Investigate Feasibility of User-Connect vs Backend-Connect for ChatGPT Playground

[kiritowu]

To-Dos:

• [x] Find MVP code for both approach
• [x] Evaluate Pros and Cons for both approach
• [x] Discuss and finalize the option

[kiritowu]

According to this (js) and this (py) example, the only authentication to call openai's API is just the apiKey.

import { Configuration, OpenAIApi } from "openai";

const configuration = new Configuration({
  apiKey: process.env.OPENAI_API_KEY,
});
const openai = new OpenAIApi(configuration);

import openai
openai.api_key = "sk-..."

# list engines
engines = openai.Engine.list()

# print the first engine's id
print(engines.data[0].id)

# create a completion
completion = openai.Completion.create(engine="ada", prompt="Hello world")

# print the completion
print(completion.choices[0].text)

There are two options available and the attached are the pros & cons:

1. User use own APIKey:
   • Pros:
     1. Needless to worry about scalability and profitability (arguably we dont need to worry in first place since just hackathon)
     2. Needless for backend, everything happens in front-end (However, at current stage seems like Backend is needed for the tokenization and GPT-like detection)
   • Cons:
     1. Addtional front-end validation to ensure that the user must provide a APIKey and it is valid
2. Use centralised APIKey:
   • Pros:
     1. It works. Can bs about the feasibility also
   • Cons
     1. Should have no latency issue but that is one small concern
     2. Have backend

[kiritowu]

Use Python Backend for APIKey.