Use Image to Video when the scene is already decided and the next choice is camera movement. Both clips below start from the same room image; the camera sentence is the only prompt change.
Seedance 2.0 Mini · Start Frame · requested 5 seconds, 480P, 16:9 · audio off · 48 credits per clip on October 4, 2026.
The sofa and coffee table grow larger as the camera approaches. The window, door and seating arrangement remain recognizable through the clip.
Preserve the room layout, sofa, windows, table and wall colors from the uploaded start image. No people. The camera makes one slow, smooth forward push toward the seating area. Keep the furniture still and use steady daylight. No cuts, no new furniture, no text.
The sideways move reveals more of the television and changes the overlap between the table and sofa. The room stays recognizable, with small changes to object details.
Preserve the room layout, sofa, windows, table and wall colors from the uploaded start image. No people. The camera makes a small, smooth sideways orbit to the right around the seating area, revealing gentle parallax. Keep the furniture still and use steady daylight. No cuts, no new furniture, no text.
Each downloaded result is 864 × 496 pixels and 5.04 seconds long.
Choose a forward push to emphasize the seating area, or a small orbit to reveal depth and side details. Compare the full clips: the camera instruction changes framing, and furniture details can still shift. If a detail matters, reduce the move or guide the ending with a compatible end frame.
Try these camera directionsStart with an idea, a photo, or a clip. Pick what you want to create next.
Turn a written idea into a new image, from product concepts to illustrations.
Create an image
Use your photo to guide a new style, setting, or visual edit.
Edit a photoDescribe a scene, action, and camera movement to create a short clip.
Create a video
Current workflow
Animate a portrait, product photo, or artwork with a motion prompt.
Back to editorRestyle a clip using its motion and composition as creative guidance.
Transform a videoAnimate your image, edit its setting, or prepare a more precise motion prompt.
Replace the background when you want an edited image rather than a video.
AI background changerStart with a product, portrait, or landscape prompt and adapt it to your source image.
Image to video prompt templatesCompare start-frame, start-and-end-frame, and reference-image inputs before choosing a model.
Image to video input guideThis summary covers stable workflow requirements. Available controls and credit estimates vary by selected model, so review the editor before each generation.
| Capability | What to expect | Basis |
|---|---|---|
| Image role | Start frame, start and end frames, or reference images on compatible models | Workflow definition |
| Optional end frame | Available only when the selected model exposes the control | Live editor controls |
| File checks, duration, framing, and resolution | Use the requirements and choices currently shown in the editor | Live editor validation |
Answers to common workflow choices, review steps, and capability limits.
AI Image to Video generates a new clip from a still image and a motion prompt. Your photo guides the subject and composition while the wording directs action, atmosphere, and camera behavior. Choose a compatible model and available output settings, check the credit cost, and inspect the full video. Fine details can change during animation.
Upload a clear photo as the start frame, then describe one action and camera movement. Try: Leaves sway gently as the camera slowly pushes forward; keep the house and windows unchanged. Choose the available duration, ratio, and resolution, check the estimate, and generate. Review every part of the clip for drift before downloading.
Yes. Generate one video for free without logging in or signing up. Text to Video, Image to Video, and Video to Video share this trial. Guest videos use Seedance 2.0 Mini at 480P, up to 5 seconds. Duration and resolution controls require sign-in. Failed generations do not use the free try; further generations require credits.
For an edited still, use AI Background Changer to replace the setting or Image to Image for broader visual changes. Image to Video creates a moving clip from the source image. Choose the workflow by the output you need, then review its controls and cost. Animation and still-photo editing are separate generation requests.
Compatible models expose a start-and-end-frame mode; others accept a start image only. Use related endpoint images with matching subjects, scale, lighting, and camera angle, then describe the transition. Inspect both endpoints and the frames between them. An end image guides where the clip should finish; it does not lock every frame or guarantee an exact transition.
Your uploaded image guides the subject, composition, palette, and texture of the generated video; the prompt supplies motion and atmosphere. The model creates new frames rather than leaving the photo unchanged under a motion layer. Name the features to retain and compare them throughout the output. Faces, architecture, labels, or clothing may drift.
Start with a clear source and one small action, then state speed and camera direction separately. For a portrait, try a gentle head turn with a slow camera push rather than several gestures and cuts. Inspect the motion across the whole result and revise the instruction linked to the visible issue. Another submission uses credits and may still vary.
Text, logos, labels, and product geometry can change as new frames are generated. Ask to retain those details, use a clear source, and inspect the opening, middle, and closing frames at a useful size. Treat a product clip as a concept until reviewed; exact packaging or approved typography may need separate finishing work.
Image uploaders accept JPEG, PNG, and WebP within the size and dimension requirements shown for the selected model. Choose the input mode and model before checking the limits. Use a clear frame with the subject unobstructed. A file that passes validation is compatible input; validation does not guarantee the animated result's likeness or quality.
The selected model determines the durations and resolutions shown in the Image to Video editor. Choose a supported length that fits one clear action and review the estimate after changing settings. Switching models may reset incompatible choices. The requested duration does not guarantee the intended action, framing, or camera behavior across the whole clip.
Image to Video starts with a still frame and generates motion. Video to Video starts with footage that can guide motion, timing, or composition, with optional references on compatible models. Choose by whether you have an approved image or a source clip. Both generate new frames and require review for changes to important details.
Image to Video shows a credit estimate before submission for the selected model and supported settings, including duration and resolution where available. Check it after changing model, mode, or settings. Each additional direction is a new generation. Confirmed failed jobs follow the platform's credit-return process; a result you dislike is not automatically a failed job.
Select a secure payment provider to continue.
Sign in with Google or GitHub to generate and access your saved creations.
Create a new account with Google or GitHub to claim them. Unused credits expire 30 days after signup.
By continuing, you agree to the Terms of Service and Privacy Policy.
