​
Using AWS Bedrock with Flowstack

AWS Bedrock can be integrated with Flowstack, but currently, only the Python SDK implementation is available. REST API and Node.js SDK implementations are not provided.

​
Python SDK Implementation

To use AWS Bedrock with Flowstack in Python, follow these steps:

  1. Install Boto3 : Ensure you have the Boto3 library installed. Boto3 is the Amazon Web Services (AWS) SDK for Python.
pip install boto3
  1. Python Code Implementation : Use the following Python code to set up the AWS Bedrock client and make a request. Replace "YOUR_AWS_ACCESS_KEY_ID" and "YOUR_AWS_SECRET_ACCESS_KEY" with your AWS credentials.
import boto3
import json

def _add_header(request, **kwargs):
    request.headers.add_header('Flowstack-Auth', 'Bearer YOUR_FLOWSTACK_KEY')
    request.headers.add_header('AWS-Secret-Access-Key', "YOUR_AWS_SECRET_ACCESS_KEY")

bedrock = boto3.client(service_name='bedrock-runtime', 
                        region_name='us-east-1', 
                        endpoint_url='https://aws-bedrock.flowstack.ai',
                        aws_access_key_id="YOUR_AWS_ACCESS_KEY_ID",
                        aws_secret_access_key="YOUR_AWS_SECRET_ACCESS_KEY")

event_system = bedrock.meta.events
event_system.register_first('before-sign.*.*', _add_header)
modelId = 'your-model-id' # Replace with your model ID

body = json.dumps({
    "prompt": "Your prompt here", 
    "maxTokens": 200,
    "temperature": 0.5,
    "topP": 0.5
})

response = bedrock.invoke_model(
    modelId=modelId,
    body=body
)

response.get('body').read()

Replace "YOUR_FLOWSTACK_KEY" and "your-model-id" with the appropriate Flowstack key and model ID.

AI21Vertex AI