P-Video-Animate

P-Video-Animate animates a single image using the motion, timing, and camera movement from a given video.

Given a reference still and driver video, the model animates the image with the clip’s motion—output keeps the atmosphere of the still (look, lighting, wardrobe) while preserving the driver’s acting, timing, camera movement, and scene structure.

It is optimized for:

• Top visual quality

• Most efficient inference

  • Speed: 5.24 s generation time per 1 s of output video

  • Price: $0.03 and $0.06 per 1 s of 720p and 1080p video

Not sure how this differs from P-Video-Replace, P-Video, or P-Video-Avatar? See P-Video-Animate vs. P-Video-Replace and Choosing the right video model below.

Note

When using P-Video-Animate, respect the copyright of videos and images you use as input and of the video you generate.

Pricing:

Resolution	Price
720p	$0.03 per second of output video
1080p	$0.06 per second of output video

Tip

Test it in the P-Video Playground.

Prompt formula

Fast pass

Driver clip carries motion—good for first renders and timing checks.

[p-image still] [video] [instruction_prompt]

one line: who, outfit, aspect, soft light, upload winning ad / avatar take, usually empty

Locked-in

Spell out identity, motion beats, and audio sync for repeatable runs.

[p-image still] [video] [instruction_prompt]

single subject, pose, wardrobe, lens + aspect spelled out, same driver clip every run, name subject + motion beats + match lip sync and audio

Fast pass — still ↔ output compare; driver-led motion with a light still prompt and empty ``instruction_prompt``.

Locked-in — still ↔ output compare; detailed still + ``instruction_prompt`` with explicit motion and lip-sync rules.

1. Still / reference image — The API uses the first entry in images. reference_image_prompt is not a request field; generate the still with P-Image (see Domain Use Cases there), then upload the file as image.

   • Examples: “Photorealistic woman, 9:16 chest-up, soft window light, single subject”, “2D fitness mascot, cel-shaded, gym background, mouth open mid-hook”

2. video — Required driver clip. Motion, acting, timing, and camera movement come from this file.

   • Examples: a winning UGC take, avatar output, or storyboard reference performance.

3. instruction_prompt — Optional text that names the subject and the motion beats to preserve from the driver. Usually empty on a fast pass; add detail when identity drifts or lip sync needs tightening.

   • Fast pass: leave empty and let the driver carry performance.

   • Locked-in: “Animate the woman in the terracotta robe using the source video: she speaks to camera, lifts a serum bottle toward the lens, opens her palm, and nods. Match lip sync and audio from the source video.”

Slot	Fast pass (enough to run)	Locked-in (stronger control)
Still prompt (P-Image → images[0])	One line: who, outfit, basic light; single subject implied.	Single subject explicit; pose, wardrobe, expression, light direction, set, lens + aspect—stable identity across reruns.
video	Any clean driver with the motion you want to reuse.	Same clip every run when comparing stills or prompt tweaks.
instruction_prompt	Empty — driver carries motion and audio.	Subject + motion beats + sync line; call out hands, props, and match lip sync and audio from the source video.

Tip

For comprehensive video prompting (motion, framing, atmosphere), see the Video Generation guide.

How does it differ from other models?

There exist multiple models aiming at animating an image with the motion of a reference video. Among those models, P-Video-Animate is the fastest and most cost-efficient model without compromising on quality.

We provide benchmark numbers below to support it. Benchmark numbers below are directional and may vary depending on resolution, clip length, settings, provider, queue time, and test date.

P-Video-Animate vs. P-Video-Replace

Both models take a source video and reference image(s), but they are built for different workflows—not interchangeable substitutes.

	P-Video-Animate	P-Video-Replace
What it does	Animates one image using motion, timing, and camera movement from a driver clip.	Replaces the character(s) in a video with the character(s) from reference stills.
What atmosphere you keep	The image’s atmosphere—look, lighting, wardrobe, and world of the still drive the output.	The video’s atmosphere—background, blocking, camera, and scene of the footage drive the output.
When to use it	You have an approved hero still and want it to perform like an existing take.	You have finished footage and want different people in the same shot.

