text, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a reference with the traveler and a recognizable location cue in frame.
What to change
A location-to-location montage requires multiple clips and external assembly; one prompt cannot guarantee every stop.
Watch for
Unseen landmarks, signage, and geographic continuity may be invented or altered.
Use this prompt

Generate a clip

Twist comedy

Work backward from one visible, nonviolent punchline.

A serious office worker places a sealed envelope on a desk, opens it with growing concern, pauses, and discovers a single cheerful birthday sticker inside. Hold on the awkwardly delighted reaction before a coworker quietly slides over a cake. Keep the office, faces, envelope, and cake consistent; no gore, injury, teleportation, duplicated people, subtitles, or watermark.

Inputs and workflow
Photo to use
Use an ordinary location and a visible object that can carry the final reveal.
What to change
The pause and final reaction may need external editing if exact comedic timing matters.
Watch for
The model may reveal the twist early or change the object; keep the gag visually simple.
Use this prompt

Generate clips, then edit

Nonviolent suspense

Build unease through a chain of small visible changes, without gore.

Create nonviolent suspense from small visible changes: a quiet room, a noticed detail, a changed object, a cautious approach, and an unresolved final state. Keep the protagonist, location, light direction, and geography consistent. Let tension come from what is seen and withheld; no gore, wounds, attacks, explicit violence, graphic threat, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a reference with a readable space, one suspicious detail, and enough negative space for tension.
What to change
A longer suspense chain may need separate clips and editing; the generator does not add unseen coverage automatically.
Watch for
Suspense depends on continuity and pacing that can drift; keep changes concrete and review for unintended violence.
Use this prompt

05

Animation & style

Keep the drawing or material style consistent as the action develops.

Prepare ordered references

Anime style lock

Preserve a measurable illustrated look while the subject moves.

Animate the supplied character in a consistent anime-inspired illustrated style. Lock line quality, cel-shaded value blocks, eye construction, hair shapes, palette, and background rendering. Describe the action with clear poses and camera direction. Keep the character readable and avoid drifting into generic 3D, photorealism, painterly blur, copied logos, subtitles, or watermark.

Inputs and workflow
Photo to use
Choose a model that accepts multiple references, then upload up to its documented limit in order; Seedance 2.0 Fast supports up to four. Use clean character references with visible linework, palette, face, and silhouette.
What to change
For a single-image model, merge the character and style cues into one reference or keep only the clearest image; use an original character instead of naming a copyrighted one.
Watch for
Complex hands, hair strands, and perspective changes may drift from the reference style.
Use this prompt

Generate a clip

Stop-motion cadence

Use stepped movement, holds, and material detail instead of smooth motion.

Create a handcrafted stop-motion scene with visible stepped poses and deliberate holds. Preserve the material identity of the clay, paper, felt, or miniature set. Move one object at a time with small physical changes, fixed lighting, and a stable table-level camera. Keep seams and handmade texture; no fluid interpolation, floating pieces, modern camera glide, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a miniature or handmade reference where material texture and the set boundary are visible.
What to change
Frame rate, hold count, and final assembly may need to be controlled outside the generator.
Watch for
The model may smooth stepped motion or invent material details between poses.
Use this prompt

Generate a clip

3D cartoon character

Give one expressive 3D character a readable physical action.

Animate one original 3D cartoon character through a small physical action with a clear start, contact, and settle. Keep its face, body proportions, costume, materials, and color palette stable. Use readable squash-and-stretch without losing object identity, with soft studio light and a simple camera move. No extra limbs, duplicate character, texture flicker, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a full character reference with silhouette, face, costume, and material cues visible.
What to change
For a sequence of actions, plan separate clips or a storyboard rather than assuming automatic shot assembly.
Watch for
Facial expressions, hands, and fine 3D materials can change during squash-and-stretch.
Use this prompt

06

Action & effects

Describe movement in causal order; reserve exact timing for editing.

Generate a clip

Nonviolent combat choreography

Plan readable contact, defense, and recovery without graphic harm.

Choreograph a nonviolent screen fight with two clearly distinct performers. Show stance, readable attack direction, defensive contact, balance recovery, and a safe ending. Keep bodies, costumes, camera geography, and contact points consistent. Use controlled impact without wounds, blood, broken bones, or graphic injury. No duplicate fighters, teleportation, floating limbs, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a wide reference where both performers, their spacing, and the safe environment are visible.
What to change
A complex fight should be planned as panels or clips and assembled externally; do not expect automatic stunt editing.
Watch for
Fast contact, hands, and occlusion can distort bodies; review for unintended injury imagery.
Use this prompt

Audio editing may be needed

Music video beat plan

Organize movement and cuts around a musical pulse without promising exact sync.

Stage a music-video performance around clear musical pulses: define the performer, formation, camera move, and visible gesture for each planned beat. Keep the lead performer centered and preserve costume, lighting, and set continuity. Use energetic but readable motion, with no random extra dancers, unreadable text, face drift, or watermark.

