Amazon Bedrock Models

A drag-and-drop component for integrating Amazon Bedrock LLMs into your workflow. Simply configure the required parameters and connect the inputs/outputs to other components.

Amazon Bedrock component interface and configuration

AWS Credentials Required: This component requires valid AWS credentials with appropriate permissions to access Amazon Bedrock services. Ensure you have the necessary IAM roles and policies configured before using this component in production.

Component Inputs

Input: Text input for the model
Example: "Explain the concept of quantum computing in simple terms."
System Message: System prompt to guide model behavior
Example: "You are a helpful assistant who explains complex topics in simple language."
Stream: Toggle for streaming responses
Example: true (for real-time token streaming) or false (for complete response)
Model ID: The Bedrock model identifier
Example: "anthropic.claude-3-haiku-20240307", "amazon.titan-text-express-v1"
AWS Access Key ID: Your AWS access key
Example: "AKIAIOSFODNN7EXAMPLE"
AWS Secret Access Key: Your AWS secret key
Example: "wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY"
AWS Session Token: Optional temporary session token
Example: "AQoEXAMPLEH4aoAH0gNCAPyJxz4BlCFFxWNE1OPTgk5TthT..."
Credentials Profile Name: AWS credentials profile (optional)
Example: "bedrock-profile"
Region Name: AWS region where Bedrock is deployed
Example: "us-east-1", "us-west-2"

Component Outputs

Text: Generated text output
Example: "Quantum computing is a type of computing that uses quantum mechanics to process information..."
Language Model: Model information and metadata
Example: model_id: anthropic.claude-3-haiku-20240307, usage: {input_tokens: 40, output_tokens: 120, total_tokens: 160}

Additional Settings

Model Kwargs

Key-value pairs for additional model parameters

{
  "temperature": 0.7,
  "max_tokens": 500,
  "top_p": 0.9
}

Endpoint URL

Custom endpoint URL (optional)

Default: AWS regional endpoint based on Region Name
Example: "https://bedrock-runtime.us-east-1.amazonaws.com"
Use case: For VPC endpoints or custom AWS endpoint configurations

Supported Models

Anthropic Claude Models

High-performance models from Anthropic available via Bedrock

- anthropic.claude-3-sonnet-20240229
- anthropic.claude-3-haiku-20240307
- anthropic.claude-3-opus-20240229
- anthropic.claude-instant-v1

Amazon Titan Models

Amazon's proprietary LLMs

- amazon.titan-text-express-v1
- amazon.titan-text-lite-v1

Other Provider Models

Additional models available through Bedrock

- meta.llama2-13b-chat-v1
- meta.llama2-70b-chat-v1
- cohere.command-text-v14
- ai21.j2-ultra-v1

Implementation Example

// Basic configuration with environment variables
const bedrockClient = {
  modelId: "anthropic.claude-3-haiku-20240307",
  regionName: "us-east-1",
  // AWS credentials from environment variables
  // AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY automatically used
};

// Advanced configuration
const advancedBedrockClient = {
  modelId: "anthropic.claude-3-sonnet-20240229",
  regionName: "us-west-2",
  awsAccessKeyId: process.env.AWS_ACCESS_KEY_ID,
  awsSecretAccessKey: process.env.AWS_SECRET_ACCESS_KEY,
  awsSessionToken: process.env.AWS_SESSION_TOKEN,
  stream: true,
  modelKwargs: {
    temperature: 0.5,
    max_tokens: 2000,
    top_p: 0.9
  },
  endpointUrl: "https://bedrock-runtime.us-west-2.amazonaws.com"
};

// Usage example
async function generateResponse(input) {
  const response = await bedrockComponent.generate({
    input: input,
    systemMessage: "You are a helpful assistant that explains complex concepts clearly.",
    modelId: "anthropic.claude-3-haiku-20240307"
  });
  
  return response.text;
}

Use Cases

Enterprise Applications: Build AI solutions that leverage AWS security and compliance standards
Multi-Model Workflows: Create applications that can switch between different model providers
Content Generation: Generate articles, summaries, and creative content
Conversational Agents: Build chatbots and virtual assistants using streaming responses
AWS Integration: Integrate with other AWS services in a unified environment

Useful Resources

Best Practices

Use environment variables for AWS credentials
Test with small inputs first to validate your setup
Monitor token usage and costs through AWS billing dashboard
Use appropriate regional endpoints to minimize latency
Implement error handling in workflows to manage API failures
Consider using AWS IAM roles instead of access keys for production deployments

Documentation