AWS CLI Guide for Bedrock

This tutorial explains how to use the AWS Command Line Interface (CLI) to work with Amazon Bedrock services.

Prerequisites

AWS CLI installed (version 2.x recommended)
Configured AWS profile with Bedrock permissions
Basic familiarity with command line interfaces

Setting Up AWS CLI for Bedrock

Configure Your Profile

For local testing, we always use the aws profile as documented in our CLAUDE.md file:

aws configure --profile aws

Enter your AWS Access Key ID, Secret Access Key, default region (use a region where Bedrock is available, such as us-west-2), and preferred output format (json recommended).

Verify Bedrock Access

Confirm you can access Bedrock services:

aws bedrock list-foundation-models --profile aws

If successful, you’ll see a JSON response listing available models.

Common Bedrock CLI Commands

List Available Models

aws bedrock list-foundation-models --profile aws

Get Model Details

aws bedrock get-foundation-model \
  --model-identifier anthropic.claude-3-sonnet-20240229-v1:0 \
  --profile aws

Check Model Throughput Capacity

aws bedrock get-foundation-model-throughput-capacity \
  --model-id anthropic.claude-3-sonnet-20240229-v1:0 \
  --profile aws

List Custom Models

If you have custom models:

aws bedrock list-custom-models --profile aws

Check Service Quotas

To view your current quotas:

aws service-quotas list-service-quotas \
  --service-code bedrock \
  --profile aws

To request a quota increase:

aws service-quotas request-service-quota-increase \
  --service-code bedrock \
  --quota-code L-12345678 \
  --desired-value 100 \
  --profile aws

Run Inference (Text Example)

For simple testing, you can invoke a model directly via CLI:

aws bedrock-runtime invoke-model \
  --model-id anthropic.claude-3-sonnet-20240229-v1:0 \
  --body '{"anthropic_version":"bedrock-2023-05-31","max_tokens":1000,"messages":[{"role":"user","content":"Explain quantum computing in simple terms."}]}' \
  --profile aws \
  output.json

This saves the response to output.json.

Streaming Inference

For streaming responses (requires further processing):

aws bedrock-runtime invoke-model-with-response-stream \
  --model-id anthropic.claude-3-sonnet-20240229-v1:0 \
  --body '{"anthropic_version":"bedrock-2023-05-31","max_tokens":1000,"messages":[{"role":"user","content":"Write a short story about robots."}]}' \
  --profile aws \
  output_stream.json

Working with Different Model Formats

Different models require different request formats. Here are examples for major model families:

Anthropic Claude

aws bedrock-runtime invoke-model \
  --model-id anthropic.claude-3-sonnet-20240229-v1:0 \
  --body '{
    "anthropic_version": "bedrock-2023-05-31",
    "max_tokens": 1000,
    "messages": [
      {"role": "user", "content": "Write a poem about machine learning."}
    ]
  }' \
  --profile aws \
  claude_output.json

Meta Llama 2

aws bedrock-runtime invoke-model \
  --model-id meta.llama2-13b-chat-v1 \
  --body '{
    "prompt": "<s>[INST] Write a poem about machine learning. [/INST]",
    "max_gen_len": 1000,
    "temperature": 0.7,
    "top_p": 0.9
  }' \
  --profile aws \
  llama_output.json

Amazon Titan

aws bedrock-runtime invoke-model \
  --model-id amazon.titan-text-express-v1 \
  --body '{
    "inputText": "Write a poem about machine learning.",
    "textGenerationConfig": {
      "maxTokenCount": 1000,
      "temperature": 0.7,
      "topP": 0.9
    }
  }' \
  --profile aws \
  titan_output.json

Processing Responses

To extract just the generated text from the response JSON:

cat claude_output.json | jq -r '.content[0].text'
cat llama_output.json | jq -r '.generation'
cat titan_output.json | jq -r '.results[0].outputText'

Creating Shell Scripts for Common Tasks

Here’s an example shell script that tests multiple models with the same prompt:

#!/bin/bash

PROFILE="aws"
PROMPT="Explain the concept of cloud computing."

# Claude
echo "Testing Claude 3 Sonnet..."
aws bedrock-runtime invoke-model \
  --model-id anthropic.claude-3-sonnet-20240229-v1:0 \
  --body "{\"anthropic_version\":\"bedrock-2023-05-31\",\"max_tokens\":1000,\"messages\":[{\"role\":\"user\",\"content\":\"$PROMPT\"}]}" \
  --profile $PROFILE \
  claude_response.json

# Llama 2
echo "Testing Llama 2..."
aws bedrock-runtime invoke-model \
  --model-id meta.llama2-13b-chat-v1 \
  --body "{\"prompt\":\"<s>[INST] $PROMPT [/INST]\",\"max_gen_len\":1000}" \
  --profile $PROFILE \
  llama_response.json

# Titan
echo "Testing Titan..."
aws bedrock-runtime invoke-model \
  --model-id amazon.titan-text-express-v1 \
  --body "{\"inputText\":\"$PROMPT\",\"textGenerationConfig\":{\"maxTokenCount\":1000}}" \
  --profile $PROFILE \
  titan_response.json

echo "All tests complete. Results saved to *_response.json files."

Make the script executable with chmod +x test_models.sh and run it with ./test_models.sh.

Next Steps

Now that you’re familiar with the basic AWS CLI commands for Bedrock, you can:

Create more advanced scripts for automation
Integrate these commands into your development workflow
Set up monitoring scripts to track quota usage
Explore advanced topics like batch processing