SDK Download

Download pre-configured SDKs with your credentials embedded. No API keys to manage, no configuration files to write. Just download, extract, and start building secure AI agents with cryptographic identity verification built-in.

Zero Configuration: Your downloaded SDK comes with credentials already embedded. No need for API keys, environment variables, or configuration files!

Available SDKs

Python SDK

Full-featured Python SDK with automatic agent registration, action verification, and secure credential storage using keyring.

Python 3.8+•~2.5 MB

Node.js SDK

Coming Soon

TypeScript SDK with full type safety, async/await support, and native cryptographic operations.

Node.js 18+•~1.8 MB

Download Process

  1. 1

    Navigate to SDK Download

    In your AIM dashboard, go to Settings → SDK Download

  2. 2

    Choose Your SDK

    Select "Download Python SDK" (currently available)

  3. 3

    Extract Package

    Extract the downloaded aim_sdk.zip file

  4. 4

    Start Building

    Import and use immediately - no configuration needed!

SDK Download API

Download Python SDK

Request
curl -X GET https://api.opena2a.org/v1/sdk/download \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Accept: application/zip" \
  --output aim_sdk.zip

Download Agent-Specific SDK

Download an SDK pre-configured for a specific agent:

Request
curl -X GET https://api.opena2a.org/v1/agents/agent_2KL9m3nX8fY5pQr7/sdk \
  -H "Authorization: Bearer YOUR_TOKEN" \
  -H "Accept: application/zip" \
  --output agent_sdk.zip

What's Included

SDK Package Contents

aim_sdk/
├── aim_sdk/
│   ├── __init__.py          # Main SDK module
│   ├── agent.py             # Agent management
│   ├── auth.py              # Authentication
│   ├── crypto.py            # Ed25519 operations
│   ├── verification.py      # Action verification
│   └── config.py            # Embedded credentials
├── examples/
│   ├── basic_agent.py       # Simple agent example
│   ├── secure_actions.py    # Action verification
│   └── mcp_integration.py   # MCP server example
├── requirements.txt         # Python dependencies
├── README.md               # Getting started guide
└── LICENSE                 # MIT License

Embedded Credentials

Security Note: Your SDK contains embedded credentials specific to your organization. Keep your SDK package secure and never commit it to public repositories.

Each SDK download includes credentials automatically configured:

# aim_sdk/config.py (auto-generated)
CONFIG = {
    "api_url": "https://api.opena2a.org",
    "organization_id": "org_7hJ3m5nY8qM4pS6u",
    "sdk_token": "sdk_9kL4n6oZ0rN5sT8v",  # Refresh token
    "token_expires": "2025-01-15T00:00:00Z",
    "auto_refresh": True,
    "keyring_backend": "system"
}

# Credentials are encrypted and stored in system keyring
# Never hardcode sensitive data in your code!

Using the SDK

Auto-Detection: The SDK automatically detects your agent type from Python imports (langchain, crewai, anthropic, openai, etc.). No configuration needed!

Full-Featured Registration

Python Example
from aim_sdk import secure, AgentType

# Full-featured agent registration
agent = secure(
    "my-ai-assistant",
    agent_type=AgentType.LANGCHAIN,  # CREWAI, AUTOGEN, GPT, CLAUDE, etc.
    capabilities=["db:read", "api:call"],
    version="1.0.0",  # Note: version defaults to "1.0.0" if undeclared
    description="Customer support AI agent",
    tags=["production", "customer-facing", "gpt-4", "support-team"],
    metadata={
        "model": "gpt-4",
        "department": "support"
    },
    mcp_servers=["github", "filesystem"],  # Remove to use mcp auto detect
)

# All actions are now verified, logged, and monitored
# Risk level auto-detected from capability name (db:read → low)
@agent.perform_action(capability="db:read")
def get_customer(customer_id: str):
    return {"id": customer_id, "name": "Jane Doe"}

result = get_customer("cust-123")

Auto Risk Detection

Risk levels are automatically detected from capability patterns - no need to specify manually:

  • • db:read, weather:fetch → low
  • • db:write, api:call → medium
  • • file:delete, admin:modify → high
  • • payment:process, crypto:sign → critical

System Requirements

Python SDK

  • • Python 3.8 or higher
  • • keyring package (for secure credential storage)
  • • cryptography >= 41.0.0 (Ed25519 support)
  • • requests >= 2.31.0 (API communication)
  • • Operating System: Windows, macOS, Linux
  • • Memory: 50 MB minimum
  • • Disk Space: 10 MB for SDK files

SDK Token Management

SDK tokens are long-lived refresh tokens that automatically rotate for security. The SDK handles token refresh transparently:

# Token refresh happens automatically
agent = secure("my-agent")

# SDK checks token expiration before each API call
# If expired, it automatically refreshes using the refresh token
# New tokens are stored securely in the system keyring

# View token status
print(agent.token_status())
# Output: {"valid": true, "expires_in_hours": 168}

# Force token refresh (rarely needed)
agent.refresh_token()

Troubleshooting

Common Issues

Keyring not available

Install keyring: pip install keyring

On Linux, you may need: pip install keyrings.alt

Token expired

SDK tokens auto-refresh. If refresh fails, download a new SDK package.

Agent exists but no local credentials

If you run the SDK from a different machine or directory, credentials are automatically regenerated. The SDK will generate a new keypair and update the server - no action needed!

Import errors

Ensure you're in the correct directory and all dependencies are installed:

pip install -r requirements.txt

Permission denied

Your user role may lack permissions. Contact your admin for access.

Next Steps

Quick Start Guide

Complete walkthrough of using the SDK to build your first agent.

SDK Token Management

Manage, rotate, and monitor your SDK tokens.

Python Integration

Deep dive into Python SDK features and best practices.

Agent Management

Register and manage agents using the SDK.