Skip to main content
Athina AI simplifies the integration of custom Large Language Models (LLMs) into your projects on its platform. By connecting with AWS Bedrock, you can access a variety of LLMs from providers like Amazon, Anthropic (Claude), Liquid AI, and more. Custom LLMs enable you to adapt these models to your specific requirements, whether you’re running prompts, conducting evaluations, or building multi-step workflows. This guide will walk you through the process of adding custom LLMs using AWS Bedrock to the Athina platform and show you how to use them effectively in your projects.

​
About AWS Bedrock

Amazon Bedrock is a managed service that provides access to various foundation models (FMs) from leading AI companies through a single API. It allows you to choose and customize models for your specific needs, integrate seamlessly with other AWS services, and ensure your data and credentials remain secure and private. This flexibility and security make it an efficient choice for incorporating large language models (LLMs) into your applications. Now, let’s walk through the steps to add Amazon Titan Text Lite v1 to Athina AI.

​
Add Custom Model

​
Step 1: Get AWS Credentials

1
First, get the AWS Access Key ID, Secret Access Key, Region (e.g., us-east-1), and the Model Identifier from the Model Catalog in AWS Bedrock. The Model Identifier will look something like this:

​
Step 2: Add a Custom Model

1
Log in to Athina AI, go to the Settings page, open the Custom Models tab, and click on the Add Custom Model button, as shown below.

​
Step 3: Configure the Custom Model

1
Next, select AWS Bedrock as the provider, enter the Model Identifier, AWS credentials, and optionally, token cost details, then save the configuration.
Input Token Cost and Output Token Cost won’t be tracked in Analytics if you don’t provide them.

​
Step 4: Test the Model

Go to the Dataset section, select a dataset, run a test prompt with the custom model, and review the output, as shown in the following steps:
1
Open the Run Prompt option to start testing the model:
2
Choose your custom AWS Bedrock model from the model list:
3
Enter a test prompt, then click Save & Run to execute the model:
4
Review and confirm the output to ensure the model works as expected:
With these steps, you can now fully integrate your AWS Bedrock custom model into Athina’s platform, unlocking advanced capabilities for tasks such as running prompts, evaluations, and workflows. This seamless integration allows you to customize workflows to suit your specific needs, improving efficiency and results. Whether you’re processing large datasets, automating tasks, or testing multi-step workflows, your custom model ensures precision and flexibility. If you encounter any issues during the process, double-check your AWS credentials, model identifier, and region configuration. For further support, refer to Athina or AWS Bedrock’s official documentation.