Inputs and workflow
Photo to use
Use a reference with the lead performer and enough surrounding space for movement or formation.
What to change
Exact BPM cuts, beat markers, and final music edit require an audio track and external post-production.
Watch for
Text prompts cannot guarantee frame-accurate beat sync or consistent group choreography.
Use this prompt

Generate a clip

Freeze and rewind

Use a clear physical event, suspended peak, and readable return.

Build a physical set piece in five readable states: normal motion, the initiating contact, a freeze at the peak, a controlled camera observation, and a precise return to motion. Keep one chosen subject exempt from the freeze while the environment remains suspended. Preserve object positions and gravity; no teleportation, duplicated objects, injury, gore, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a reference with a clear impending contact, suspended objects, and enough depth for camera observation.
What to change
A precise rewind or freeze timing may need panel planning and external editing instead of one automatic pass.
Watch for
Object trajectories, gravity, and the exempt subject may not remain exact through the reversal.
Use this prompt

Generate a clip

Vehicle and camera action

Keep one vehicle coherent while the camera supplies the energy.

Keep the vehicle as one continuous machine: preserve body panels, wheels, lights, paint, and scale. Let the camera create energy with a controlled tracking move, low angle, or safe orbit while the vehicle follows one believable path. Keep road contact, reflections, and surrounding traffic coherent. No impossible wheel changes, extra cars appearing, collisions, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a clear vehicle reference with wheels, distinctive panels, and its travel surface visible.
What to change
A cut-heavy driving sequence needs multiple clips and external assembly; this prompt does not create a shot list automatically.
Watch for
Fast motion can warp wheels, reflections, road markings, and distant traffic.
Use this prompt

Prepare ordered references

Science-fiction reveal

Reveal scale through grounded objects, lighting, and camera movement.

Follow an astronaut in a white suit as she steps into a glass observation bay and sees a ring-shaped space station unfolding beyond the planet's blue horizon. Start on her visor reflection, track past her shoulder, and reveal the station's lights and slow rotation. Keep the suit, visor, bay structure, planet, and light direction consistent; use original designs with no copied franchise characters, unreadable text, gore, subtitles, or watermark.

Inputs and workflow
Photo to use
Choose a model that accepts multiple references, upload them in order within its documented limit, and include one foreground scale object plus one environment reference.
What to change
For a single-image model, combine the astronaut and station cues into one reference or keep the strongest image; larger reveals can use separate clips and external editing.
Watch for
Unseen world details and exact scale can be invented; inspect silhouettes and object continuity.
Use this prompt

Generate a clip

Sports action loop

Keep run-up, action, contact, and landing physically readable.

Show one athlete completing a readable action loop: preparation, approach, central move, contact or takeoff, and controlled landing or finish. Preserve body proportions, equipment, ground contact, direction, and safe surroundings. Use camera movement that follows the action without hiding its key link. No impossible hovering, teleportation, duplicate limbs, unsafe graphic injury, subtitles, or watermark.

Inputs and workflow
Photo to use
Use a reference that shows the athlete's equipment, body line, ground, and direction of travel.
What to change
A full routine with several tricks needs separate clips or a storyboard and external assembly.
Watch for
Fast limbs, equipment, contact, and landing physics can drift; inspect every link of the loop.
Use this prompt

Two real runs from our existing showcase

Published AwakePix runs with the original input, prompt and recorded settings. Expand each example for its original prompt; check the generator for current pricing.

Product motion from one reference image: original reference image
Input photo
Generated result

MiniMax H3 Max ·

Product motion from one reference image

5s · 768p · 9:16 · 40 credits

A product sequence with its original reference image and prompt. Adapt the composition to your own product.

Read the exact original prompt

Begin from the uploaded frame, keeping the woman, her floral dress, headscarf, the yellow sunglasses, the coral phone, the yellow handbag and the sunny street background consistent. A man in a dark jacket walks in from the right edge and collides with her shoulder. On impact her sunglasses fly off her face, the phone leaves her hand and the handbag is knocked out of her grip; all three products launch outward through the air in different directions, tumbling in slow motion, staying sharp and undistorted. The camera pushes in at a moderate amplitude and a slow steady pace. Street ambience with a sudden thud on impact.

View model details
A cat by the window: original reference image
Input photo
Generated result

Gemini Omni 1.1 Flash ·

A cat by the window

5s · 720p · auto aspect · with audio · 55 credits

An earlier single-image generation with a fixed camera and small animal movements. Changing the photo or model can change the result.

Read the exact original prompt

Begin from the uploaded photo and keep the ginger cat, the windowsill, the basil pot and the morning light unchanged. The cat slowly turns its head toward the window, its ears twitch and its tail curls gently; the basil leaves stir in a light breeze. The camera stays still. Soft kitchen ambience with distant birdsong.

View model details
Open the video generatorSeedance features on AwakePix

Some templates adapted from LearnPrompt / awesome-seedance (CC BY 4.0)