OpenAI’s ChatGPT is an synthetic intelligence (AI) chatbot that may generate human-like responses primarily based on textual content enter utilizing the Generative Pre-trained Transformer (GPT) massive language mannequin. OpenAI launched ChatGPT in December 2022, and in two months, it garnered over 100 million month-to-month lively customers, making it the fastest-growing client utility in historical past. Many individuals use ChatGPT for interactive training, mock interviews, language translation, automated customer support and extra.

ChatGPT is obtainable as an utility programming interface (API) service, which suggests you’ll be able to combine it into your Android utility. With ChatGPT, you’ll be able to create an Android utility to right your typing grammar, as an illustration.

On this tutorial, you’ll discover ways to construct Chatty Bot, a ChatGPT-based chatbot Android utility. Within the utility, you’ll be able to select a persona to your chatbot, akin to a journey information.

Through the course of, you’ll study to:

• Design chat messages.
• Create typing animation.
• Combine the ChatGPT OpenAI API service.
• Perceive the parameters of the API name.

Be aware: This text assumes you have got earlier expertise with Kotlin and Jetpack Compose in Android. To know Kotlin, leap into this Introduction to Kotlin for Android tutorial and the guide Kotlin Apprentice. For studying Android improvement, take a look at the guide Android Apprentice. Lastly, to grasp Jetpack Compose, you can begin with Jetpack Compose by Tutorials.

And with that, it’s time to speak it up!

Getting Began

Obtain the starter mission by clicking the Obtain Supplies button on the prime or backside of the tutorial. Open Android Studio Electrical Eel or import and open the starter mission later.

Browse the mission’s recordsdata to familiarize your self with the codebase. There are two most important folders underneath the com.kodeco.android.chattybot folder: the ui folder and the mannequin folder.

Six recordsdata are underneath the ui folder. The Settings.kt file is the Settings display, a type for getting into the OpenAI API key. The TabScreen.kt file is a container for the chatting display, and the PersonaScreen.kt file permits customers to decide on the ChatGPT persona, akin to a journey information. You’ll create the consumer interface (UI) for the chat message packing containers within the AIMessage.kt, UserMessage.kt and ChattingScreen.kt recordsdata.

The mannequin listing accommodates a number of recordsdata. The primary file, Persona.kt, shops the persona textual content for PersonaScreen.kt. Constants.kt shops mission constants. ChatModel.kt accommodates the information fashions for exchanging messages between the OpenAI API server with the Retrofit library. OpenAIService.kt is the interface for calling the OpenAI API server, and ChatRetriever.kt receives messages for ChattingScreen.kt and connects to the server utilizing the Retrofit library.

Opening the Starter Mission

Construct and run the starter mission. You’ll see the next display:

The app shows three tabs: Chat, Persona and Settings. The Chat tab is chosen by default. This tab display has a textual content discipline with a button on the backside , like a chat window. On this tab, you’ll show message packing containers from the consumer and reply packing containers from the API service.

Click on the Persona tab. You’ll see 4 persona decisions. Personas are methods of characterizing the chatbot’s responses: A journey information presents the responses in a different way than, say, a motivational coach. For those who click on one in every of them, the field border turns cyan to point that that is your chatbot persona:

Do you see the “Journey Information in Tokyo” persona? This appears like an journey ready to occur! :-]

The final tab is the Settings tab, the place you’ll be able to enter your OpenAI key:

Making a Chat Messages Checklist

Add some messages to the chat display to simulate a chatbot. Right here, you’ll create a UI resembling fashionable chat purposes like WhatsApp, Telegram or Fb Messenger. A chatting window UI sometimes consists of an inventory of messages, a textual content discipline for typing messages and a button to ship messages. The textual content discipline and button have already been created. You’ll discover the container for the listing of chat messages within the ChattingScreen.kt file. Search for the road // TODO: Create dialog packing containers, which is inside an empty Field.

For every consumer message, create a Field that doesn’t take up the whole horizontal house and is aligned to the best facet of the container. Then, add a Textual content contained in the Field. The Textual content will show the message the consumer varieties to ChatGPT. To point that this message comes from a consumer, add an icon on the best facet of the message.

Add the next code beneath the road // TODO: Create dialog packing containers:

        
val content material = "Whats up, AI!"
Row(
  modifier = Modifier.fillMaxWidth(),
  horizontalArrangement = Association.Finish,
  verticalAlignment = Alignment.Backside
) {
    Field(
      modifier = Modifier
        .padding(16.dp)
        .fillMaxWidth(0.8f)
        .heightIn(120.dp)
        .background(Shade.LightGray, RoundedCornerShape(16.dp))
    ) {
      Textual content(
        textual content = content material,
        modifier = Modifier
          .align(Alignment.Middle)
          .padding(8.dp)
      )
    }
    Icon(
      imageVector = Icons.Default.Individual,
      contentDescription = "Consumer Icon"
    )
  }

Construct and run the mission, and it is best to see the next display:

It appears good! The chat message field is appropriately proper aligned. The Row takes up the total width, and the Field takes up solely 80 % of the width, as indicated by the .fillMaxWidth(0.8f) modifier. The one concern is that the chat message field is on the prime, relatively than the underside, near the typing TextField. For those who’ve used a chat utility, you’re conversant in message packing containers popping up from the underside, however you’ll cowl that later.

Displaying the Chat Response

For now, you’ll wish to refactor the code. Delete the code, and transfer it to the UserMessage.kt file. Open the UserMessage.kt file and add the next code beneath the road // TODO: Create the consumer dialog field:

  Row(
    modifier = Modifier.fillMaxWidth(),
    horizontalArrangement = Association.Finish,
    verticalAlignment = Alignment.Backside
  ) {
    Field(
      modifier = Modifier
        .padding(16.dp)
        .fillMaxWidth(0.8f)
        .heightIn(120.dp)
        .background(Shade.LightGray, RoundedCornerShape(16.dp))
    ) {
      Textual content(
        textual content = content material,
        modifier = Modifier
          .align(Alignment.Middle)
          .padding(8.dp)
      )
    }
    Icon(
      imageVector = Icons.Default.Individual,
      contentDescription = "Consumer Icon"
    )
  }

Then, within the ChattingScreen.kt file, beneath the road // TODO: Create dialog packing containers, add the next code:

UserMessage("Whats up, AI!")

For those who construct and run the mission once more, you see the identical display. Now, you wish to add a chat message from ChatGPT. This chat message field ought to be aligned to the left of the chatting container. Make sure the field colour is completely different so customers can differentiate the messages extra simply. Don’t overlook so as to add an Icon on the left of the message.

Open the AIMessage.kt file and add the next code beneath the road // TODO: Create the AI dialog field:

  Row(
    modifier = Modifier.fillMaxWidth(),
    horizontalArrangement = Association.Begin,
    verticalAlignment = Alignment.Backside
  ) {
    Icon(
      imageVector = Icons.Default.Face,
      contentDescription = "AI Icon"
    )
    Field(
      modifier = Modifier
        .padding(16.dp)
        .fillMaxWidth(0.8f)
        .heightIn(120.dp)
        .background(Shade.DarkGray, RoundedCornerShape(16.dp))
    ) {
      Textual content(
        textual content = content material,
        modifier = Modifier
          .align(Alignment.Middle)
          .padding(8.dp),
        colour = Shade.White
      )
    }
  }