memra 0.2.0

Declarative framework for enterprise workflows with MCP integration

Memra SDK

A declarative orchestration framework for AI-powered business workflows. Think of it as "Kubernetes for business logic" where agents are the pods and departments are the deployments.

🚀 Team Setup

New team member? See the complete setup guide: TEAM_SETUP.md

This includes:

• Database setup (PostgreSQL + Docker)
• Local development environment
• Testing instructions
• Troubleshooting guide

Quick Start

from memra.sdk.models import Agent, Department, Tool

# Define your agents
data_extractor = Agent(
    role="Data Extraction Specialist",
    job="Extract and validate data",
    tools=[Tool(name="DataExtractor", hosted_by="memra")],
    input_keys=["input_data"],
    output_key="extracted_data"
)

# Create a department
dept = Department(
    name="Data Processing",
    mission="Process and validate data",
    agents=[data_extractor]
)

# Run the workflow
result = dept.run({"input_data": {...}})

Installation

pip install memra

API Access

Memra requires an API key to execute workflows on the hosted infrastructure.

Get Your API Key

Contact [e-mail] for API access.

Set Your API Key

# Set environment variable
export MEMRA_API_KEY="your-api-key-here"

# Or add to your shell profile for persistence
echo 'export MEMRA_API_KEY="your-api-key-here"' >> ~/.zshrc

Test Your Setup

python examples/accounts_payable_client.py

Documentation

Documentation is coming soon. For now, see the examples below and in the examples/ directory.

Example: Propane Delivery Workflow

See the examples/propane_delivery.py file for a complete example of how to use Memra to orchestrate a propane delivery workflow.

Contributing

We welcome contributions! Please see our contributing guide for details.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Examples

├── examples/
│   ├── accounts_payable_client.py  # API-based example
│   ├── accounts_payable.py         # Local example
│   ├── invoice_processing.py       # Simple workflow
│   └── propane_delivery.py         # Domain example
├── memra/                  # Core SDK
├── logic/                  # Tool implementations  
├── local/dependencies/     # Database setup & schemas
└── docker-compose.yml      # Database setup

✨ New: MCP Integration

Memra now supports Model Context Protocol (MCP) integration, allowing you to execute operations on your local infrastructure while leveraging Memra's cloud-based AI processing.

Key Benefits:

• 🔒 Keep sensitive data local - Your databases stay on your infrastructure
• ⚡ Hybrid processing - AI processing in the cloud, data operations locally
• 🔐 Secure communication - HMAC-authenticated requests between cloud and local
• 🛠️ Easy setup - Simple bridge server connects your local resources

Quick Example:

# Agent that uses local database via MCP
agent = Agent(
    role="Data Writer",
    tools=[{
        "name": "PostgresInsert",
        "hosted_by": "mcp",  # Routes to your local infrastructure
        "config": {
            "bridge_url": "http://localhost:8081",
            "bridge_secret": "your-secret"
        }
    }]
)

📖 Complete MCP Integration Guide →