Z.ai Open Platform Java SDK

中文文档 | English

Not yet released.

The official Java SDK for Z.ai platforms, providing a unified interface to access powerful AI capabilities including chat completion, embeddings, image generation, audio processing, and more.

✨ Features

• 🚀 Type-safe API: All interfaces are fully type-encapsulated, no need to consult API documentation
• 🔧 Easy Integration: Simple and intuitive API design for quick integration
• ⚡ High Performance: Built with modern Java libraries for optimal performance
• 🛡️ Secure: Built-in authentication and token management
• 📦 Lightweight: Minimal dependencies for easy project integration

📦 Installation

Requirements

• Java 1.8 or higher
• Maven or Gradle
• Not supported on Android platform

Maven

Add the following dependency to your pom.xml:

<dependency>
    <groupId>ai.z.openapi</groupId>
    <artifactId>zai-sdk</artifactId>
    <version>0.0.1</version>
</dependency>

Gradle

Add the following dependency to your build.gradle (for Groovy DSL):

dependencies {
    implementation 'ai.z:zai-sdk:0.0.1'
}

Or build.gradle.kts (for Kotlin DSL):

dependencies {
    implementation("ai.z:zai-sdk:0.0.1")
}

📋 Dependencies

This SDK uses the following core dependencies:

Library	Version
OkHttp	4.12.0
Java JWT	4.4.0
Jackson	2.17.2
Retrofit2	2.11.0
RxJava	3.1.8
SLF4J	2.0.16

🚀 Quick Start

ZHIPU AI API PATH https://open.bigmodel.cn/api/paas/v4/

Z.ai API PATH https://api.z.ai/api/paas/v4/

Basic Usage

1. Create a ZaiClient with your API credentials
2. Access services through the client
3. Call API methods with typed parameters

import ai.z.openapi.ZaiClient;
import ai.z.openapi.service.model.*;
import ai.z.openapi.core.Constants;

// Create client with API key, recommend export the ENV api-key
// export ZAI_API_KEY=your.api.key

// For Z.ai platform https://api.z.ai/api/paas/v4/
ZaiClient client = ZaiClient.builder().build();

// For ZHIPU AI platform https://open.bigmodel.cn/api/paas/v4/
ZaiClient zhipuClient = ZaiClient.builder().ofZHIPU().build();


// Or set the api-key by code
ZaiClient client = ZaiClient.builder()
        .apiKey("your.api.key.your.api.secret")
        .build();

// For ZHIPU AI platform https://open.bigmodel.cn/api/paas/v4/
ZaiClient zhipuClient = ZaiClient.ofZHIPU("your.api.key").build();

Client Configuration

The SDK provides a flexible builder pattern for customizing your client:

ZaiClient client = ZaiClient.builder()
    .apiKey("your.api.key.your.api.secret")
    .baseUrl("https://api.z.ai/api/paas/v4/")
    .enableTokenCache()
    .tokenExpire(3600000) // 1 hour
    .connectionPool(10, 5, TimeUnit.MINUTES)
    .build();

💡 Examples

Chat Completion

import ai.z.openapi.ZaiClient;
import ai.z.openapi.service.model.*;
import ai.z.openapi.core.Constants;
import java.util.Arrays;

// Create client
ZaiClient client = ZaiClient.builder()
    .apiKey("your.api.key.your.api.secret")
    .build();

// Create chat request
ChatCompletionCreateParams request = ChatCompletionCreateParams.builder()
    .model(Constants.ModelChatGLM4)
    .messages(Arrays.asList(
        ChatMessage.builder()
            .role(ChatMessageRole.USER.value())
            .content("Hello, how are you?")
            .build()
    ))
    .stream(false)
    .temperature(0.7f)
    .maxTokens(1024)
    .build();

// Execute request
ChatCompletionResponse response = client.chat().createChatCompletion(request);

if (response.isSuccess()) {
    String content = response.getData().getChoices().get(0).getMessage().getContent();
    System.out.println("Response: " + content);
} else {
    System.err.println("Error: " + response.getMsg());
}

