Getting Started: Installation¤
Prerequisites¤
- Python 3.10+
piporpoetry
Installation via pip¤
pip install crewmaster
Optional extras¤
CrewMaster provides optional dependency groups for specific use cases:
# Install with the sandbox UI (Streamlit-based interactive evaluation)
pip install "crewmaster[sandbox-ui]"
# Install with PydanticAI runtime driver support
pip install "crewmaster[pydantic-ai]"
# Install all optional dependencies
pip install "crewmaster[sandbox-ui,pydantic-ai]"
Installation via poetry¤
poetry add crewmaster
With extras in Poetry¤
poetry add crewmaster -E sandbox-ui -E pydantic-ai
Verifying the Installation¤
To ensure CrewMaster is correctly installed, open a Python interpreter and run a simple import:
import crewmaster
print("Crewmaster is ready!")
If this runs without an error, you're all set to start building your AI crews.
Environment Variables¤
CrewMaster v2.0.0 is framework-agnostic — the environment variables you need depend on which runtime driver you choose.
PydanticAIDriver (recommended)¤
Create a .env file at the root of your project:
# Required: your LLM provider API key
OPENAI_API_KEY=<your-api-key>
# Optional: override model defaults
PDAI_MODEL_NAME=openai:gpt-4o
PDAI_TEMPERATURE=0.7
PydanticAI supports many providers (OpenAI, Anthropic, Groq, Ollama, etc.). See PydanticAI documentation for provider- specific environment variable names.
LangChainDriver (legacy)¤
If using the LangChain driver, set the standard LangChain environment
variables such as OPENAI_API_KEY, LANGSMITH_API_KEY, etc. Refer to the
LangChain documentation
for your chosen provider.