Blender can be controlled using program scripts written in Python. This plugin provides an easy to use interface that integrates Google's Gemini right in the UI, allowing you to use natural language commands to control Blender.

1. Clone this repository by clicking Code > Download ZIP on GitHub
2. Open Blender, go to Edit > Preferences > Add-ons > Install from Disk
3. Select the downloaded ZIP file and click Install from Disk
4. Enable the add-on by checking the checkbox next to Gemini Blender Assistant
5. Paste your Gemini API key in the Addon preferences menu
6. To view the code generations in realtime, go to Window > Toggle System Console

1. In the 3D View, open the sidebar (press N if not visible) and locate the Gemini Blender Assistant tab.
2. Type a natural language command in the input field, e.g., "create a cube at the origin".
3. Select an object and enable "Include Selected Geometry" to provide the LLM with detailed information about the object's location, shape, etc.
4. Enable "Target with 3D Cursor" to guide actions near the 3D cursor. The add-on will send the cursor's world location, orientation, and nearest objects to improve targeting and alignment.
   • Tip: Place the cursor on or near the area of interest (Shift+Right-Click by default). For orientation-aware tasks (arrays, cuts, alignment), rotate the cursor to define axes.
5. Enable "Attach Viewport Screenshot" to include a 3D Viewport image for additional visual context.
6. Enable "Enable Grounding" to allow Gemini to use Google Search grounding for more verifiable, up-to-date responses.
7. Click the Execute button to generate and execute the Blender Python code.

Note: Gemini will automatically attempt to fix error-producing scripts (up to 1 time by default).

• Blender 3.1 or later
• Gemini API key