- step1: Install via pip with `pip install llama-index-supervisor`
- step2: Import Supervisor from `llama_index_supervisor` in your Python code
- step3: Define workflow nodes (retrieval, LLM call, summarizer) and connect them
- step4: Configure supervision rules, validation schemas, and retry policies
- step5: Invoke `supervisor.run_workflow()` to execute the agent pipeline
- step6: Monitor logs and metrics for performance and error reports