Skip to content
CentralMind

Amazon Bedrock

The Amazon Bedrock provider enables integration with Claude 3.7 Sonnet and other Anthropic models hosted on AWS Bedrock. This guide covers setup, configuration, authentication, and usage options.

Overview

Amazon Bedrock is a fully managed service that offers a selection of high-performing foundation models (FMs), including Anthropic’s Claude models. The provider leverages AWS services to securely access these models with robust authentication and flexible configuration options.

AWS Authorization and Authentication

The provider uses standard AWS authentication mechanisms to manage access to Bedrock services. To ensure secure and seamless integration, make sure you have the necessary IAM permissions. AWS supports multiple methods for credential management, including:

  • Environment Variables: Set AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY to provide your credentials.
  • AWS Credentials File: Configure your credentials in the ~/.aws/credentials file.
  • EC2 Instance Profiles: For applications running on EC2, IAM roles can provide temporary credentials.

For more details on configuring AWS credentials and best practices, refer to the following resources:

CLI Installation

The AWS CLI is a crucial tool for managing your AWS credentials and interacting with AWS services. If you haven’t installed it yet, follow the instructions in the official guide:

After installation, configure your credentials using:

Terminal window
aws configure

Example Usage

Below is a basic example of how to use the provider with a connection configuration file:

Terminal window
./gateway discover \
  --ai-provider bedrock \
  --config connection.yaml

Region Configuration

The provider determines which AWS region to use in the following order:

  1. Explicitly configured region via the --bedrock-region flag
  2. Environment variable BEDROCK_REGION
  3. Environment variable AWS_REGION
  4. Default fallback to us-east-1

Example with an explicit region:

Terminal window
./gateway discover \
  --ai-provider bedrock \
  --bedrock-region us-west-2 \
  --config connection.yaml

Model Selection

By default, the Bedrock provider uses us.anthropic.claude-3-7-sonnet-20250219-v1:0. You can specify a different model using one of the following methods:

  1. Command-line Flag: Use the --ai-model flag.
  2. Environment Variable: Set the BEDROCK_MODEL_ID.

Examples:

Terminal window
# Specify model via command line
./gateway discover \
  --ai-provider bedrock \
  --ai-model us.anthropic.claude-3-7-sonnet-20250219-v1:0 \
  --config connection.yaml


# Or via environment variable
export BEDROCK_MODEL_ID=us.anthropic.claude-3-7-sonnet-20250219-v1:0
./gateway discover \
  --ai-provider bedrock \
  --config connection.yaml

Advanced Configuration

Reasoning Mode

Enable Claude’s “thinking” mode for complex reasoning tasks:

Terminal window
./gateway discover \
  --ai-provider bedrock \
  --ai-reasoning=true \
  --config connection.yaml

When reasoning mode is enabled, the provider:

  • Activates Claude’s thinking capability
  • Allocates a thinking token budget of 4096 tokens
  • Sets the temperature to 1.0

Response Length Control

Control the maximum token count in responses:

Terminal window
./gateway discover \
  --ai-provider bedrock \
  --ai-max-tokens 8192 \
  --config connection.yaml

If not specified, the default maximum token count is 64000 tokens.

Temperature Adjustment

Adjust the randomness of responses with the temperature parameter:

Terminal window
./gateway discover \
  --ai-provider bedrock \
  --ai-temperature 0.5 \
  --config connection.yaml

Lower values produce more deterministic outputs, while higher values increase creativity and randomness.

Usage Costs

The Bedrock provider includes cost estimation for Claude Sonnet models:

  • Input tokens: $3.75 per 1 million tokens
  • Output tokens: $15.00 per 1 million tokens

Costs are calculated based on the actual token usage for each request.

  • Start with Claude 3.7 Sonnet: Offers a balance of performance and cost.
  • Manage Token Count: Set a reasonable maximum token count to control costs.
  • Enable Reasoning Mode: Use for complex analytical tasks.
  • Secure Configuration: Use environment variables or AWS credentials files for sensitive settings.
  • Regularly Update IAM Policies: Follow the principle of least privilege to minimize risks.

Additional Resources