agentic-crew

Framework-agnostic AI crew orchestration - declare once, run on CrewAI, LangGraph, or Strands.

Why agentic-crew?

The AI agent ecosystem is fragmented:

• CrewAI - Full-featured but heavyweight
• LangGraph - Great for flows but different API
• Strands - Lightweight but AWS-specific

agentic-crew solves this with a universal crew format that decomposes to any framework.

Installation

# Core (no framework - auto-detects at runtime)
pip install agentic-crew

# With specific framework
pip install agentic-crew[crewai]      # CrewAI (recommended)
pip install agentic-crew[langgraph]   # LangGraph
pip install agentic-crew[strands]     # AWS Strands

# All frameworks
pip install agentic-crew[ai]

Quick Start

1. Define a Crew (YAML)

# .crewai/manifest.yaml
name: my-package
version: "1.0"

crews:
  analyzer:
    description: Analyze codebases
    agents: crews/analyzer/agents.yaml
    tasks: crews/analyzer/tasks.yaml

# crews/analyzer/agents.yaml
code_reviewer:
  role: Senior Code Reviewer
  goal: Find bugs and improvements
  backstory: Expert at code analysis

# crews/analyzer/tasks.yaml
review_code:
  description: Review the provided code for issues
  expected_output: List of findings with severity
  agent: code_reviewer

2. Run It

from agentic_crew import run_crew

# Auto-detects best framework
result = run_crew("my-package", "analyzer", inputs={"code": "..."})

Or from CLI:

agentic-crew run my-package analyzer --input "Review this code: ..."

Framework Decomposition

The magic happens in core/decomposer.py:

from agentic_crew.core.decomposer import detect_framework, get_runner

# See what's available
framework = detect_framework()  # "crewai", "langgraph", or "strands"

# Get a runner
runner = get_runner()  # Auto-selects best
runner = get_runner("langgraph")  # Force specific

# Build and run
crew = runner.build_crew(config)
result = runner.run(crew, inputs)

Framework Priority

1. CrewAI (if installed) - Most features, best for complex crews
2. LangGraph (if CrewAI unavailable) - Good for flow-based logic
3. Strands (fallback) - Lightweight, minimal deps

Package Integration

Any package can define crews in a .crewai/ directory:

my-package/
├── .crewai/
│   ├── manifest.yaml
│   ├── knowledge/
│   │   └── domain_docs/
│   └── crews/
│       └── my_crew/
│           ├── agents.yaml
│           └── tasks.yaml
└── src/

Then run:

agentic-crew run my-package my_crew --input "..."

Use Cases

1. Connector Builder (vendor-connectors)

A crew that scrapes API docs and generates HTTP connectors:

agentic-crew run vendor-connectors connector_builder \
  --input '{"api_docs": "https://docs.meshy.ai/en"}'

2. Code Generation (any project)

Define crews for your specific domain and run them on any framework.

Development

# Install with dev deps
uv sync --extra dev --extra tests --extra crewai

# Run tests
uv run pytest tests/ -v

# Lint
uvx ruff check src/ tests/ --fix

Related Projects

• vendor-connectors - HTTP connector library
• CrewAI - Original crew framework
• LangGraph - Graph-based agents
• Strands - AWS agent framework

License

MIT