Amazon Bedrock

How to set up Amazon Bedrock with Msty Studio

Amazon Bedrock is a fully managed service from AWS that makes it easy to build and scale generative AI applications using foundation models (FMs) from leading AI startups and Amazon. It’s beneficial because it provides a secure and scalable environment for deploying, managing, and fine-tuning AI models, making it ideal for businesses to utilize LLM models and AI applications safely and efficiently.

This guide will walk you through the steps to add Bedrock as an LLM provider in Msty Studio.

Amazon Bedrock Steps
Aurum Perk

These steps cover how to setup Amazon Bedrock in the AWS Management Console. These steps assume you have an AWS account and have access to the Amazon Bedrock service.

  1. Go to AWS Management Console
  2. Search for Amazon Bedrock in the main search bar at the top and select it
  3. Review the Model Catalog to see what models are available
  4. Go to Model Access and request the models you want to use - it may take a few minutes before your request is granted
  5. Grab the AWS Region the model is being enabled in, you will need this for the setup in Msty
  6. Go to API Keys, you'll see that you can either create short-term API keys (recommended by AWS) or long-term API keys

If you wish to use short-term API keys, then you'll need to create temporary access session tokens.

The aws cli can be used for this purpose by running the following command:

aws sts get-session-token

Grab the session token which you'll need when setting up Bedrock in Msty.

  1. Ensure that the user is setup in IAM with the appropriate permissions
  2. Grab the IAM user's Access Key ID and Secret Access Key
  3. Back in Bedrock, verify the model(s) you requested access to have been granted

You can now continue to Msty for adding Bedrock as a provider.

Msty Studio Steps

Remember to grab the IAM user's Access Key ID, Secret Access Key, AWS Region ID, and Session Token if using short-term access.

  1. Go to Msty Studio
  2. Select Model Hub from the left menu
  3. Select Add Provider
  4. Select Amazon Bedrock from the list of providers
  5. Enter the Access Key ID
  6. Enter the Secret Access Key
  7. If using short-term access, enter the Session Token
  8. Enter the AWS Region (ex: us-east-1)
  9. From the Models dropdown, select the models that you enabled in Bedrock that you want to make available in Msty
  10. Click Add Provider

Now, start a new chat and say hello to your new Amazon Bedrock model!

Azure OpenAI
How to set up Azure OpenAI with Msty Studio
Get the latest Ollama version
How to get the latest version of Ollama for Sidecar