Workflow: from a drawing (photo, PDF, DXF, DWG) to a 3D model

Which way to use

• I have a DXF or DWG: convert it directly (way A).
• I have a photo or PDF with dimensions: trace and draw over it (way B) or ask the AI (way C).
• I only have an idea or a sketch: AI (way C).

A. DXF/DWG → 3D

1. File → Open… the DXF/DWG. It opens as a drawing; the geometry shows in Model.
2. At the top choose Convert to 3D (2D → 3D) or Open as sketch in a part.
3. Select the outline layers. Choose the method: Recognize JIS three views (aligned views: depth comes from the views), Extrude closed profile (one outline + depth in mm) or Structural members (section) (each line becomes a library profile).
4. Check unit, thickness/depth and click Create solid. The original DXF is not changed; Ctrl+Z undoes.

B. Image reference

1. New part → Sketch → Image reference → JPG/PNG/PDF.
2. Mark two points with a known distance, type the mm, Apply and trace.
3. Draw and dimension over the image; then extrude.

C. Generate with AI / Bob

1. Advanced modeling → Generate with AI (or ask Bob in Support). Describe what you want and attach up to 4 images (PNG/JPEG/WebP).
2. Choose the engine. In + Options: Vectorize to 2D, 2D → 3D directly, Initial parametric design or Background reference.
3. Generate. The stages show (reading the drawing, dimensions, building the sketch). Cancel keeps text and images.
4. Check the preview, sizes and warnings. Nothing changes in the document until Apply draft (one undoable operation). The DXF button saves the draft profiles.

Tips

• The AI does not invent what the drawing does not say: missing sizes become questions or editable assumptions. Review before manufacturing.
• Drawing views do not show hidden cavities: check the generated solid.

Common mistakes

• Depth highlighted in red: the value is missing. The suggestion is only a reference; type the real size.
• Image → DXF with imperfect reading: review the dimensions before continuing.