Adjust / save

01 / Load a photo

＋ Choose an image

You can also drag and drop. Transparent PNGs and photos with a plain background work best.

Your original photo stays private
The original photo is displayed and processed only in your device’s browser. It is not sent to or stored on the site’s server or shared library, and is not visible to the operator or other users. Only the exported line drawing is shared.

Processing resolution (applies to the next image)

Standard uses about 44% of the previous maximum pixels. For fine legs or branches, choose High detail before importing. PC AI uses 512px input. Mobile portrait AI uses a longest edge up to 512px, while other subjects use 320px.

Processing mode

Select a photo, open 3 Lines and tap Generate line drawing for automatic AI cutout and linework. Processing is free and stays on your device. The first download is about 15–18MB. People uses dedicated portrait AI; other subject types use general AI. These smaller models differ from PC AI, so fine details and complex backgrounds may need corrections. If processing fails, switch to Colour / brush.

02 / Recognize the subject

Recognition methodAI execution

Choose the recognition target

Mobile AI cuts out the photo without prompts. For multiple subjects, draw a box around the whole subject before generating. Points select separate objects; use Keep / Remove brushes for fine corrections. PC AI supports box and point prompts.

Recognize the whole photo

Use Keep to paint missing areas, even before recognition.

Extract the subject → correct its mask → generate lines.

Lightweight subject selection downloads about 14MB on first use. Standard whole-image recognition downloads about 192MB on CPU or 100MB on GPU. Photos are processed on your device. Correct fine parts with subject points or brushes. Additional taps on the same photo and box reuse image embeddings.

Mask threshold128

Mask brush radius (px)

Paint Keep over missing legs and Remove between them. Regenerate after changing the mask.

Import / export mask

Import a mask image

White is subject; black is background. Use the same aspect ratio as the original image.

03 / Line settings

Lines to keep after recognition

Choosing Furniture selects part boundaries. Candidate parts are segmented to create boundaries; Internal line amount is not used in this mode. For an existing line drawing, choose Classic recognition.

Subject type

Line preset

Adjust regions of a person

Bottom of head (%)Top of shoes (%)

Regions are divided by full-body height. This is not semantic body-part recognition.

1. Create outline → 2. Separate parts by photo colors → 3. Add boundaries only between parts. You can add boundaries after outline generation while preserving the outline.

Refine the lines

Pick two points to repair · regeneration resets repairs

Finish lines before export

Smoothness45 Preserve major corners

Resamples points, reduces jitter and fits cubic Bézier curves. Segments with excessive deviation retain their original lines. Zoom in to check the result. SVG saves Bézier curves; DXF uses fine line segments. Reapplying uses the original, unfinished lines.

Trim unwanted lines

Deletion method

To delete a section, click its start and end on the same line.

Selection removes whole selected lines; the eraser removes only the painted section. Section deletion removes the shorter arc of a closed outline. Press Esc to finish. Regeneration or boundary updates reset deletions. Export bounds and scale are preserved.

Add gray fill inside the outline

Outline fillFill transparency50%

0% is opaque; 100% is transparent. Updates the Line drawing and Overlay views immediately. Fills closed outlines while preserving holes. Open outlines are not filled.

SVG preserves transparency. DWG and DXF use gray hatches to approximate the appearance on white paper.

04 / Export for drawings

Actual height (mm)Scale 1 :

Export mode

SVG is an editable vector format. Open it in Illustrator, then use Save As to save an AI file. SVG respects export mode and scale.

Workspace

100%Wheel: zoom / Space + drag or middle button: pan

Choose an image to get started.