nativeblend 0.1.2

NativeBlend CLI - Generate 3D models in Blender using AI

NativeBlend CLI

Open-source command-line interface for generating 3D models in Blender using AI.

Features

• 🎨 Text & Image to 3D - Generate models from natural language descriptions
• ⚡ Local Blender Execution - Fast rendering on your machine, no bandwidth waste
• 🔄 Real-time Progress - Live streaming updates from the AI agent
• 🎯 Multiple Quality Modes - Express, Standard, and Pro generation modes
• 🔐 Secure - API keys stored in system keychain

Installation

pip install nativeblend

Or install from source:

git clone https://github.com/addy999/nativeblend-cli.git
cd nativeblend-cli
uv sync
alias nativeblend="uv run nativeblend --"

Then use nativeblend command in your terminal.

Quick Start

1. Initialize the CLI:

nativeblend init

2. Get your API key from nativeblend

3. Configure your API key:

nativeblend auth login
# Enter your API key when prompted

4. Generate your first model:

nativeblend generate "a low poly spaceship"

Usage

Generate a 3D Model

# Basic generation
nativeblend generate "a red cube"

# Use different quality modes
nativeblend generate "a spaceship" --mode pro

# Use a reference image
nativeblend generate "a car" --image reference.jpg

# Enable verbose output
nativeblend generate "a tree" --verbose

Modes

• express - Fast generation, single iteration (~30s)
• standard - Balanced quality with refinement (~2min)
• pro - High quality with multiple refinement passes (~5min)

Change output dir

By default renders and outputs are saved to ./outputs. Change this in your config:

nativeblend config set output.default_dir /path/to/outputs

Authentication

# Login with API key
nativeblend auth login

# Check authentication status
nativeblend auth status

# Logout
nativeblend auth logout

How It Works

1. CLI sends prompt - Your prompt is sent to NativeBlend's cloud API
2. AI generates code - NativeBlend creates Blender Python scripts
3. Inline execution - Blender tasks (rendering, exporting) run locally on your machine during the generation stream
4. Iterative refinement - NativeBlend reviews renders and improves until perfect

Your Blender installation stays local - only prompts and small preview images are sent to the cloud. No background workers needed; everything runs inline in your terminal.

Configuration

Configuration file location: ~/.config/nativeblend/config.toml

[api]
endpoint = "https://blender-ai.fly.dev"
timeout = 300

[blender]
executable = "/Applications/Blender.app/Contents/MacOS/Blender"  # macOS
# executable = "/usr/bin/blender"  # Linux
# executable = "C:\\Program Files\\Blender Foundation\\Blender\\blender.exe"  # Windows

[output]
default_dir = "./outputs"
save_renders = true

Requirements

• Python 3.12 or higher
• Blender 4.5 or higher installed locally
• NativeBlend API key

Pricing

• Free Tier: 5 generations/month (express or standard mode)
• Indie: $29/month - Unlimited generations (express + standard + pro)
• Team: $79/month - Unlimited generations + 3 seats + team collaboration features

See pricing for details.

Troubleshooting

Blender Not Found

If you get "Blender executable not found":

# Set blender path explicitly
nativeblend config set blender.executable /path/to/blender

Development

# Clone the repository
git clone https://github.com/addy999/nativeblend-cli.git
cd nativeblend-cli

# Install development dependencies
uv sync

# Format code
black src/
ruff check src/

License

MIT License - see LICENSE for details.

Support

• 📖 Documentation (coming soon)
• 💬 Discord Community
• 🐛 Issue Tracker
• 📧 Email: [e-mail]

Powered By

NativeBlend CLI is powered by NativeBlend's proprietary AI agent infrastructure, combining industry knowledge with advanced vision models for iterative 3D generation.

---

Made with ❤️ by the NativeBlend team