AWS Lambda with Docker

Deploy cryptocom-agent-client as a serverless AWS Lambda function using Docker.

Overview

This example demonstrates how to:

• Deploy a cryptocom-agent-client powered AI agent to AWS Lambda

• Use Docker containers for Lambda deployment

• Create custom tools (functions) the agent can invoke

• Query blockchain data through the Crypto.com Developer Platform

Features

• AI Agent: Uses OpenAI GPT-4o-mini with cryptocom-agent-client

• Custom Tools:

  • get_current_time() - Returns current local and UTC time

  • fibonacci(n) - Calculates the nth Fibonacci number

• Blockchain Integration: Query Cronos blockchain data via Developer Platform Dashboard API

• Serverless: Auto-scaling, pay-per-use AWS Lambda deployment

Prerequisites

• AWS CLI configured with appropriate credentials

• Docker Desktop running

• jq for JSON processing

• AWS account with permissions for Lambda, ECR, and IAM

• OpenAI API key

• Crypto.com Developer Platform API key

Quick Start

1. Configure Environment

cp .env.example .env

Edit .env with your credentials:

# AWS (use profile or access keys)
AWS_PROFILE=default
AWS_DEFAULT_REGION=us-east-1

# API Keys
OPENAI_API_KEY=sk-...
DASHBOARD_API_KEY=your-dashboard-api-key

2. Deploy to AWS

./deploy.sh

This will:

• Build Docker image (ARM64 optimized)

• Create ECR repository and push image

• Create IAM role with basic execution permissions

• Deploy Lambda function (300s timeout, 1024MB memory)

3. Test the Function

# Default prompt
./run.sh

# Custom prompts
./run.sh "What is the current time?"
./run.sh "Calculate fibonacci of 10"
./run.sh "Get the latest block on Cronos"

4. Clean Up

./cleanup.sh

Local Testing

Test the Lambda function locally before deploying:

./test-local.sh "What is the current time?"

Project Structure

.
├── handler.py          # Lambda handler with AI agent
├── requirements.txt    # Python dependencies
├── Dockerfile          # Container configuration
├── deploy.sh           # Deploy to AWS Lambda
├── run.sh              # Invoke the Lambda function
├── test-local.sh       # Test locally with Docker
├── cleanup.sh          # Remove all AWS resources
└── .env.example        # Environment template

Customization

Adding Custom Tools

Edit handler.py to add your own tools:

from crypto_com_agent_client import tool

@tool
def my_custom_tool(param: str) -> str:
    """
    Description of what this tool does.

    Args:
        param: Description of parameter

    Returns:
        Description of return value
    """
    # Your implementation
    return result

Register the tool in the agent initialization:

agent = Agent.init(
    llm_config={...},
    blockchain_config={...},
    plugins={
        "tools": [get_current_time, fibonacci, my_custom_tool],
    },
)

Changing the LLM Model

Modify the llm_config in handler.py:

llm_config={
    "provider": Provider.OpenAI,
    "model": "gpt-4o",  # or "gpt-3.5-turbo", etc.
    "provider-api-key": openai_api_key,
}

Using Different Providers

The cryptocom-agent-client supports multiple LLM providers:

from crypto_com_agent_client.lib.enums.provider_enum import Provider

# AWS Bedrock
llm_config={
    "provider": Provider.AWSBedrock,
    "model": "anthropic.claude-3-haiku-20240307-v1:0",
}

# Google AI
llm_config={
    "provider": Provider.GoogleAI,
    "model": "gemini-pro",
    "provider-api-key": google_api_key,
}

Architecture

┌─────────────┐    ┌─────────────────┐    ┌──────────────┐
│   Client    │───>│  AWS Lambda     │───>│   OpenAI     │
│  (run.sh)   │    │  (handler.py)   │    │   GPT-4o     │
└─────────────┘    └────────┬────────┘    └──────────────┘
                            │
                            v
                   ┌─────────────────┐
                   │   Crypto.com    │
                   │   Dashboard     │
                   │   (Blockchain)  │
                   └─────────────────┘

Lambda Configuration

Setting	Value
Runtime	Python 3.12 (Docker)
Architecture	ARM64
Timeout	300 seconds
Memory	1024 MB

AWS Resource Identifiers

What is ARN?

ARN (Amazon Resource Name) is a unique identifier for any resource in AWS.

Format:

arn:aws:<service>:<region>:<account-id>:<resource-type>/<resource-name>

Example for this Lambda:

arn:aws:lambda:us-east-1:123456789012:function:cryptocom-agent-lambda
         │       │          │              │
         │       │          │              └─ Function name
         │       │          └─ Your AWS account ID
         │       └─ Region (e.g., us-east-1)
         └─ Service (lambda)

Resource Identifiers Used

Resource	Identifier
Lambda Function	cryptocom-agent-lambda
ECR Repository	cryptocom-agent-lambda
IAM Role	lambda-exec-cryptocom-agent
Docker Image	cryptocom-agent-lambda:latest

When invoking the Lambda in the same account/region, scripts use just the function name. AWS resolves the full ARN automatically from your credentials.

Security Notes

• API keys are passed at runtime in the request payload, not stored in Lambda environment

• The IAM role has minimal permissions (AWSLambdaBasicExecutionRole only)

• Never commit .env files with real credentials

Troubleshooting

Docker Build Issues

• Ensure Docker Desktop is running

• On macOS, the build uses ARM64 for Apple Silicon optimization

Lambda Timeout

• Default timeout is 300 seconds

• First invocation may take longer due to cold start

• Subsequent invocations use cached agent instance

ECR Push Failures

• Verify AWS credentials are configured: aws sts get-caller-identity

• Check ECR permissions in your AWS account

Related Resources

• cryptocom-agent-client on PyPI

• Crypto.com Developer Platform

• AWS Lambda Container Images

License

Apache-2.0