Manual pose control
Drag the body handles to adjust the head, torso and limbs. Connected body parts follow the movement as you shape the pose.
Pose first. Create next.
AI pose generator
Arrange people, animals and props in a 3D scene. Apply pose presets, adjust joints and frame the camera, then use your scene as a reference for AI image generation.
Inside Pose StudioScene-to-image walkthrough
From pose to image
How to turn a 3D pose into an AI image.
A pose reference gives the image model visual direction for the body and framing. Your prompt supplies the character, setting, and style.
Pose and frame
Add your characters and props, then apply a pose preset. Refine the body controls and adjust the camera until the scene reads clearly.
3D pose editorCapture the reference
Select Pose to Image to capture the visible scene without the editing handles. The capture is attached as a reference in the image workspace.
Scene → reference imageDescribe and generate
Add your subject, clothing, environment, and lighting to the prompt. Check the displayed credit cost, sign in, and generate an image. Review the result and adjust as needed.
Reference + prompt → image
Original examples
Pose references and AI image results.
Compare a posed scene with its generated image. Explore how camera angle, body position and framing shape the result.
Crossed arms
Compare the front and low camera views
Kneeling
Compare the oblique and high camera views
Jogging
Compare the arm swing and leg placement
First attempts, without retouching. Pose, anatomy, and framing can differ from the reference. See more examples with people, animals and props →
Direct the scene
Pose presets, manual control and camera framing.
Work out the silhouette, limb placement, and camera angle before asking the image model to interpret your idea.
Explore the controlsFind your framing
Drag empty space to orbit the scene and scroll to zoom. The captured camera view becomes part of the reference you send to image generation.
Capture a clean reference
Send the visible characters, props and camera view to the image workspace. Editing handles stay out of the capture, leaving a clear reference for your prompt.
Start, adjust, undo
Try crossed arms, kneeling, or jogging presets in the editor. Refine them with the handles, use undo and redo, or reset the selected mannequin.
A useful starting point
Guide the pose. Review the result.
A captured mannequin scene is a visual reference. The image model interprets it alongside your prompt, so a generated image may differ in anatomy, pose, camera angle, or placement.
Start with a readable silhouette and clear limb placement. Check hands, overlapping limbs, and body proportions in every result. Exact pose matching and identity preservation are not guaranteed.
AI Pose Generator FAQ
Before your first pose.
What does this AI pose generator do?
Seedance combines a manual 3D pose editor with reference-guided image generation. You pose the mannequin and choose the camera view, then attach a capture in the image workspace and describe the image you want. The image model interprets that reference when generating.
Can I use the pose editor without generating an image?
Yes. You can open the editor, adjust the mannequin and camera, and prepare a reference without submitting a generation request. Image generation is a separate action that requires sign-in and sufficient credits.
Can I pose multiple characters at once?
Yes. Add people and animals to the same scene, select each character to adjust its pose, and arrange props around them. Capture the composition as a shared reference. Browse the editable pose scenes for a starting point.
Will the image match the pose exactly?
Exact matching is not guaranteed. The reference guides body pose and camera framing, but the model can reinterpret anatomy, overlaps, and placement. The mannequin also does not establish a person's identity. Refine the reference and prompt if the output misses your intention.
Does the editor export OpenPose data?
No. This workflow captures a rendered image of the 3D scene. It does not provide an OpenPose skeleton, keypoint file, or ControlNet pose map export.
Can I turn my generated image into a video?
Yes. Generate an image from your pose scene, then use the image result's video action to carry it into the image-to-video workspace as a reference. Describe the action and camera movement you want before generating the video.