Spring AI Introduction - Build AI Applications with Java & Spring Boot

Learn Spring AI from scratch. Understand its architecture, core concepts, supported AI models, and build your first AI-powered Spring Boot application step by step.

Spring AI is an official Spring ecosystem project that makes it easy for Java developers to build AI-powered applications using familiar Spring programming models.

Instead of learning different SDKs for every AI provider, Spring AI provides a consistent abstraction that works across multiple AI platforms.

Think of it like Spring Data JPA.

Instead of writing database-specific code for MySQL, PostgreSQL, Oracle, or SQL Server, Spring Data provides one common API.

Similarly,

Instead of writing different code for OpenAI, Azure OpenAI, Anthropic Claude, Google Gemini, Amazon Bedrock, or Ollama, Spring AI provides one unified API.

Why Spring AI?

Before Spring AI, developers had to:

Learn each AI provider SDK
Handle REST API calls manually
Manage authentication
Parse JSON responses
Handle retries
Handle token limits
Build prompt templates manually

Spring AI simplifies all of this.

Traditional AI Integration

flowchart LR

A[Spring Boot Application]

B[OpenAI SDK]

C[Gemini SDK]

D[Claude SDK]

E[Bedrock SDK]

F[REST APIs]

A --> B
A --> C
A --> D
A --> E

B --> F
C --> F
D --> F
E --> F

Problems:

Vendor lock-in
Different APIs
Different request formats
More maintenance

Spring AI Approach

flowchart LR

App[Spring Boot Application]

AI[Spring AI]

OpenAI[OpenAI]

Gemini[Google Gemini]

Claude[Anthropic Claude]

Bedrock[Amazon Bedrock]

Ollama[Ollama Local]

App --> AI

AI --> OpenAI
AI --> Gemini
AI --> Claude
AI --> Bedrock
AI --> Ollama

One API.

Multiple AI providers.

Where Can Spring AI Be Used?

Spring AI can build:

ChatGPT-like applications
AI chatbots
Customer support bots
Knowledge assistants
Document Q&A
PDF search
Resume analyzers
Email generators
Code generators
SQL generators
AI agents
Multi-agent systems
Recommendation engines
Enterprise copilots

Real Enterprise Example

Imagine a banking application.

Customer asks:

"Why was my credit card payment declined?"

Without AI:

Customer support manually checks multiple systems.

With Spring AI:

Customer
     │
     ▼
Spring Boot
     │
     ▼
Spring AI
     │
     ▼
LLM
     │
     ▼
Knowledge Base
     │
     ▼
Answer

The AI can search documentation, transaction history (with proper authorization), and FAQs to generate a human-readable response.

Spring AI Architecture

flowchart TD

User

Controller

Prompt

ChatClient

AIModel

OpenAI

Response

User --> Controller

Controller --> Prompt

Prompt --> ChatClient

ChatClient --> AIModel

AIModel --> OpenAI

OpenAI --> Response

Spring AI Modules

Spring AI consists of multiple modules.

Chat Models

Used for conversations.

Examples:

ChatGPT
Claude
Gemini
Llama

Embedding Models

Convert text into vectors.

Used for:

Semantic search
Similarity search
Recommendation systems
RAG

Image Models

Generate images from prompts.

Example:

Generate a futuristic city skyline.

Audio Models

Speech-to-text

Text-to-speech

Voice assistants

Moderation Models

Detect:

Toxic content
Hate speech
Violence
Unsafe content

Vector Stores

Store embeddings.

Supported stores include:

PostgreSQL + PGVector
ChromaDB
Milvus
Redis
Elasticsearch
Pinecone
Weaviate
Azure AI Search

Core Components

Prompt

A prompt is the instruction sent to the AI.

Example:

Explain Java Streams with examples.

ChatClient

The main API developers use.

String response =
chatClient.prompt("Hello")
.call()
.content();

ChatModel

Responsible for communicating with the LLM.

Application

↓

ChatClient

↓

ChatModel

↓

AI Provider

EmbeddingModel

Converts text into vectors.

