This article was initially published on my Medium Page.

In my last article, I showed how I built a talking chat thanks to ChatGPT, AWS, and Unity3D. We will reuse the main components to create a chat showing images generated by Dall-E, the OpenAI images generation system.

Here is a video of the final result:

General Architecture

This time, it will be much more straightforward. We use API Gateway to expose our endpoint, Cognito for the authentication, and Lambda to perform the OpenAI API requests:

Notes:

• OpenAI stores the image created by Dall-E in its repository and returns the URL.
• The Unity client app reads the image directly from the URL.

AWS Implementation

This is our new Lambda function:

Notes:

• We re-use the openai layer we created in my previous article, and the same environment variable openai_api_key.
• We use the image completion endpoint of OpenAI, as specified in the documentation.
• The default image size generation by the endpoint is 1024×1024. We will show the image inside a chat message; then, we can work with a lower resolution (512×512).

Unity implementation

Compared with my previous article, we use the same application with a few changes.

First, our Lambda function only returns the image URL, so we need to change the result data class:

Then, we need to show the image generated by the OpenAI API, so we replace the message Text component by an empty Image in the message object:

Note that the Image width is 512px, the same size we specify in the Lambda function.

And we retrieve the image thanks to the GetTexture function:

Note: Once we have received the texture, we convert it into a Sprite and assign it to the Image of the new instantiated chat message.

Costs

Let’s check with the AWS Calculator what our system cost would be for a very pessimistic scenario: you love the app and perform 50 daily, 1,500 requests monthly.

• Cognito: This project only has one MAU (monthly active user). Cost: 0.00 USD
• API Gateway: With 1,500 requests to our REST API, the monthly cost is 0,00 USD
• Lambda: With 1,500 requests with an average time of 5 seconds and a 1,024 MB of memory allocation, the monthly cost is 0.00 USD; excellent!
• OpenAI: According to the OpenAI pricing sheet, each image generated costs 0.018 USD for a 512×512 resolution. Then, for 1,500 images generated monthly, we have a cost of 27.00 USD.

Total: The total bill for our system would be 27.00 USD monthly, with a pessimistic scenario. That is expensive, but image generation is a complex process requiring high resources, so it can be understandable.

Closing Thoughts

In this article, we could figure out how to build an entire cloud architecture on AWS and how easy it is to integrate the OpenAI API with a Lambda function. We could also evaluate the cost of the whole system thanks to the AWS Calculator.

Every code of this article has been tested using Unity 2021.3.3 and Visual Studio Community 2022 for Mac. The mobile device I used to run the Unity app is a Galaxy Tab A7 Lite with Android 11.

All ids and tokens shown in this article are fake or expired; if you try to use them, you will not be able to establish any connections.

You can download the Unity package of the client app specially designed for this article.

A special thanks to Gianca Chavest for designing the fantastic illustration.