Introducing Spring AI: Building a Simple Chat Application with DeepSeek

Spring AI is a Java framework within Spring that simplifies AI integration by providing unified interfaces and portable APIs for various AI models.

Key features include support for major AI model providers (Anthropic, OpenAI, Microsoft, Amazon, Google, Ollama), multiple model types (chat, embedding, text‑to‑image, audio transcription, text‑to‑speech), cross‑provider API, vector‑database integration, observability, retrieval‑augmented generation (RAG), and Spring Boot starters.

In the engineering example, a Spring Boot project is created (via start.spring.io) with Maven, adding the spring-ai-openai-spring-boot-starter dependency (version 1.0.0‑M6) and configuring DeepSeek API credentials in application properties.

Configuration snippet:

spring.ai.openai.api-key=<insert deepseek api key here>
spring.ai.openai.base-url=https://api.deepseek.com
spring.ai.openai.chat.options.model=deepseek-chat

A REST controller injects OpenAiChatModel and defines two endpoints: /ai/call for synchronous chat using chatModel.call(message) , and /ai/stream for reactive streaming using chatModel.stream(prompt) .

Supporting classes such as Prompt , ChatResponse , and related metadata are shown, illustrating how messages, options, and model responses are modeled.

The article concludes that the simple DeepSeek chat application demonstrates Spring AI’s capability to accelerate AI engineering and that future releases will expand functionality.