py-orc: AI Pipeline Orchestration Framework

Python Version License

py-orc is a powerful AI pipeline orchestration framework that provides a unified interface for executing AI workflows defined in YAML. It serves as an intelligent wrapper around LangGraph, MCP (Model Context Protocol), and other AI agent control systems.

🎯 Input-Agnostic Pipelines

Create reusable pipelines that adapt to different inputs dynamically

🤖 Intelligent Ambiguity Resolution

Let AI models resolve ambiguous values using <AUTO> tags

🔧 Tool Integration

Seamlessly integrate real-world tools via MCP protocol

🔄 State Management

Built-in checkpointing and recovery for robust execution

Quick Example

import orchestrator as orc

# Initialize models
orc.init_models()

# Compile a pipeline
pipeline = orc.compile("pipelines/research-report.yaml")

# Execute with different inputs
result = pipeline.run(
    topic="quantum_computing",
    instructions="Focus on error correction"
)

Indices and tables