_______  _______ _________
(  ___  )(  ____ )\__   __/
| (   ) || (    )|   ) (   
| (___) || (____)|   | |   
|  ___  ||  _____)   | |   
| (   ) || (         | |   
| )   ( || )      ___) (___
|/     \||/       \_______/

ChatGPT-Swift-API

Key Features

The OpenAI ChatGPT API can be applied to virtually any task that involves understanding or generating natural language or code.

  • Create your personal AI
    • You need to create an accout on OpenAi and get an API key.
  • Get response and store in a struct
  • Code adaptable to every need
    • Content generation
    • Summarization
    • Classification, categorization, and sentiment analysis
    • Data extraction
    • Translation
    • Many more!
  • Integrated with swiftUI

How To Use

You can build the project using Xcode. To get the Api key go to your account on https://beta.openai.com/account/api-keys and copy the key.

In this example i’ve used Davinci “gpt-3.5-turbo” model with chat completions, you can adapt the code, create your custom AI

Change the value with your OpenAi Api key

urlRequest.httpMethod = "POST"
        
urlRequest.addValue("Bearer YOUR-API-KEY", forHTTPHeaderField: "Authorization")
        
urlRequest.addValue("application/json", forHTTPHeaderField: "Content-Type")

You can change and add or remove all OpenAI parameters:

guard let url = URL(string: "https://api.openai.com/v1/chat/completions") else {
            fatalError("Missing URL")
        }
.        
.
.
let message = Messages(role: "user", content: prompt)
let requestData = RequestData(model: "gpt-3.5-turbo", messages: [message])
.
.
.
struct RequestData: Codable {
    var model: String
    var messages: [Messages]
    
}
struct Messages: Codable {
    let role: String
    let content: String
}
extension TranslationResponse {
    struct TextCompletionChoice: Decodable{
        var index: Int
        var message: Messages
        var finish_reason: String
    }
}

Please read the official OpenAi API Documentation: https://beta.openai.com/docs if you need more info !

In the view you can easy call the Api using:

  .onAppear {
          network.prompt = "Hello my friend how are you?"

          network.getResponse()
        }

To get response text in the view, you can use:

Text("Assistant: "+network.response.resultText)

The Struct must be declared with the same parameters of the API Response

If you change the Request parameters, the response may change, so you need to edit the struct!

struct TranslationResponse: Decodable {
    var id: String
    var object: String
    var created: Int
    var choices: [TextCompletionChoice]
    
    var resultText: String {
        choices.map(\.message.content).joined(separator: "\n")
    }
}

extension TranslationResponse {
    struct TextCompletionChoice: Decodable{
        var index: Int
        var message: Messages
        var finish_reason: String
    }
}

Note

You can found more information on: https://platform.openai.com/docs/api-reference/chat/create

Credits

THIS IS NOT THE OFFICIAL API CODE

Russo Giovanni M.

OpenAI

For educational purposes

License

MIT

GitHub

View Github