Rule of thumb: Choose Animate when the still defines the world; choose Replace when the clip defines the world.

Choosing the right video model

Pruna ships four performance video models. They share the same prediction API, but each solves a different production problem. P-Video-Animate requires an existing source video to drive your still.

	P-Video	P-Video-Avatar	Animate (this page)	P-Video-Replace
One-line job	Generate new footage from prompts	Speak from one still (script or audio)	Retarget one still with clip motion	Swap characters in existing footage
You start with	Text prompt (+ optional image refs)	Portrait still + voice_script or audio	Source video + one still	Source video + 1–4 identity stills
You keep from the source	N/A (new scene)	Aspect ratio of the still	Motion, timing, camera movement, and optionally audio	Camera, timing, blocking, background
Typical ask	“Make a 10 s product ad in this style.”	“This spokesperson says this line in French.”	“Animate this catalog still using our winning ad take.”	“Put our creator in this UGC b-roll.”

Quick decision guide

• No source video yet → use P-Video to create the plate, or P-Video-Avatar if you only need a talking head from a still.

• Footage exists and the hero still should move like the driver → P-Video-Animate (only the first entry in images is used). See P-Video-Animate vs. P-Video-Replace.

• Footage exists and you need different people in the same shot → P-Video-Replace (Model: p-video-replace; use instruction_prompt when multiple people are on screen).

Note

P-Video-Avatar vs. animate: Avatar creates speech from a still (TTS or uploaded audio, lip-sync-focused). Animate copies motion from a driver video onto your still—it does not write a new script. Use avatar for new lines; use animate when the timing and performance of an existing take should drive the still.

Tip

Common pipelines: Generate stills with P-Image → P-Video-Avatar for new spokesperson clips → P-Video-Replace to drop talent into b-roll → P-Video-Animate to apply a hero still to motion from an avatar or ad clip.

Speed and throughput

Metric	P-Video-Animate (720p benchmark)
Generation time per 1 s of output	5.24 s
Wall-clock for 5 s output	26.2 s
Typical motion-transfer alternatives (1 s output)	~36.0–43.0 s
Typical motion-transfer alternatives (5 s output)	~180.0–215.0 s
Price per second (720p)	$0.03 (vs. ~$0.07–$0.35 for comparable tools)

Key features

P-Video-Animate fits the same Pruna API patterns as P-Video and P-Video-Avatar:

Fastest cost-efficient motion transfer

Benchmarked at 5.24 s per 1 s of output—roughly 7× faster than typical motion-transfer alternatives at a fraction of the per-second cost.

Single-image motion transfer

Upload a video and one image to retarget an approved still from a driver clip.

Long-form friendly at 720p

Supports videos up to ~2 minutes at 720p (subject to platform limits; confirm in your account).

Reliable everyday motion

Strong on normal movement and slow, controlled action—walking, talking heads, presenters, product demos.

Motion and scene structure preservation

Follows the source clip’s acting, timing, camera movement, and layout while applying the reference image’s look.

Audio-aware output

Control source audio with save_audio and ignore_audio (see Configuration).

Practical constraints

• Output length follows the source video duration (within platform max length).

• Output aspect ratio follows the source video.

• Very fast action, heavy occlusion, or extreme camera motion may reduce consistency.

• Use a clean, front-facing reference still when possible.

• Only the first image in images is used.

Examples

One tab per Hugging Face folder (ugc_ads, film_casting, gaming, meme_remixes), with side-by-side cards showing the reference image (image), driver (video), a driver ↔ output compare clip, resolution, and copy-ready prompts. The still text is a P-Image prompt (see Domain Use Cases there for tone and structure). Use it in P-Image to create the frame, then upload that file as image in the API. The label reference_image_prompt in the copy blocks is documentation shorthand—it is not a request field.

