Skeleton rig · part slicing · mesh posing · WindSeeker
1. Load Image
📷 Which picture gives the best result?
Whole body, facing the camera, arms held away from the body (T or A pose), hands and feet in the picture.
What is not in the picture cannot appear later. No hands in the picture means no hands when the performer raises a hand.
Not a full-body picture? Set Figure (under Save Rig) to Waist up or Head & shoulders.
Waist up: keep the whole arms with the hands inside the frame, or have the arms down along the body. An arm that is cut off by the edge moves like a stump — better erase the half arm (Paint → Remove) or choose Head & shoulders.
Head & shoulders: only the head and the face move.
Plain background in a color the character does not have. A large, clear face (eyes and mouth easy to see).
Photos of people: face the camera with a neutral face and a closed mouth, even light, no hat, glasses or hair over the face; fitted, plain clothes. Set Picture (under Save Rig) to Photo of a person before you split. The head, neck, chest, shoulder and face points are then placed for you (on this computer, nothing is uploaded) — check them and drag any point that is off. Photos of real people are portrait-only for now: use a head-and-shoulders picture — the head turns, the face moves with its own lips, eyelids and eyebrows, the body leans; arms and legs stay still.
Drawings and dolls with clear limbs move more cleanly than photos of people in wide clothes.
Making a portrait with an AI image generator? Copy the prompt and change the parts in [brackets]:
Realistic studio close-up portrait of [a young woman], head and shoulders, facing the camera straight on, head upright and not tilted, the whole head with all of the hair and both shoulders inside the frame with margin above the head. Neutral face, mouth closed, eyes open looking at the camera, [hair tied back] so ears and neck are clear, no glasses, no hat. Sharp, large, clearly visible eyes and mouth. Plain seamless light grey background, soft even frontal lighting.
Background tolerance12
A solid-color background is removed automatically on load. Raise the tolerance if a halo remains; lower it if dark hair or clothing gets eaten. If part of the background stays (brush strokes, a shadow, a second color), press the button above and click what is left.
2. Mode
Skeleton
Paint
Face
Pose
Drag the skeleton points onto the character.
3. Skeleton Points
Save Rig and Split into Parts also remember your work in this browser: open the same image again and the skeleton, mask and face points come back by themselves. Nothing is uploaded.
Finds the face and shoulders for you. You can still drag every point afterwards.
Next: drop the downloaded files into Cutout Animator to animate with keyframes, or into Live Puppet to perform with your camera.
Joint overlap / cap radiusauto
🎯 Per-part reach
How far each part reaches from its bones. Higher = the part claims a wider area (e.g. the head covering the hair).
View
Welcome!
Load a PNG from the left panel. Skeleton points are placed automatically; then you correct them.
Tips
Mode 1: Edit Skeleton
• Drag the points to the right places
• "Auto Fit" gives you starting positions
• When done, switch to "Pose" mode to animate
🖌️ Manual Mask Touch-up
Correct the automatic split. Press Preview first (left panel), then pick a part here and fix it with the brush, or click an area with Smart select.
Tool
Color similarity24
Click an area on the character: everything connected to it with a similar color gets the action below in one go.
• Hair lying on a shoulder: choose Head, click the hair.
• Clothing showing through the hair: choose Torso, click the cloth.
• A leftover patch of background: choose Remove, click it.
Too much or too little? Move the slider — the last click updates.
Action
Paint: force pixels into the selected part. Revert: return pixels to the automatic split. Remove: make pixels fully transparent (they disappear from the parts and from Pose mode).
Brush Size30px
Paint into which part?
🙂 Face Points
Optional — lets Live Puppet blink, look around and move the mouth. Mark both eyes and the mouth; nothing is cut here, the points are saved with the parts.
Blue oval — the eye opening. Drag its center to move it, the side and bottom squares to size it. Fit it to the visible eye, just inside the outline. Yellow circle — the iris (the colored part). Drag its center, or its square to size it. Red oval — the mouth. Put its cross on the middle of the mouth and size the oval to the widest, most open mouth you want (a big laugh). Bigger oval = bigger talking mouth.
No face points yet.
Selected Bone
— Select a bone —
↻:0°
Click a bone to select it — then use the slider or drag the bone directly.
⊞:24×24
Mesh resolution — higher = smoother deformation, slower. The mesh is re-baked when it changes.