Streaming Chat

// Create streaming request
ChatCompletionCreateParams streamRequest = ChatCompletionCreateParams.builder()
    .model(Constants.ModelChatGLM4)
    .messages(Arrays.asList(
        ChatMessage.builder()
            .role(ChatMessageRole.USER.value())
            .content("Tell me a story")
            .build()
    ))
    .stream(true) // Enable streaming
    .build();

// Execute streaming request
ChatCompletionResponse response = client.chat().createChatCompletion(streamRequest);

if (response.isSuccess() && response.getFlowable() != null) {
    response.getFlowable().subscribe(
        data -> {
            // Handle streaming chunk
            if (data.getChoices() != null && !data.getChoices().isEmpty()) {
                String content = data.getChoices().get(0).getDelta().getContent();
                if (content != null) {
                    System.out.print(content);
                }
            }
        },
        error -> System.err.println("\nStream error: " + error.getMessage()),
        () -> System.out.println("\nStream completed")
    );
}

Function Calling

// Define function
ChatTool weatherTool = ChatTool.builder()
    .type(ChatToolType.FUNCTION.value())
    .function(ChatFunction.builder()
        .name("get_weather")
        .description("Get current weather for a location")
        .parameters(ChatFunctionParameters.builder()
            .type("object")
            .properties(Map.of(
                "location", Map.of(
                    "type", "string",
                    "description", "City name"
                )
            ))
            .required(Arrays.asList("location"))
            .build())
        .build())
    .build();

// Create request with function
ChatCompletionCreateParams request = ChatCompletionCreateParams.builder()
    .model(Constants.ModelChatGLM4)
    .messages(Arrays.asList(
        ChatMessage.builder()
            .role(ChatMessageRole.USER.value())
            .content("What's the weather like in Beijing?")
            .build()
    ))
    .tools(Arrays.asList(weatherTool))
    .toolChoice("auto")
    .build();

ChatCompletionResponse response = client.chat().createChatCompletion(request);

Embeddings

import ai.z.openapi.service.embedding.*;

// Create embedding request
EmbeddingCreateParams request = EmbeddingCreateParams.builder()
    .model(Constants.ModelEmbedding3)
    .input(Arrays.asList("Hello world", "How are you?"))
    .build();

// Execute request
EmbeddingResponse response = client.embeddings().create(request);

if (response.isSuccess()) {
    response.getData().getData().forEach(embedding -> {
        System.out.println("Embedding: " + embedding.getEmbedding());
    });
}

Image Generation

import ai.z.openapi.service.image.*;

// Create image generation request
CreateImageRequest request = CreateImageRequest.builder()
    .model(Constants.ModelCogView3Plus)
    .prompt("A beautiful sunset over mountains")
    .size("1024x1024")
    .quality("standard")
    .n(1)
    .build();

// Execute request
ImageResponse response = client.images().generate(request);

if (response.isSuccess()) {
    response.getData().getData().forEach(image -> {
        System.out.println("Image URL: " + image.getUrl());
    });
}

Spring Boot Integration

@RestController
public class AIController {
    
    private final ZaiClient zaiClient;
    
    public AIController() {
        this.zaiClient = ZaiClient.builder()
            .apiKey("your.api.key.your.api.secret")
            .enableTokenCache()
            .build();
    }
    
    @PostMapping("/chat")
    public ResponseEntity<String> chat(@RequestBody ChatRequest request) {
        ChatCompletionCreateParams params = ChatCompletionCreateParams.builder()
            .model(Constants.ModelChatGLM4)
            .messages(Arrays.asList(
                ChatMessage.builder()
                    .role(ChatMessageRole.USER.value())
                    .content(request.getMessage())
                    .build()
            ))
            .build();
            
