Azure OpenAI Express App

This is an example app using Express and Azure OpenAI to stream chat completions.

Installation

git clone https://github.com/mpodwysocki/openai-express.git

cd openai-express
npm install

You will need to create a .env file with the credentials such as what is in the sample.env file for the Express port as well as the Azure OpenAI Credentials.

# Express
PORT=7669

# Azure OpenAI
AZURE_OPENAI_API_KEY=
AZURE_OPENAI_ENDPOINT=
AZURE_OPENAI_DEPLOYMENT=

To run the server, use the following:

npm run dev

Chat Completions

Using cURL

To interact with the API to do chat completions, you can use cURL or Postman such as the following cURL where you can set the role and content to be completed.

curl --location 'http://localhost:7669/api' \
--header 'Content-Type: application/json' \
--data '[{"role":"user","content":"Write me a function that adds two numbers in TypeScript?"}]'

Using the Fetch API

Using the fetch API, you can do the following:

const messages = [
  {
    "role": "user",
    "content":"Write me a function that adds two numbers in TypeScript?"
  }
];

const response = await fetch("/api", {
  method: "POST",
  headers: {
  "Content-Type": "application/json"
  },
  body: JSON.stringify(get(messages))
});

const reader = response.body.pipeThrough(new TextDecoderStream()).getReader();
while (true) {
  const { done, value } = await reader.read()
  if (done) {
    break
  }

  console.log(value);
}

LICENSE

MIT