AWS Bedrock

AWS Bedrock provides access to multiple foundation models through a unified API. Use it as Mona's LLM backend for enterprise deployments.

Prerequisites

• AWS account with Bedrock access
• IAM permissions for bedrock:InvokeModel and bedrock:Converse

Installation

The Bedrock extra is included in the [all] install. If you installed the minimal bundle:

cd ~/.monoclaw/monoclaw-runtime
uv pip install -e ".[bedrock]"

Authentication

Option 1: IAM role (recommended for EC2/ECS/Lambda)

If running on AWS infrastructure, attach an IAM role with Bedrock permissions.

Option 2: AWS CLI credentials

aws configure

Enter your Access Key ID, Secret Access Key, and default region.

Option 3: Environment variables

export AWS_ACCESS_KEY_ID="your-key"
export AWS_SECRET_ACCESS_KEY="your-secret"
export AWS_REGION="us-east-1"

Configure MonoClaw

monoclaw model

Choose AWS Bedrock and select your model.

Or manually:

monoclaw config set model.default "bedrock/anthropic.claude-sonnet-4"

Available models

Cross-region inference

Bedrock supports cross-region inference for high availability:

model:
  bedrock:
    region: "us-east-1"
    fallback_regions:
      - "us-west-2"
      - "eu-west-1"

Guardrails

Apply AWS Bedrock Guardrails for content filtering:

model:
  bedrock:
    guardrail_id: "your-guardrail-id"
    guardrail_version: "1"

Configuration reference

model:
  default: "bedrock/anthropic.claude-sonnet-4"
  bedrock:
    region: "us-east-1"
    max_tokens: 4096
    temperature: 0.7
    top_p: 0.9

Troubleshooting