Portrait push-in
A still portrait gains a slow cinematic push-in and a hint of breeze — no description prompt needed.
Turn any image into AI video
Image to video, the simple way: drop a still, add one motion prompt, get a cinematic 6–15 second clip.
How do I turn an image into a video with AI?
Drag a still image onto the Canvas — this is the image-to-video step — write one motion prompt describing what moves, and Seedance 2.0 animates it into a 6, 10, or 15 second clip. "Slow camera push-in", "leaves drifting in wind" — skip describing the image; Seedance reads it. Each aspect ratio renders natively, not a center-crop.
Building a full music video instead? See the AI music video guide
Image to video in 3 steps
The standard 3-step Seedance 2.0 image-to-video flow. First/last frame control is optional.
Step 1Drop photo
Drag any JPG/PNG onto the Canvas surface. Seedance i2v handles portraits, products, and scenes; any size, upscaled internally.
Step 2Write motion prompt
Describe what moves: "slow camera push-in", "leaves drifting in gentle wind". Skip the photo description — Seedance reads the image.
Step 3Set duration, aspect
Duration: 6s, 10s, or 15s. Check 9:16 / 1:1 / 16:9 boxes — each aspect renders natively, not a center-crop.
From still photo to clip
One image plus a motion prompt — Seedance 2.0 i2v handles portraits, products, and scenes.
Product showcase
A flat product shot turns into a rotating CGI reveal with lighting that holds throughout.
Ambient motion
Keep the subject still and let the world move around them — crowds blur, the frame stays composed.
Lifelike detail
Small natural motion — a glance, a shift in posture — animated from a single everyday photo.
What image-to-video actually does
Image-to-video starts from a single still and treats it as the first frame, then generates the seconds that follow. That first-frame control is the whole reason the output looks like your photo and not a vaguely similar scene: the subject, composition, and lighting you fed in are locked as the opening frame, and the model only invents what comes after. You are not describing the image — Seedance 2.0 reads it directly — you are describing the change.
So the prompt is a motion brief, not a scene description. The clip gains two kinds of movement the still never had: camera motion and subject motion. Camera motion is how the frame itself moves — "slow push-in", "left dolly", "tilt-down" — and subject motion is what moves inside it — "leaves drifting in wind", "a slow glance", "hair shifting in a breeze". Separating the two in your head is the fastest way to get the result you pictured.
In Renoise this runs on Seedance 2.0 i2v. Keep the prompt to verb-noun motion pairs and let the model handle physics and continuity across the 6–15 second clip. When you need the ending pinned rather than chosen — a product reveal, a lighting shift — set a last-frame anchor and Seedance interpolates between your start and end states instead of improvising the finish.
Renoise capabilities used
Three pieces matter most for photo-to-video work.
Seedance i2v
Hand it one photo plus a motion prompt; get a coherent 6–15 second clip.
Frame control
Pin start and end frames; Seedance interpolates for transitions, lighting shifts, and reveals.
Multi-aspect render
Generate 9:16, 1:1, and 16:9 from one job, each rendered natively.
Choose your plan
One plan unlocks Seedance 2.0 and every other model.
Ends in
Current Plan
StarterFor first-time AI content creators
$20/mo
Upgrade Plan1,200 ©/mo
≈ 400 GPT Image 2 Generations≈ 60 Seedance 2.0 videos
$1 = 60©
Generation Discount
Seedance 2.0$0.083/s
Kling 3.0$0.267/s
Nano Banana 2$0.133/img
All other models
✓
GPT Image 250% OFF
✓
Watermark-free exports
✓
Image Models
✓
Video Models
Current Plan
StandardFor creators shipping content every week.
$60/mo
Upgrade Plan3,600 ©/mo
≈ 1,200 GPT Image 2 Generations≈ 211 Seedance 2.0 videos
$1 = 60©
15% Generation Discount
Seedance 2.0$0.071/s
Kling 3.0$0.227/s
Nano Banana 2$0.113/img
All other models
✓
Seedance 2.0 Series15% OFF
✓
GPT Image 250% OFF
✓
Watermark-free exports
✓
Latest Image Models
GPT Image 2
Seedream 5.0 Lite
Midjourney V8.1
Nano Banana Pro
Grok Imagine Image Quality
✓
Latest Video Models
Seedance 2.0
Gemini Omni Flash
Kling 3.0 Omni
Grok Imagine Video 1.5
HappyHorse 1.0
Best ValueCurrent Plan
AdvancedFor studios and pros producing at commercial scale.
$200/mo
Upgrade Plan14,000 ©/mo2,000© BONUS
≈ 4,666 GPT Image 2 Generations≈ 1,000 Seedance 2.0 videos
$1 = 70©17% MORE
30% Generation Discount
Seedance 2.0$0.050/s
Kling 3.0$0.160/s
Nano Banana 2$0.080/img
All other models
✓
Seedance 2.0 Series30% OFF
✓
GPT Image 250% OFF
✓
Watermark-free exports
✓
Latest Image Models
GPT Image 2
Seedream 5.0 Lite
Midjourney V8.1
Nano Banana Pro
Grok Imagine Image Quality
✓
Latest Video Models
Seedance 2.0
Gemini Omni Flash
Kling 3.0 Omni
Grok Imagine Video 1.5
HappyHorse 1.0
Animate a photo in minutes
Drop an image, add a motion prompt, get watermark-free clips.
Frequently asked questions
1.How long can the output clip be?
Seedance 2.0 i2v supports 6, 10, or 15 second outputs from a single image. For longer videos, generate multiple clips with shared first/last frames and stitch them on the Canvas Timeline.
2.Does the photo need to be high resolution?
No. iPhone quality (12MP+) is fine; older smartphones (5MP+) also work. Renoise upscales internally before passing the image to Seedance. The bottleneck is subject clarity, not pixel count.
3.Can I control the camera motion?
Yes — through the motion prompt. Use verb-noun pairs: "slow push-in", "left dolly", "tilt-down", "static frame with subject motion only". Seedance interprets these directly, so you don't need camera-language jargon.
4.Last-frame anchor versus motion prompt?
Motion prompts let Seedance choose the ending; last-frame anchors lock the exact end state. Use last-frame for controlled transitions like product opens or lighting reveals, and prompt-only for natural camera and ambient motion.
5.Will faces in the photo stay consistent across the clip?
Yes within a single clip. Seedance 2.0 preserves facial features across the 6–15 second output. For consistency across multiple clips (multi-shot videos), @-reference the same image in each clip prompt. If that image contains a real human face, clear it through FacePass first — video models block real-face reference images, and FacePass is the compliant path once you hold the rights to that likeness.
6.Can I animate a product photo for ecommerce?
Yes — and it's one of the most common uses. See the AI product video guide for a full ecommerce workflow with batch generation across SKUs.
7.What is image to video AI and how does it work?
Image to video AI takes a single still image as the opening frame, then generates the seconds that follow based on a motion prompt you write. Renoise runs this on Seedance 2.0 (ByteDance), which reads the image directly — you only describe the motion — and outputs a 6–15 second clip at your chosen aspect ratio.
By Max, RenoiseLast reviewed Models verified: Seedance 2.0