Build a ChatGPT on AWS

Introduction
ChatGPT is based on OpenAI's GPT (Generative Pre-trained Transformer) architecture. It leverages deep learning to understand and generate human-like text based on the input it receives. Building a ChatGPT involves setting up a similar environment where your model can learn from data and interact with users in a conversational manner.

Service used:

API Gateway
lambda function
Bedrock
IAM Role

1. Create a bedrock:

It provides a foundation models:

• base models
• custom models
• imported models the request access should be provided before using.

click on base model

• no access to any model

Step 1:
To providing access click on 'enable all models'

Step 2:
uncheck the Anthropic models.

2.create an IAM role

• click roles
• choose lambda function as service.

• add permissions (bedrockFullAccess)(cloudwatchlogFullAccess)

• provide role name.

3.create a lambda function:

• create a function name.
• set the run time
• choose existing role. function is created and click configuration to increase the timeout.

this is our request body the AI model

create a lambda code and paste it on the lambda_function.

4.create a API Gateway:
-create a API gateway to create a endpoints

• click on restAPI
• provide a name.

click regional and create a resource.

• create a method.
• choose method type "POST"
• check the lambda proxy
• choose the lambda function.
• click create the method.

Deploy the APT

• deploy the API to a new stage the API endpoint is create!

this API endpoint will call the lambda function to generate the response.

• copy the lambda endpoint and paste it in the POSTMAN tool.
• select the method as post.