Agent Client

What ChatClient did for completion endpoints, AgentClient does for agent CLIs.

Agent Client provides a unified Java API for autonomous CLI agents — Claude Code, Codex, Gemini, Amazon Q, and Amp — with Spring Boot auto-configuration support.

📖 Documentation | Getting Started | Reference | Tutorial

Quick Start

Add the dependency for your provider:

<dependency>
    <groupId>org.springaicommunity.agents</groupId>
    <artifactId>agent-claude</artifactId>
    <version>0.15.0</version>
</dependency>

Build a model, create a client, run a goal — no Spring Boot required:

ClaudeAgentModel model = ClaudeAgentModel.builder()
    .defaultOptions(ClaudeAgentOptions.builder()
        .model("claude-sonnet-4-5")
        .yolo(true)
        .build())
    .build();

AgentClient client = AgentClient.create(model);
AgentClientResponse response = client.run("Create hello.txt with 'Hello from Agent Client!'");

With Spring Boot

Use a starter for auto-configuration:

<dependency>
    <groupId>org.springaicommunity.agents</groupId>
    <artifactId>agent-starter-claude</artifactId>
    <version>0.15.0</version>
</dependency>

@Component
public class MyAgent implements CommandLineRunner {
    private final AgentClient.Builder agentClientBuilder;

    public MyAgent(AgentClient.Builder agentClientBuilder) {
        this.agentClientBuilder = agentClientBuilder;
    }

    @Override
    public void run(String... args) {
        AgentClient client = agentClientBuilder.build();
        AgentClientResponse response = client.run("Fix the failing test");
    }
}

Supported Providers

Provider	Starter	Status
Claude Code	agent-starter-claude	Production
Codex	agent-starter-codex	Production
Gemini CLI	agent-starter-gemini	Production
Amazon Q	agent-starter-amazon-q	Beta
Amp	agent-starter-amp	Beta

Multi-Provider Support

Switch providers without changing code — use Maven profiles or swap the starter:

// This code works with ANY provider
AgentClient client = AgentClient.create(model);
AgentClientResponse response = client.run("Create hello.txt");

See Switching Providers for the Maven profile pattern.

Configuration

spring:
  ai:
    agents:
      mode: loose  # or strict
      claude-code:
        model: claude-sonnet-4-5
        timeout: PT5M
        yolo: true
      codex:
        model: gpt-5-codex
        full-auto: true
      gemini:
        model: gemini-2.5-flash
        yolo: true

See the Reference pages for all configuration options.

Architecture

agent-client/
├── agent-client-core/               # AgentClient fluent API
├── agent-models/                    # Provider adapters
│   ├── agent-model/                 # Core abstractions (AgentModel, AgentOptions)
│   ├── agent-tck/                   # Provider parity test kit
│   ├── agent-claude/                # Claude Code adapter
│   ├── agent-codex/                 # Codex adapter
│   ├── agent-gemini/                # Gemini CLI adapter
│   ├── agent-amazon-q/              # Amazon Q adapter
│   └── agent-amp/                   # Amp adapter
├── provider-sdks/                   # CLI client libraries
├── agent-starters/                  # Spring Boot auto-configuration
└── agents/                          # JBang-compatible agents

Two-Layer Design

• AgentClient — High-level fluent API (like ChatClient)
• AgentModel — Low-level provider interface (like ChatModel)

Provider selection happens at construction time. Everything after AgentClient.create(model) is portable.

Documentation

Type	Link
Getting Started	Quick start guide
Tutorial	Step-by-step lessons
Reference	Configuration options
Provider Reference	Claude · Codex · Gemini
Defaults Philosophy	LOOSE vs STRICT modes
Sessions	Multi-turn conversations

Building

./mvnw clean compile          # Compile
./mvnw clean test             # Unit tests
./mvnw clean verify -Pfailsafe  # Integration tests (requires CLIs + API keys)

License

Apache 2.0 — see LICENSE.