Chat Model Overview

Conversational AI using Google Vertex AI chat models (chat-bison, gemini). Implements LangChain4j ChatModel interface.

Class

public class VertexAiChatModel implements ChatModel {
    public ChatResponse chat(ChatRequest chatRequest);
    public static Builder builder();
}

Imports

import dev.langchain4j.model.vertexai.VertexAiChatModel;
import dev.langchain4j.data.message.UserMessage;
import dev.langchain4j.data.message.AiMessage;
import dev.langchain4j.model.chat.request.ChatRequest;
import dev.langchain4j.model.chat.response.ChatResponse;

Minimal Example

VertexAiChatModel model = VertexAiChatModel.builder()
    .endpoint("https://us-central1-aiplatform.googleapis.com/v1/")
    .project("your-project-id")
    .location("us-central1")
    .publisher("google")
    .modelName("chat-bison@001")
    .build();

ChatRequest request = ChatRequest.builder()
    .messages(UserMessage.from("What is LangChain4j?"))
    .build();

ChatResponse response = model.chat(request);
System.out.println(response.aiMessage().text());

Supported Models

• chat-bison@001 - PaLM 2 chat model
• chat-bison@002 - PaLM 2 chat model (updated)
• gemini-pro - Gemini Pro multimodal
• gemini-ultra - Gemini Ultra (most capable)

Configuration Parameters

Required

• endpoint - API endpoint URL
• project - Google Cloud Project ID
• location - GCP region
• publisher - Model publisher ("google")
• modelName - Model name/version

Optional

• temperature (Double) - Randomness 0.0-1.0 (default: varies by model)
• maxOutputTokens (Integer) - Max response length (default: 200)
• topK (Integer) - Top-K sampling
• topP (Double) - Nucleus sampling 0.0-1.0
• maxRetries (Integer) - Retry attempts (default: 2)
• credentials (GoogleCredentials) - Custom auth

Use Cases

• Chatbots and virtual assistants
• Multi-turn conversations with context
• Question answering with follow-ups
• Interactive tutorials and guides
• Customer support automation

See Also

• Examples - Detailed usage examples
• API Reference - Complete API documentation