UGC adsFilm castingGamingMeme remixes

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the Filipina woman in the terracotta robe using the source video: she speaks to camera, lifts a serum bottle toward the lens, opens her palm, raises her eyebrows, and nods. Match lip sync and audio from the sour… reference_image_prompt: Vertical 9:16 three-quarter angle from slightly below, chest-up in a bathroom, Chest-up facing camera, both hands empty and relaxed at sides, mouth slightly open as if mid-sentence, shoulders squared. Photorealistic Fil…

instruction_prompt: Animate the Filipina woman in the terracotta robe using the source video: she speaks to camera, lifts a serum bottle toward the lens, opens her palm, raises her eyebrows, and nods. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 three-quarter angle from slightly below, chest-up in a bathroom, Chest-up facing camera, both hands empty and relaxed at sides, mouth slightly open as if mid-sentence, shoulders squared. Photorealistic Filipina woman mid-twenties, warm golden-brown skin, faint beauty mark above left lip, wavy black hair in tortoiseshell claw clip, terracotta linen robe with rolled cuffs, thin gold hoop earrings and delicate chain necklace, sage-green subway tile bathroom with brass faucet, fogged mirror edge, and folded white towels on a bamboo shelf, amber ring-light catchlight in eyes, both hands empty and relaxed at sides, no product visible, soft steam haze near mirror edge, mouth slightly open as if speaking to camera, one person only. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the 2D illustrated fitness woman in the coral visor and black windbreaker using the source video: she leans toward camera, points forward, opens both hands in a listen-up gesture, then shakes her head with a hal… reference_image_prompt: Vertical 9:16 three-quarter front chest-up in a gym, Three-quarter front toward camera, slight lean inward, both arms empty and relaxed at sides, mouth open mid-hook delivery. Clean 2D flat-vector animation portrait of…

instruction_prompt: Animate the 2D illustrated fitness woman in the coral visor and black windbreaker using the source video: she leans toward camera, points forward, opens both hands in a listen-up gesture, then shakes her head with a half-smirk. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 three-quarter front chest-up in a gym, Three-quarter front toward camera, slight lean inward, both arms empty and relaxed at sides, mouth open mid-hook delivery. Clean 2D flat-vector animation portrait of an original fictional Latina fitness woman early thirties, warm tan skin, dark ponytail through coral visor, open black windbreaker with thick white sleeve stripes over grey crop tank, thin gold chain necklace, cel-shaded flat colors with crisp bold outlines, modern fitness-app mascot illustration style, not photorealistic, tight chest-up crop from mid-torso up only, shoulders and head fill the frame, legs and feet not visible, no full-body shot, three-quarter front toward camera with slight lean inward, orange resistance tubes and kettlebells blurred behind on rubber mat, charcoal gym walls with faded motivational stencil text, harsh overhead fluorescents with green rim light from left, both hands empty at sides, mouth open mid-sentence, one illustrated character only. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the Black man in the grey tank top using the source video: he keeps both hands on the shaker bottle at his chest throughout, lifts the bottle toward the camera with both hands still gripping it without shaking,… reference_image_prompt: Vertical 9:16 straight-on chest-up on an outdoor running track at midday under open sky, Chest-up facing camera, exactly one white shaker bottle centered at chest with both hands gripping the same single bottle, forearm…

