🦞
AI Model Configuration

Configure OpenClaw with Amazon Bedrock

Access multiple AI models through AWS Bedrock with OpenClaw (formerly Moltbot). Enterprise-grade security with your existing AWS infrastructure.

AWS Why Choose Amazon Bedrock?
  • +Multiple models, one API - Access Claude, Llama, Mistral, and Titan through a single AWS endpoint
  • +AWS integration - Uses your existing IAM, VPC, and security policies
  • +Enterprise compliance - SOC2, HIPAA, and other certifications through AWS
  • +Data privacy - Your data never leaves your AWS account
  • +Consolidated billing - All AI costs on your AWS bill

Available Bedrock Models

Claude 3.5 Sonnet (Bedrock)

Recommended

Anthropic's best model, available through AWS Bedrock.

anthropic.claude-3-5-sonnet-20241022-v2:0

Input: $3.00/MTok

Output: $15.00/MTok

Claude 3.5 Haiku (Bedrock)

Fast and affordable Claude through Bedrock.

anthropic.claude-3-5-haiku-20241022-v1:0

Input: $0.80/MTok

Output: $4.00/MTok

Llama 3.1 70B (Bedrock)

Meta's open model, hosted on AWS infrastructure.

meta.llama3-1-70b-instruct-v1:0

Input: $0.99/MTok

Output: $0.99/MTok

Mistral Large (Bedrock)

Mistral's flagship model for complex tasks.

mistral.mistral-large-2407-v1:0

Input: $4.00/MTok

Output: $12.00/MTok

Amazon Titan Text

Amazon's own foundation model. Cost-effective for simple tasks.

amazon.titan-text-express-v1

Input: $0.20/MTok

Output: $0.60/MTok

* Prices per million tokens as of January 2026. Check AWS Bedrock pricing for current rates.

Setup Instructions

1

Enable Bedrock Models

Request access to models in the AWS Console:

  • - Go to AWS Bedrock Console
  • - Navigate to "Model access"
  • - Request access to desired models (Claude, Llama, etc.)
  • - Wait for approval (usually instant for most models)
2

Configure AWS Credentials

Set up AWS credentials with Bedrock access:

aws configure

Or use IAM roles if running on EC2/ECS. Required permission: bedrock:InvokeModel

3

Configure OpenClaw

Add Bedrock configuration to your openclaw.json:

{
  "agent": {
    "provider": "bedrock",
    "model": "anthropic.claude-3-5-sonnet-20241022-v2:0",
    "region": "us-east-1"
  }
}
4

Restart OpenClaw

Apply the configuration:

openclaw restart
Advanced Configuration
{
  "agent": {
    "provider": "bedrock",
    "model": "anthropic.claude-3-5-sonnet-20241022-v2:0",
    "region": "us-east-1",
    "accessKeyId": "AKIA...",
    "secretAccessKey": "...",
    "maxTokens": 4096,
    "temperature": 0.7
  }
}
  • region - AWS region (us-east-1, us-west-2, etc.)
  • accessKeyId - Optional if using IAM roles or AWS CLI profile
  • secretAccessKey - Optional if using IAM roles or AWS CLI profile
  • profile - AWS CLI profile name (alternative to explicit keys)
Switch Between Models

One benefit of Bedrock: easily switch models without changing API credentials:

{
  "agent": {
    "provider": "bedrock",
    "model": "meta.llama3-1-70b-instruct-v1:0",
    "region": "us-east-1"
  }
}

Just change the model ID to switch between providers. No need to manage multiple API keys.

For detailed Bedrock configuration options, see the OpenClaw Bedrock Documentation.

Bedrock Configured!

Now connect your messaging channels and start chatting.