Installation Guide¶

This guide will help you install LlamaCanvas and set up your environment.

Requirements¶

LlamaCanvas requires the following:

Python 3.8 or later
pip (Python package installer)
For video processing: ffmpeg
For GPU acceleration: CUDA-compatible GPU (optional)

Basic Installation¶

The simplest way to install LlamaCanvas is with pip:

pip install llama-canvas

This will install the core functionality of LlamaCanvas without optional dependencies.

Installation with Options¶

LlamaCanvas offers several installation options for different use cases:

API Server¶

To use the web UI and API server:

pip install llama-canvas[api]

Claude Integration¶

To use Anthropic's Claude for multi-modal capabilities:

pip install llama-canvas[claude]

Stable Diffusion Support¶

To use Stable Diffusion for image generation:

pip install llama-canvas[diffusion]

Video Processing¶

To enable video creation and manipulation:

pip install llama-canvas[video]

Development Environment¶

For development and contributing:

pip install llama-canvas[dev]

Documentation¶

For building documentation locally:

pip install llama-canvas[docs]

Full Installation¶

To install all features and dependencies:

pip install llama-canvas[full]

Installation from Source¶

To install from source:

git clone https://github.com/llamasearch/llamacanvas.git
cd llamacanvas
pip install -e ".[full]"

Verifying Installation¶

To verify that LlamaCanvas is installed correctly:

python -c "import llama_canvas; print(llama_canvas.__version__)"

You should see the version number printed out.

Configuration¶

After installation, you may want to set up your configuration. Create a file at ~/.llama_canvas/config.json:

{
  "claude_api_key": "your-anthropic-api-key",
  "default_claude_model": "claude-3-opus-20240229",
  "default_image_width": 512,
  "default_image_height": 512
}

Alternatively, you can set environment variables:

export CLAUDE_API_KEY="your-anthropic-api-key"
export DEFAULT_CLAUDE_MODEL="claude-3-opus-20240229"

See the Configuration Guide for more details.

Troubleshooting¶

Common Issues¶

Missing dependencies¶

If you encounter errors about missing dependencies, try installing with the appropriate option:

pip install llama-canvas[full]

API key not found¶

If you get errors about missing API keys, make sure you've set them up in your configuration file or environment variables.

Video processing issues¶

For video processing issues, ensure ffmpeg is installed and available in your PATH:

# On Ubuntu/Debian
sudo apt-get install ffmpeg

# On macOS with Homebrew
brew install ffmpeg

# On Windows with Chocolatey
choco install ffmpeg

Getting Help¶

If you encounter issues that aren't covered here:

Check the GitHub Issues to see if the problem has been reported
Search the GitHub Discussions for similar problems
Open a new issue or discussion if needed