instruction_prompt: Animate the Black man in the grey tank top using the source video: he keeps both hands on the shaker bottle at his chest throughout, lifts the bottle toward the camera with both hands still gripping it without shaking, raises his eyebrows with a half-smirk, then nods. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 straight-on chest-up on an outdoor running track at midday under open sky, Chest-up facing camera, exactly one white shaker bottle centered at chest with both hands gripping the same single bottle, forearms bent inward, bottle vertical and still, mouth open mid-hook, shoulders relaxed. Photorealistic Black man late twenties, deep brown skin, close-cropped fade haircut, grey performance tank top with sweat sheen, silver chain necklace, one white shaker bottle with black lid only, both forearms bent gripping one bottle at center chest, not one bottle per hand, no second bottle, no bottle in left hand alone, no bottle in right hand alone, no floating hands, no arm at side, outdoor running track with red lane lines on dark rubber surface, chain-link fence and empty metal bleachers behind, bright natural sunlight and blue sky, open-air sports field environment, not inside any building, no kitchen no cabinets no countertop no oven no indoor room, straight-on chest-up, mouth open mid-sentence, one person only. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the South Asian woman in the cream cardigan using the source video: she taps her phone screen, lifts the phone toward the camera, opens her left palm, and nods with a half-smile. Match lip sync and audio from th… reference_image_prompt: Vertical 9:16 over-shoulder three-quarter chest-up in a sunlit apartment living room, Over-shoulder three-quarter toward camera, phone held at chest in right hand with screen facing lens, left hand empty at side, mouth…

instruction_prompt: Animate the South Asian woman in the cream cardigan using the source video: she taps her phone screen, lifts the phone toward the camera, opens her left palm, and nods with a half-smile. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 over-shoulder three-quarter chest-up in a sunlit apartment living room, Over-shoulder three-quarter toward camera, phone held at chest in right hand with screen facing lens, left hand empty at side, mouth open mid-demo line, shoulders relaxed. Standing upright chest-up in front of linen sofa not sitting, torso and face three-quarter toward camera, eyes toward lens, right hand holds phone vertically at chest with budget app screen facing camera, left arm hanging relaxed empty at side, no pointing, no cross-legged pose, no profile looking down at phone. Photorealistic South Asian woman late twenties, warm brown skin, black hair in low sleek ponytail, cream ribbed cardigan over white tee, thin gold huggie earrings, linen sofa with sage throw pillow and fiddle-leaf fig blurred behind, warm afternoon window light from left, phone screen shows blurred colorful budget chart UI without readable text, standing app-demo pose matching driver opening frame, mouth open mid-sentence, one person only. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the East Asian man in the navy apron using the source video: he lifts the meal box lid halfway, peels back the paper liner to reveal portions, raises his eyebrows with a grin, then shrugs. Match lip sync and aud… reference_image_prompt: Vertical 9:16 high downshot chest-up over a kitchen counter, High downshot chest-up over a kitchen counter, both palms flat on a closed meal box lid at center frame, elbows on counter edge, mouth open mid-hook, shoulder…

instruction_prompt: Animate the East Asian man in the navy apron using the source video: he lifts the meal box lid halfway, peels back the paper liner to reveal portions, raises his eyebrows with a grin, then shrugs. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 high downshot chest-up over a kitchen counter, High downshot chest-up over a kitchen counter, both palms flat on a closed meal box lid at center frame, elbows on counter edge, mouth open mid-hook, shoulders slightly hunched toward camera. Clean 2D flat-vector animation portrait of an original fictional Black woman home chef early thirties, deep brown skin, natural hair in a high puff, sunflower-yellow apron over teal long-sleeve tee, cel-shaded flat colors with crisp bold outlines, modern meal-kit app mascot illustration style, not photorealistic, tight chest-up crop from mid-torso up only, shoulders and head fill the frame, closed coral meal-kit box flat on counter with illustrated food icons and no readable brand name, box never lifted off counter, pastel kitchen counter with copper kettle and herb jar blurred behind, warm under-cabinet light, high downshot chest-up, mouth open mid-sentence, one illustrated character only. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the 3D cel-shaded wellness mascot in the sage wrap top using the source video: she speaks with lip sync in a tight face close-up, leans slightly toward camera, tilts her head with a smile, shrugs her shoulders w… reference_image_prompt: 3D cel-shaded CGI mascot portrait only, Pixar-style mature adult woman age 25, not a photograph, not photorealistic, not a child not a toddler not a baby. Vertical 9:16 macro face-only close-up on a sunrise rooftop terr…

