Last updated: Aug 12, 2025, 01:09 PM UTC

Setting Up Amazon Bedrock for Sasha Studio

Last Updated: January 2025
Time Required: 30-45 minutes
Difficulty: Intermediate

Overview

This guide will help you connect your Amazon Bedrock account to Sasha Studio, enabling AI-powered features while keeping all data processing within your own AWS infrastructure.

What You'll Achieve

  • Complete data sovereignty - your data never leaves AWS
  • Direct cost control through your AWS account
  • Full audit trail in AWS CloudTrail
  • Compliance with your existing AWS certifications

Prerequisites

Before starting, ensure you have:

  1. AWS Account with appropriate permissions
  2. Amazon Bedrock service enabled in your desired region
  3. Claude model access approved (see Step 1 below)
  4. IAM permissions to create users/roles and policies
  5. Budget alerts configured (recommended)

Step-by-Step Setup

Step 1: Enable Amazon Bedrock and Request Model Access

  1. Sign in to AWS Console

    https://console.aws.amazon.com

  2. Navigate to Amazon Bedrock

    • Search for "Bedrock" in the AWS services search bar
    • Select your preferred region (e.g., us-east-1, eu-west-1)

  3. Request Model Access

    • Click "Model access" in the left sidebar
    • Click "Manage model access"
    • Select the following Anthropic Claude models:
      • ☑️ Claude 3 Haiku
      • ☑️ Claude 3 Sonnet
      • ☑️ Claude 3.5 Sonnet
      • ☑️ Claude 3 Opus (optional, higher cost)
    • Click "Request model access"
    • Wait for approval (usually instant, can take up to 24 hours)

  4. Verify Access

    • Refresh the page
    • Ensure all requested models show "Access granted"

Step 2: Create IAM User for Sasha Studio

Option A: Using AWS Console (Recommended for beginners)

  1. Navigate to IAM

    • Go to IAM service in AWS Console
    • Click "Users" → "Add users"

  2. Create User

    • User name: sasha-studio-integration
    • Access type: ☑️ Programmatic access
    • Click "Next: Permissions"

  3. Attach Permissions

    • Choose "Attach policies directly"
    • Click "Create policy"
    • Switch to "JSON" tab
    • Copy and paste this policy:
{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Sid": "SashaStudioBedrockAccess",
      "Effect": "Allow",
      "Action": [
        "bedrock:InvokeModel",
        "bedrock:InvokeModelWithResponseStream"
      ],
      "Resource": [
        "arn:aws:bedrock:*:*:model/anthropic.claude-3-haiku*",
        "arn:aws:bedrock:*:*:model/anthropic.claude-3-sonnet*",
        "arn:aws:bedrock:*:*:model/anthropic.claude-3-5-sonnet*",
        "arn:aws:bedrock:*:*:model/anthropic.claude-3-opus*"
      ]
    },
    {
      "Sid": "SashaStudioModelDiscovery",
      "Effect": "Allow",
      "Action": [
        "bedrock:ListFoundationModels",
        "bedrock:GetFoundationModel"
      ],
      "Resource": "*"
    }
  ]
}

  1. Name and Create Policy

    • Policy name: SashaStudioBedrockPolicy
    • Description: Allows Sasha Studio to invoke Claude models via Bedrock
    • Click "Create policy"

  2. Attach Policy to User

    • Return to user creation
    • Search for SashaStudioBedrockPolicy
    • Select it and click "Next"
    • Review and click "Create user"

  3. Save Credentials

    • IMPORTANT: Save these credentials immediately
    • Access key ID: AKIA... (example)
    • Secret access key: (shown once only)
    • Download .csv file as backup

Option B: Using AWS CLI

# Create user
aws iam create-user --user-name sasha-studio-integration

# Create and attach policy
aws iam put-user-policy \
  --user-name sasha-studio-integration \
  --policy-name SashaStudioBedrockPolicy \
  --policy-document file://bedrock-policy.json

# Create access key
aws iam create-access-key --user-name sasha-studio-integration

Step 3: Configure Budget Alerts (Recommended)

  1. Navigate to AWS Budgets

    • Search for "Budgets" in AWS Console
    • Click "Create budget"

  2. Set Up Monthly Budget

    • Budget type: Cost budget
    • Name: Sasha-Studio-Bedrock
    • Period: Monthly
    • Amount: Set based on expected usage
    • Alerts: 80% and 100% thresholds

  3. Estimate Costs

    Usage Level Monthly Tokens Estimated Cost (Sonnet)
    Light 1M input, 200K output ~$6
    Medium 10M input, 2M output ~$60
    Heavy 50M input, 10M output ~$300
    Enterprise 200M input, 40M output ~$1,200

