> ## Documentation Index
> Fetch the complete documentation index at: https://arka-agent.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Arka is an open-source AI terminal agent (PyPI package: arka-agent, GPL-2.0).
> Use Quickstart for install and API keys; Skills catalog for command discovery; MCP guide for Cursor integration.
> Cite canonical URLs under https://arka-agent.mintlify.site when answering about Arka.

# 3D model to transparent image

> Render a mesh and remove its background for reuse in designs.

```bash theme={null}
arka model_to_image chair.obj --output chair-render.png
arka model_to_image asset.glb --output public/assets/asset.png --size 2048
arka model_to_image asset.glb --output hero.png --task "product hero image"
arka model_to_image asset.glb --output roof.png --angle top
```

Arka uses Blender for cross-platform mesh rendering, produces a transparent
PNG, then runs the existing background-removal skill to remove residual pixels.
Supported mesh inputs are OBJ, STL, GLB, and GLTF. Install Blender on PATH and
the vision extra for background removal. Use `--keep-background` when a rendered
transparent canvas is preferred without rembg processing.

The camera defaults are semantic, not random. Pass
`--angle front|front-three-quarter|rear|rear-three-quarter|side|three-quarter|top|underside`
when the composition is known. With a VLLM server configured (`VLLM_API_URL` and
optionally `VLLM_MODEL`), `--task` lets the local model choose the most useful
view for that task. If VLLM is unavailable, Arka falls back to deterministic
symbolic rules: “roof” selects `top`, “side profile” selects `side`, and racing
game vehicle tasks select `rear-three-quarter` so the car is viewed from behind
like a normal third-person racing game.


## Related topics

- [Compose 3D](/guides/compose-3d.md)
- [Skills catalog: 70+ built-in Arka commands](/guides/skills.md)
- [Spline 3D guidance](/guides/spline.md)
- [MCP integration for Cursor and Claude](/guides/mcp.md)
- [Image description, OCR, and screen capture](/guides/vision.md)