instruction_prompt: Animate the 3D cel-shaded wellness mascot in the sage wrap top using the source video: she speaks with lip sync in a tight face close-up, leans slightly toward camera, tilts her head with a smile, shrugs her shoulders with an eyebrow pop, then nods. No hands or arms visible at any point—face and shoulders only. Never show lotion, cream, or white product on her face. Match lip sync and audio from the source video. reference_image_prompt: 3D cel-shaded CGI mascot portrait only, Pixar-style mature adult woman age 25, not a photograph, not photorealistic, not a child not a toddler not a baby. Vertical 9:16 macro face-only close-up on a sunrise rooftop terrace, Macro face-only close-up facing camera, forehead to chin fills entire frame, no shoulders visible, no neck below jawline, no hands, no arms, no torso, mouth slightly open mid-sentence, face completely clean with uniform peachy skin and no white cream or cosmetic marks. Original fictional wellness app mascot, not based on any existing franchise. Stylized CGI with soft subsurface skin, large expressive eyes, lavender hair in a loose side braid with small jade hair clip, sage-green wrap top collar barely visible at bottom edge, soft peach sunrise sky and blurred bamboo bokeh behind, premium mobile game cel-shaded render, one 3D character only, single subject, no group. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the woman in the sand hijab and camel coat using the source video: she gives a small fist pump, turns her head slightly toward camera, opens her palm outward, then nods with a smile. Match lip sync and audio fro… reference_image_prompt: Vertical 9:16 profile chest-up on a commuter train window seat, Profile chest-up facing right, wireless earbuds visible, both hands empty at sides, mouth open mid-sentence, chin slightly lifted. Photorealistic Middle Ea…

instruction_prompt: Animate the woman in the sand hijab and camel coat using the source video: she gives a small fist pump, turns her head slightly toward camera, opens her palm outward, then nods with a smile. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 profile chest-up on a commuter train window seat, Profile chest-up facing right, wireless earbuds visible, both hands empty at sides, mouth open mid-sentence, chin slightly lifted. Photorealistic Middle Eastern woman early twenties, olive skin, hijab in soft sand linen, white wireless earbuds, camel wool coat over cream turtleneck, blurred city lights through rain-streaked train window, cool blue ambient light, profile chest-up facing right, mouth open mid-sentence, one person only. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the redhead woman in the sage robe using the source video: she speaks to camera with lip sync, runs her fingers through her hair with one hand, flips her hair toward the camera, briefly frames her face with both… reference_image_prompt: Vertical 9:16 eye-level straight-on chest-up in a bathroom vanity nook, Eye-level straight-on chest-up facing camera, both arms hanging straight down at sides with hands relaxed beside hips, no hands raised, no hands to…

instruction_prompt: Animate the redhead woman in the sage robe using the source video: she speaks to camera with lip sync, runs her fingers through her hair with one hand, flips her hair toward the camera, briefly frames her face with both hands then drops them, and raises her eyebrows with a nod. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 eye-level straight-on chest-up in a bathroom vanity nook, Eye-level straight-on chest-up facing camera, both arms hanging straight down at sides with hands relaxed beside hips, no hands raised, no hands touching hair or face, mouth slightly open mid-sentence, shoulders squared. Photorealistic white woman mid-twenties, fair skin with light freckles, copper-red wavy hair past shoulders with visible volume and shine, sage-green satin robe, single thin gold chain necklace, pink vanity with round Hollywood mirror and skincare bottles blurred behind her, warm soft vanity light catchlights, anatomically correct hands with five fingers each resting beside hips below chest line, no product visible, one person only. Copy full example prompts