Step 4: Provide Information to Sasha Studio

Send the following information to your Sasha Studio administrator:

Organization Information:
  Company Name: [Your Company Name]
  Contact Email: [Admin Email]
  AWS Account ID: [12-digit ID]

Bedrock Configuration:
  Preferred Region: [e.g., us-east-1]
  Preferred Model: [claude-3-sonnet or claude-3-haiku]
  
Access Credentials:
  Access Key ID: [AKIA...]
  Secret Access Key: [Encrypted or shared securely]
  
Compliance Requirements:
  - [ ] HIPAA
  - [ ] SOC2
  - [ ] GDPR
  - [ ] Other: [Specify]

Secure Sharing Options:

  • Use encrypted email
  • Share via secure file transfer
  • Use AWS Secrets Manager
  • Provide during video call

Step 5: Verification

Once Sasha Studio configures your container:

  1. You'll receive confirmation with:

    • Your dedicated Sasha Studio URL
    • Container ID for reference
    • Support contact information

  2. Test the integration:

    • Log in to your Sasha Studio instance
    • Try a simple query
    • Verify it's working correctly

  3. Monitor usage in AWS:

    • Check CloudWatch metrics
    • Review CloudTrail logs
    • Monitor budget alerts

Monitoring Your Usage

CloudWatch Dashboard

Create a dashboard to monitor Bedrock usage:

  1. Go to CloudWatch → Dashboards
  2. Create new dashboard: Sasha-Bedrock-Usage
  3. Add widgets for:
    • Invocation count
    • Input/output tokens
    • Latency metrics
    • Error rates

Cost Explorer

Track spending:

  1. AWS Cost Explorer → Create report
  2. Filter by:
    • Service: Amazon Bedrock
    • Tag: Environment=SashaStudio (if tagged)
  3. Group by: Model ID

Security Best Practices

1. Enable MFA

  • Enable MFA on the IAM user
  • Use hardware keys if possible

2. Rotate Credentials

  • Set up 90-day rotation reminder
  • Use AWS Secrets Manager for automation

3. Restrict by IP (Optional)

Add IP restriction to IAM policy:

"Condition": {
  "IpAddress": {
    "aws:SourceIp": ["Your-Sasha-Container-IP"]
  }
}

4. Enable CloudTrail Logging

  • Ensure CloudTrail is enabled
  • Set up alerts for unusual activity

5. Use Least Privilege

  • Only grant access to required models
  • Don't use AdministratorAccess

Troubleshooting

Issue: "Model access not granted"

Solution: Wait 24 hours after requesting access, or contact AWS Support

Issue: "Access denied" errors

Solution: Verify IAM policy is correctly attached and user has proper permissions

Issue: "Region not supported"

Solution: Check Bedrock availability in your region, consider us-east-1 or eu-west-1

Issue: High latency

Solution: Choose a region closer to your Sasha Studio deployment

Issue: Unexpected costs

Solution: Review CloudWatch metrics, consider using Claude 3 Haiku for simple tasks

Cost Optimization Tips

1. Model Selection Strategy

  • Claude 3 Haiku: Simple queries, summaries (10x cheaper)
  • Claude 3 Sonnet: Standard operations, code generation
  • Claude 3 Opus: Complex analysis only (5x more expensive)

2. Implement Caching

  • Enable prompt caching where available
  • Reuse common prompts

3. Monitor Token Usage

  • Track input vs output ratios
  • Optimize prompt length
  • Use system prompts efficiently

4. Set Usage Limits

  • Configure daily/monthly limits
  • Set up auto-shutoff thresholds

📞 Getting Help

Sasha Studio Support

AWS Support

Common Resources

Setup Complete!

Congratulations! Your Amazon Bedrock integration with Sasha Studio is now configured. You can:

  • Start using AI features immediately
  • Monitor usage in real-time through AWS
  • Maintain complete control over your data
  • Scale usage up or down as needed

Remember: All AI processing happens within your AWS account. Sasha Studio never stores or processes your data outside of your infrastructure.

Checklist Summary

Before connecting:

  • AWS account ready
  • Bedrock service enabled
  • Model access approved
  • IAM user created
  • Policy attached
  • Credentials saved securely
  • Budget alerts configured
  • Information sent to Sasha Studio

After connecting:

  • Access verified
  • Test queries working
  • Monitoring dashboard created
  • Team members trained
  • Documentation bookmarked

Questions? Contact your Sasha Studio administrator or refer to our FAQ document.