Conjure - Blender Add-on

A Blender add-on that uses AI to generate 3D objects from text prompts!

Workflow

1. Gemini 3: Refines your simple text prompt into a detailed description. Uses Gemini 3 Pro.
2. Nano-Banana Pro: Generates a high-quality reference image from the refined prompt along with 3 more reference images from different angles.
3. Meshy.ai: Converts the reference image into a 3D model (.glb) and imports it into Blender.

Installation

1. Open Blender.
2. Go to Edit > Preferences > Add-ons.
3. Click Install... and select the conjure_addon folder (or zip it first).
   • Note: If installing as a folder doesn't work, zip the conjure_addon directory into conjure_addon.zip and install that.
4. Enable the add-on "Conjure 3D AI".
5. Enter your API keys for Gemini and Meshy.ai in the add-on settings.
6. Click Install Dependencies to install the required Python dependencies.

Blender Add-On Store

Watch this space! Once the add-on is approved, we will update this space with instructions to download the extension from the Blender Extensions Store directly.

Dependencies

This add-on requires the requests and google-genai Python libraries. Install them into your Blender Python environment by going to Edit > Preferences > Add-ons > Conjure > Install Dependencies.

In the future, we will move to bundle these dependencies in the extension itself to remove this manual step. More details here.

Usage

1. Open the 3D Viewport.
2. Press N to open the sidebar.
3. Click on the Conjure tab.
4. Type your prompt (e.g., "A magical potion bottle").
5. Click Generate 3D Object.
6. Wait for the process to complete (check the Status tab for more).

Paid Components Disclaimer

1. The Gemini API uses gemini-3-pro-preview, which is a paid API, to refine the prompt.
2. The Gemini 3 Pro Image preview (Nano Banana Pro) model is then used to generate multi-view images of the object. This generates 4 images.
3. The Meshy.ai API is then used to generate a 3D model from the images. This is a paid API as well.

Configuration

• Gemini API: Uses Google Generative AI.
• Nano-Banana: Currently configured to use a placeholder generator for demonstration if a private endpoint is not set. You can edit utils.py to point to your specific Nano-Banana endpoint.
• Meshy: Requires a valid Meshy.ai API key.

Demo Video

Click here to watch the demo video. Note that this video is sped up 2.5x for the sake of brevity.

Development

This project uses pre-commit with ruff for linting and formatting.

uv add --dev pre-commit ruff  # Install dev dependencies
uv run pre-commit install     # Set up git hooks
uv run pre-commit run --all-files  # Run manually on all files

Contributing

Please feel free to refer to the Feature List to contribute to this project. We welcome all kinds of contributions, such as improving the docs or adding more models and coverage.

License

Apache License 2.0