Reference image image

Driver video

Driver ↔ output

1080p, 9:16

Example prompts

instruction_prompt: Animate the anime-style streamer in the violet hoodie using the source video: he slaps the desk edge, points toward the monitor, leans back laughing with palms up, then gives a quick double fist pump with lip sync. Matc… reference_image_prompt: Vertical 9:16 low-angle chest-up at a neon-lit gaming desk, Low-angle chest-up facing camera, both hands resting on desk edge below frame, mouth open mid-reaction, shoulders pulled back in disbelief. Original fictional…

instruction_prompt: Animate the anime-style streamer in the violet hoodie using the source video: he slaps the desk edge, points toward the monitor, leans back laughing with palms up, then gives a quick double fist pump with lip sync. Match lip sync and audio from the source video. reference_image_prompt: Vertical 9:16 low-angle chest-up at a neon-lit gaming desk, Low-angle chest-up facing camera, both hands resting on desk edge below frame, mouth open mid-reaction, shoulders pulled back in disbelief. Original fictional anime-style streamer portrait, not based on any existing franchise. Clean cel-anime illustration of a young man with teal undercut hair, violet gaming hoodie with white drawstrings, black over-ear headset with magenta LED ring, crisp ink outlines and flat cel shading, not photorealistic, dual monitors with blurred victory overlay behind, low-angle chest-up toward camera, mouth open mid-sentence, one illustrated character only. Copy full example prompts

Integration

P-Video-Animate uses the same Pruna prediction API as P-Video. Upload video and one image, then poll or use sync headers as with other video models.

Tip

For more information on how to use the API, see the API Reference.

API Endpoint

Base URL: https://api.pruna.ai/v1/predictions

Authentication

-H 'apikey: YOUR_API_KEY'

Step 1: Upload source video and reference image

curl -X POST "https://api.pruna.ai/v1/files" \
  -H "apikey: YOUR_API_KEY" \
  -F "content=@/path/to/source.mp4"

curl -X POST "https://api.pruna.ai/v1/files" \
  -H "apikey: YOUR_API_KEY" \
  -F "content=@/path/to/reference.jpg"

Use the returned file URLs as video and the first entry in images.

Step 2: Create generation request

Animate mode (synchronous)

curl -X POST 'https://api.pruna.ai/v1/predictions' \
-H 'Content-Type: application/json' \
-H 'apikey: YOUR_API_KEY' \
-H 'Model: p-video-animate' \
-H 'Try-Sync: true' \
-d '{
  "input": {
    "video": "https://api.pruna.ai/v1/files/file-video123",
    "images": [
      "https://api.pruna.ai/v1/files/file-portrait"
    ],
    "resolution": "720p",
    "fps": 24,
    "save_audio": true,
    "seed": 42
  }
}'

Configuration

Required parameters

Parameter	Type	Description
video	file/string	Source RGB video (.mp4). Motion, timing, and camera source.
images	file[] / string[]	Reference image(s). Animate uses the first image only.

Optional parameters

Parameter	Type	Default	Description
instruction_prompt	string	""	Usually left empty for animate.
resolution	string	1080p	Target megapixel budget: 720p ≈ 1 MP, 1080p ≈ 2 MP (aspect ratio preserved).
fps	integer	24	Frames per second of the output video.
save_audio	boolean	true	Save the video with audio.
ignore_audio	boolean	false	Ignore source audio for prompt conditioning and return a silent output video.
disable_safety_checker	boolean	false	Disable safety checker for generated videos (platform UI may still enforce checks).
seed	integer	random	Random seed. Leave blank for random.
no_op	boolean	false	Health check mode — returns status without inference.

Supported option values

• resolution: 720p, 1080p.

Argument recommendations

Use these patterns for consistent quality:

• video: prefer stable exposure, minimal motion blur, and clear visibility of the motion you want to transfer.

• images: high-resolution, well-lit, front-facing still matching the intended framing; only the first image is used.

• resolution: iterate in 720p, then rerun finals in 1080p.

• fps: match source footage when possible; default 24 is fine for most web delivery.

• save_audio / ignore_audio: keep save_audio: true for dialogue-driven clips; set ignore_audio: true when you only need motion without sound.

• seed: set for reproducible A/B tests; change one variable at a time.

• disable_safety_checker: leave default unless your workflow includes explicit moderation.