Android Guide

This page covers how to integrate Rezolve Commerce Brain in your Android app using the Point SDK. You’ll learn about the key classes, how to create and manage chat AI sessions, and how to handle real-time responses.

Prerequisites

• Android API level 26+
• The Rezolve AI integration is already configured in your project

Creating a Chat Session

After you’ve set up the Point SDK, you can create new chat sessions at any point by accessing BrainAI through ServiceManager:

val brainAI = ServiceManager.getInstance(context).brainAI
val chat = brainAI?.createNewChat()

Every chat session has a unique sessionID in URN format (e.g., urn:uuid:9a652678-4616-475d-af12-aca21cfbe06d).

If you need to retrieve an existing session:

val existingChat = brainAI?.getChatWithSessionID("urn:uuid:9a652678-4616-475d-af12-aca21cfbe06d")

And when you’re done:

brainAI?.closeChatWithSessionID("urn:uuid:9a652678-4616-475d-af12-aca21cfbe06d")

Sending Messages & Receiving Streaming Responses

With a Chat object, you can send messages to Rezolve Commerce Brain and receive chunks of responses as they’re generated. The response stream is represented by StreamingResponseDto objects.

Kotlin

A coroutine-based approach is recommended for asynchronous operations, preventing blocking the main thread.

// Use an appropriate coroutine scope (e.g., viewModelScope or a custom scope)
launch(Dispatchers.IO) {
    try {
        val responseStream = chat?.sendMessage("Tell me about the latest deals")
        responseStream?.forEach { chunk ->
            // Each chunk in the partial response
            println("Chunk received: ${chunk.response}")
        }
    } catch (e: Exception) {
        // Handle network or streaming errors
        println("Error: ${e.localizedMessage}")
    }
}

Java

If you’re using Java, you can handle the response stream using either Java 8+ streams or a callback-based approach:

try {
    // Potentially run on a background thread
    Stream<StreamingResponseDto> streamOfResponses =
        (Stream<StreamingResponseDto>) chat.sendMessage("Tell me about the latest deals", new CustomContinuation<Stream<StreamingResponseDto>>());

    streamOfResponses.forEach(responseChunk -> {
        System.out.println("Chunk received: " + responseChunk.getResponse());
    });

} catch (Exception e) {
    // Handle errors
    e.printStackTrace();
}

If you prefer callbacks, you can define a custom interface that’s invoked each time a new chunk of data arrives, similar to the closure approach on iOS.

Best Practices

• Session Persistence: Keep the same Chat session for a single conversation flow to maintain context.
• Error Handling: Network or streaming errors can occur.
• User Experience: Consider updating your UI as chunks arrive for a more dynamic and responsive experience.