Seedance 2.0 Prompt Templates for Real Person Video

Apr 5, 2026

Need better prompts for Seedance 2.0 real person image-to-video? The fastest way to improve results is not writing longer prompts. It is using shorter prompts with cleaner motion logic.

This guide gives you practical prompt templates you can paste, adapt, and test. If you need the bigger strategy first, read these two related guides:

What makes a good Seedance 2.0 real-person prompt?

A strong prompt usually does four jobs:

  1. identifies the reference image role
  2. defines one main movement
  3. defines one camera behavior
  4. sets the realism boundary

Prompt structure that works

Use @image1 as the identity anchor. [subject movement]. [camera movement]. [lighting / mood]. Keep facial proportions stable and avoid exaggerated motion.

Quick rules before the templates

  • use one primary motion per prompt
  • use one camera instruction per prompt
  • keep identity language short and practical
  • do not over-describe beauty details
  • if the face drifts, reduce action before you add adjectives

Prompt templates

1. Talking head prompt

Use @image1 as the identity anchor. The subject speaks calmly to camera with subtle lip movement, natural blinking, and a slight head tilt. Soft studio lighting, gentle push-in camera, realistic skin texture, stable facial proportions, no sudden zoom.

Best for:

  • explainer clips
  • AI spokesperson videos
  • creator intros

2. UGC ad selfie prompt

Use @image1 as the portrait reference. Create a vertical 9:16 handheld-style selfie video with natural eye contact, a light smile, subtle shoulder movement, and realistic daylight. Keep the face stable, casual, and believable, with no extreme expression changes.

Best for:

  • short ads
  • creator promos
  • app and SaaS testimonials

3. Beauty portrait prompt

Use @image1 as the identity anchor. The subject slowly turns from a three-quarter angle toward the camera, blinks naturally, and gives a soft confident smile. Editorial beauty lighting, clean background, smooth cinematic camera drift, consistent face, hair, and makeup.

Best for:

  • beauty campaigns
  • skincare creatives
  • premium editorial portrait clips

4. Fashion model prompt

Use @image1 as the first frame and identity anchor. The subject takes one slow step forward, turns slightly at the waist, and lets the fabric move gently. Premium fashion lighting, elegant camera arc, consistent face and outfit details, no fast movement.

Best for:

  • apparel launches
  • fashion lookbooks
  • creator brand shoots

5. Fitness coach prompt

Use @image1 for identity and @image2 for upper-body framing. The subject demonstrates one simple exercise motion with clear posture and controlled pacing. Bright indoor lighting, minimal camera shake, stable face and hands, no aggressive body movement.

Best for:

  • trainer content
  • tutorial shorts
  • wellness and rehab clips

6. Travel creator prompt

Use @image1 as the identity anchor. The subject looks toward the scenery, then glances back at camera with a relaxed smile. Gentle breeze in the hair, soft golden-hour light, slow cinematic pan, realistic skin and stable facial structure.

Best for:

  • travel reels
  • lifestyle promos
  • destination campaigns

7. Product presenter prompt

Use @image1 as the identity anchor and @image2 as the product reference. The subject lifts the product slightly toward camera, smiles naturally, and holds a stable presentation pose. Clean commercial lighting, subtle push-in camera, consistent face, hands, and product placement.

Best for:

  • ecommerce demo videos
  • cosmetics launches
  • affiliate clips

8. First-frame to last-frame prompt

Use @image1 as the first frame and @image2 as the last frame. Generate a smooth transition where the subject slowly turns, shifts posture naturally, and ends in the pose shown in @image2. Maintain the same identity, hairstyle, outfit, and realistic facial proportions throughout.

Best for:

  • before/after scenes
  • short cinematic transformations
  • continuity-driven portrait clips

9. Reference video + portrait prompt

Use @image1 as the identity anchor and @video1 as the motion reference. Replicate the pacing and camera language of @video1 while keeping the face, hairstyle, and outfit from @image1 stable. Realistic human motion, no over-exaggerated facial changes.

Best for:

  • dance-lite motion
  • cinematic pacing transfer
  • creator-style replication

10. Safe fallback prompt when results keep drifting

Use @image1 as the identity anchor. The subject remains mostly still with natural blinking, a slight head movement, and a calm confident expression. Soft neutral lighting, fixed camera with a subtle push-in, stable facial proportions, no dramatic motion.

This is the prompt to use when everything else is breaking.

Negative phrasing that often helps

If your platform accepts natural negative phrasing, add one short line like this at the end:

Avoid exaggerated expressions, sudden zoom, face distortion, and unstable hands.

Do not turn the prompt into a giant negative list. One short correction is usually enough.

Prompt upgrades by scenario

ScenarioAdd thisAvoid this
talking headnatural blinking, subtle lip movementdramatic head turns
beautyeditorial lighting, smooth camera driftbig emotion swings
UGChandheld-style, natural daylighthyper cinematic camera moves
fashiongentle fabric motion, one body turnwalking + spinning + zooming together
trainercontrolled posture, stable handsfast repetitive action

How to iterate prompts correctly

Most users make two mistakes:

  1. they rewrite the whole prompt after every bad result
  2. they change three variables at once

Use this better workflow instead.

Pass 1: test identity

Keep motion tiny.

Pass 2: test camera

Only change the camera behavior.

Pass 3: test action

Only increase subject movement if identity still holds.

That method tells you what actually broke the output.

Prompt patterns to avoid

Overwritten beauty prompt

A gorgeous super realistic perfect beautiful woman with flawless face, perfect lips, ultra smooth skin, model face, highly detailed eyes, hyper realistic beauty portrait.

Why it fails:

  • too generic
  • too aesthetic-heavy
  • weak motion logic
  • encourages average-face optimization

Chaos prompt

She runs toward camera, flips her hair, laughs, points to the viewer, fast zoom in, spinning side camera, dramatic lighting change, emotional performance.

Why it fails:

  • too many actions
  • too many camera instructions
  • identity cannot stay anchored

Best final advice

For Seedance 2.0 prompt templates, the winning principle is simple:

reference images define the person, prompts define the movement.

If a result looks wrong, do not immediately add more adjectives. First reduce motion, simplify the camera, and make the prompt easier for the reference to survive.

Seedance 2 Video Team

Seedance 2 Video Team

Seedance 2.0 Prompt Templates for Real Person Video | Seedance 2 Blog | AI Video Tutorials, Prompts & Updates