Skip to main content

​
Add AWS Bedrock to PraisonAI

AWS Bedrock provides access to high-performing foundation models from leading AI companies like Anthropic, Cohere, Meta, Stability AI, and Amazon through a single API.

​
Setup

​
Prerequisites

Make sure you have AWS credentials configured: 
pip install boto3

​
Environment Variables

Set up your AWS credentials: 
export AWS_ACCESS_KEY_ID=your_access_key_id
export AWS_SECRET_ACCESS_KEY=your_secret_access_key
export AWS_REGION=us-east-1

​
Using AWS Bedrock Models

​
Available Models

AWS Bedrock supports various model providers:
  • Anthropic Claude: bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0
  • Anthropic Claude Instant: bedrock/anthropic.claude-instant-v1
  • Amazon Titan: bedrock/amazon.titan-text-express-v1
  • Cohere Command: bedrock/cohere.command-text-v14
  • Meta Llama: bedrock/meta.llama2-70b-chat-v1

​
agents.yaml Configuration

framework: crewai
topic: create movie script about cat in mars
roles:
  researcher:
    backstory: Skilled in finding and organizing information, with a focus on research efficiency.
    goal: Gather information about Mars and cats
    role: Researcher
    llm:
      model: "bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0"
      temperature: 0.7
    tasks:
      gather_research:
        description: Research and gather information about Mars, its environment, and cats, including their behavior and characteristics.
        expected_output: Document with research findings, including interesting facts and information.
    tools:
    - ''

​
Python Code Example

from praisonaiagents import Agent

# Using Anthropic Claude via Bedrock
agent = Agent(
    instructions="You are a helpful assistant",
    llm={
        "model": "bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0",
        "temperature": 0.7
    }
)

# Using Amazon Titan via Bedrock
titan_agent = Agent(
    instructions="You are a helpful assistant",
    llm={
        "model": "bedrock/amazon.titan-text-express-v1",
        "temperature": 0.7
    }
)

response = agent.ask("What is artificial intelligence?")
print(response)

​
IAM Permissions

Ensure your AWS IAM user/role has the necessary permissions to access Bedrock: 
{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "bedrock:InvokeModel",
                "bedrock:InvokeModelWithResponseStream"
            ],
            "Resource": "*"
        }
    ]
}

​
Regional Availability

AWS Bedrock is available in the following regions:
  • us-east-1 (N. Virginia)
  • us-west-2 (Oregon)
  • ap-southeast-1 (Singapore)
  • ap-northeast-1 (Tokyo)
  • eu-central-1 (Frankfurt)
  • eu-west-3 (Paris)
Make sure to set your AWS_REGION environment variable to a supported region.

​
Cost Optimization

AWS Bedrock charges are based on:
  • Input tokens: Text sent to the model
  • Output tokens: Text generated by the model
Consider using smaller models for development and testing to optimize costs.

​
Error Handling

Common errors and solutions:
  • AccessDeniedException: Check your IAM permissions
  • ResourceNotFoundException: Verify the model ID is correct and available in your region
  • ThrottlingException: Implement retry logic with exponential backoff
  • ValidationException: Check your input parameters and format

​
Advanced Configuration

​
Custom Endpoint

For specific regions or custom endpoints: 
agent = Agent(
    instructions="You are a helpful assistant",
    llm={
        "model": "bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0",
        "aws_region": "us-west-2",
        "temperature": 0.7
    }
)

​
Streaming Responses

For real-time responses: 
agent = Agent(
    instructions="You are a helpful assistant",
    llm={
        "model": "bedrock/anthropic.claude-3-5-sonnet-20241022-v2:0",
        "stream": True
    }
)
PraisonAI ChatPraisonAI CodePraisonAI (Multi-Agents)
LiteLLMLiteLLMModels