        ChatCompletionResponse response = zaiClient.chat().createChatCompletion(params);
        
        if (response.isSuccess()) {
            String content = response.getData().getChoices().get(0).getMessage().getContent();
            return ResponseEntity.ok(content);
        } else {
            return ResponseEntity.badRequest().body(response.getMsg());
        }
    }
}

🔧 Available Services

The ZaiClient provides access to comprehensive AI services:

Service	Description	Key Features
Chat	Text generation and conversation	Streaming, function calling, async support
Embeddings	Text embeddings generation	Multiple embedding models
Images	Image generation and processing	CogView models, various sizes
Audio	Speech synthesis and recognition	Text-to-speech, speech-to-text
Files	File management and processing	Upload, download, batch processing
Assistants	AI assistant management	Create, configure, and manage assistants
Agents	Agent-based completions	Specialized agent interactions
Knowledge	Knowledge base operations	Document indexing and retrieval
Batch	Batch processing	Bulk operations
Web Search	Web search integration	Real-time web information
Videos	Video processing	Video analysis and generation

🎯 Supported Models

Text Generation

• glm-4-plus - Enhanced GLM-4 with improved capabilities
• glm-4 - Standard GLM-4 model
• glm-4-air - Lightweight version for speed
• glm-4-flash - Ultra-fast response model
• glm-4-0520 - GLM-4 model version 0520
• glm-4-airx - Extended Air model with additional features
• glm-4-long - Optimized for long-context conversations
• glm-4-voice - Specialized for voice interactions
• glm-4.1v-thinking-flash - Visual reasoning model with thinking capabilities
• glm-z1-air - Optimized for mathematical and logical reasoning
• glm-z1-airx - Fastest domestic inference model with 200 tokens/s
• glm-z1-flash - Completely free reasoning model service
• glm-4-air-250414 - Enhanced with reinforcement learning optimization
• glm-4-flash-250414 - Latest free language model
• glm-4-flashx - Enhanced Flash version with ultra-fast inference speed
• glm-4-9b - Open-source model with 9 billion parameters
• glm-4-assistant - AI assistant for various business scenarios
• glm-4-alltools - Agent model for complex task planning and execution
• chatglm3-6b - Open-source base model with 6 billion parameters
• codegeex-4 - Code generation and completion model

Audio Speech Recognition

• glm-asr - Context-aware audio transcription model

Real-time Interaction

• glm-realtime-air - Real-time video call model with cross-modal reasoning
• glm-realtime-flash - Fast real-time video call model

Vision Models

• glm-4v-plus - Enhanced vision model
• glm-4v - Standard vision model
• glm-4v-plus-0111 - Variable resolution video and image understanding
• glm-4v-flash - Free and powerful image understanding model

Image Generation

• cogview-3-plus - Enhanced image generation
• cogview-3 - Standard image generation
• cogview-3-flash - Free image generation model
• cogview-4-250304 - Advanced image generation with text capabilities
• cogview-4 - Advanced image generation for precise and personalized AI image expression

Video Generation

• cogvideox - Video generation from text or images
• cogvideox-flash - Free video generation model
• cogvideox-2 - New video generation model
• viduq1-text - High-performance video generation from text input
• viduq1-image - Video generation from first frame image and text description
• viduq1-start-end - Video generation from first and last frame images
• vidu2-image - Enhanced video generation from first frame image and text description
• vidu2-start-end - Enhanced video generation from first and last frame images
• vidu2-reference - Video generation with reference images of people, objects, etc.

Embeddings

• embedding-3 - Latest embedding model
• embedding-2 - Previous generation embedding

Specialized

• charglm-3 - Character interaction model
• cogtts - Text-to-speech model
• rerank - Text reordering and relevance scoring

📈 Release Notes

For detailed release notes and version history, please see Release-Note.md.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🤝 Contributing

We welcome contributions! Please feel free to submit a Pull Request.

📞 Support

For questions and support:

• Visit Z.ai Platform
• Visit ZHIPU AI Open Platform
• Check our Architecture Documentation