OpenAI Showcase

Summary

This is an example of frontend application that allows to ask to Open AI some questions or write something ( like source code, YAML, scripts...).

The application uses Next.js with React.

This is a Free, Open Source template, under Apache2 License. Feel free to propose improvements and contribute.

Local Development

To develop the service locally you need node.js v16+. To run locally the application you should complete the following tasks

Copy the env file

cp .env.example .env

Configure the OpenAI API Key variable named OPENAI_API_KEY

vi .env

and add your secret key from generated here [https://beta.openai.com/account/api-keys>]

Install and Run

npm install
npm run dev

Then you can open the application on the local url http://localhost:3000/

Build a Docker Image

docker build -t open-ai-showcase:0.0.1 .

to test it

docker run -p 3000:3000 --name open-ai-showcase open-ai-showcase:0.0.1

Then you can open the application on the local url http://localhost:3000/

Usage with Mia-Platform

Open Mia-Platform Marketplace, search for OpenAI Playground, create the microservice, expose it through and endpoint on / base path and enjoy :-)

Configuration

You can configure the following parameters using .env file or Mia-Platform Design section:

• OPENAI_API_KEY, the secret key from [https://beta.openai.com/account/api-keys]
• MODEL_FAMILY, the one of the model listed here [https://beta.openai.com/docs/models/finding-the-right-model]. Default value is text-davinci-002
• MAX_TOKENS, the max number of tokens (see [https://beta.openai.com/tokenizer]). Default values is 1000
• TEMPERATURE, the sampling temperature (see here for details [https://towardsdatascience.com/how-to-sample-from-language-models-682bceb97277]. Default value 0.6