openai-kmp

Offers a library for interacting with OpenAI APIs, featuring a clean, testable interface. Utilizes Ktor and Kotlinx Serialization for seamless integration and API communication.

#service-sdk
#openai
#ktor-client
#ktor
#kotlin-serialization
#http-client
#client

Suggest an edit

Android JVMJVMKotlin/Native

GitHub stars1

AuthorsAlderFurtado

Open issues0

LicenseMIT License

Creation date8 months ago

Last activity8 months ago

Latest release1.0.1 (8 months ago)

GitHub repository Wiki page

🤖 openai-kmp

Kotlin Multiplatform library to interact with OpenAI APIs, designed for use in Android, iOS, and other Kotlin-supported platforms with a clean and testable interface.

✨ Features

✅ Kotlin Multiplatform (KMP) support
✅ Built with Ktor + Kotlinx Serialization

📦 Installation

Add to your KMP project:

dependencies {
    implementation("com.github.alderfurtado:openai-kmp:<latest-version>")
}

🚀 Usage

First, create an instance of the OpenAiInput class. It consists of:

openAiModel: The model type to use (see available models)
content: The prompt or command you want the AI to respond to
instruction: The way or tone the AI should respond in

val input = OpenAiInput(
    openAiModel = OpenAiModel.GPT_4O,
    content = "Who discovered Brazil?",
    instruction = "Respond like a history teacher"
)

Next, create an instance of the OpenAiKmp class. It provides the sendContent function, which requires:

openAiInput: the input object you just created 👆
token: your OpenAI API key get your API key here

val response = OpenAiKmp().sendContent(
    openAiInput = input,
    token = "your-openai-api-key"
)

Then, receive the responde which consinst of:

content: result of content
error: error if something goes wrong

result.content?.let { 
            // SUCCESS CODE FLOW
        }
        
result.error?.let { 
    // ERROR CODE FLOW
}

📄 Usage Example

val openAiInput = OpenAiInput(
    openAiModel = OpenAiModel.GPT_4O,
    content = "Who discovered Brazil",
    instruction = "Talks as History teacher"
)


val result = OpenAiKmp().sendContent(
    OpenAiInput(
        openAiModel = OpenAiModel.GPT_4O,
        content = "Quem descobriu o Brasil", instruction = "Talks as History teacher"
    ),
    "token" // put your openAi token
)


result.content?.let {
    // SUCCESS CODE FLOW
}

result.error?.let {
    // ERROR CODE FLOW
}

License

This project is licensed under the MIT License.