Text

↓

Embedding Model

↓

Vector

↓

Database

Vector Store

Stores vectors.

Used for semantic search.

Supported AI Providers

Spring AI currently supports:

OpenAI
Azure OpenAI
Google Gemini
Anthropic Claude
Ollama
Amazon Bedrock
Mistral AI
Hugging Face
Cohere
DeepSeek (via compatible APIs)

This abstraction allows changing providers with minimal application code changes.

Spring AI Project Structure

spring-ai-demo

src
 ├── controller
 │      ChatController
 │
 ├── service
 │      AIService
 │
 ├── config
 │      AIConfig
 │
 └── SpringAiApplication

Create Spring Boot Project

Dependencies:

Spring Web
Spring AI
OpenAI Starter (or another provider)
Lombok (optional)

Maven Dependency

<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-starter-model-openai</artifactId>
</dependency>

application.yml

spring:

  ai:

    openai:

      api-key: ${OPENAI_API_KEY}

      chat:

        options:

          model: gpt-4.1-mini

Using environment variables keeps secrets out of source control.

Simple AI Service

@Service
public class AIService {

    private final ChatClient chatClient;

    public AIService(ChatClient.Builder builder) {
        this.chatClient = builder.build();
    }

    public String ask(String question) {

        return chatClient
                .prompt(question)
                .call()
                .content();
    }
}

REST Controller

@RestController
@RequestMapping("/chat")
public class ChatController {

    private final AIService aiService;

    public ChatController(AIService aiService) {
        this.aiService = aiService;
    }

    @GetMapping
    public String chat(@RequestParam String prompt) {
        return aiService.ask(prompt);
    }
}

Test the API

Request

GET /chat?prompt=Explain Spring Boot

Response

Spring Boot is a Java framework that simplifies the development of production-ready Spring applications by providing auto-configuration, embedded servers, starter dependencies, and production features.

Request Flow

sequenceDiagram

User->>Controller: GET /chat
Controller->>AIService: ask(prompt)
AIService->>ChatClient: prompt()
ChatClient->>OpenAI: API Request
OpenAI-->>ChatClient: AI Response
ChatClient-->>AIService: Response
AIService-->>Controller: String
Controller-->>User: JSON/Text

Why Java Developers Love Spring AI

Familiar Spring Boot programming model
Dependency Injection
Auto Configuration
Easy provider switching
Enterprise-ready architecture
Integration with Spring Security
Spring Data compatibility
Production monitoring with Micrometer
Easy testing and mocking
Consistent API across AI providers

Best Practices

✅ Keep API keys in environment variables.

✅ Use prompt templates instead of hardcoding prompts.

✅ Validate user input before sending it to AI.

✅ Log requests responsibly without exposing sensitive data.

✅ Monitor token usage and latency.

✅ Implement retries and timeouts.

✅ Cache responses where appropriate.

Common Use Cases

Use Case	Example
Chatbot	Customer support
RAG	Company documentation search
Code Generation	Generate Java code
SQL Generator	Convert English to SQL
Email Assistant	Draft business emails
PDF Assistant	Ask questions about PDFs
AI Agent	Automate multi-step workflows
Enterprise Search	Search across internal documents

What You'll Learn in This Spring AI Series

Spring AI Introduction
Building Your First AI Chat Application
Prompt Engineering
AI Chat Memory
Structured Output
Function Calling
Retrieval-Augmented Generation (RAG)
Vector Databases
AI Agents
Production Best Practices

Summary

In this article, you learned:

What Spring AI is
Why it simplifies AI development
Spring AI architecture
Core components
Supported AI providers
How to build your first AI-enabled Spring Boot application
Best practices for enterprise AI development

Spring AI brings the familiar Spring programming model to Generative AI, allowing Java developers to build intelligent applications without learning provider-specific SDKs.

In the next article, we'll build our first AI chatbot using Spring Boot and Spring AI.

Loading likes...

Comments

Share a question, correction, or practical insight about this article.

Loading approved comments...