🧠 AI with Java & Spring Boot – Part 3: Building a Memory-Aware Chatbot with LangChain4j

Hey devs! 👋

Welcome back to the third part of our AI-with-Java journey. So far, we've:

  1. Created a text summarizer with OpenAI
  2. Streamed ChatGPT-style responses in real time

Today, we’ll go even further — and introduce LangChain4j, the Java adaptation of LangChain, to create a chatbot with memory and context.

Imagine asking your bot:

"Who’s the CEO of Tesla?"

"What year was he born?"

And the bot remembers who “he” refers to. That’s what we’re building today. 🧠

🧰 Tools We’ll Use

  • LangChain4j – Java library for building LLM apps
  • Spring Boot
  • OpenAI API
  • Session Memory

🔧 Step-by-Step: AI Chatbot with Memory (LangChain4j + Spring Boot)

1. Add Dependencies (Maven)

Add this to your pom.xml:

dev.langchain4j
    langchain4j
    0.25.0


    dev.langchain4j
    langchain4j-openai
    0.25.0

2. Configure OpenAI

In application.yml:

openai:
  api-key: YOUR_OPENAI_API_KEY
  model: gpt-3.5-turbo

3. Create the ChatMemoryService 

@Component
@Scope("session") // memory per user session
public class MemoryChatService {

    private final ChatLanguageModel model;
    private final ChatMemory memory;

    public MemoryChatService(@Value("${openai.api-key}") String apiKey) {
        this.model = OpenAiChatModel.builder()
            .apiKey(apiKey)
            .modelName("gpt-3.5-turbo")
            .temperature(0.7)
            .build();

        this.memory = MessageWindowChatMemory.withMaxMessages(10);
    }

    public String chat(String userInput) {
        ChatLanguageModelChain chain = new ChatLanguageModelChain(model, memory);
        return chain.execute(userInput);
    }
}

4. Expose It via REST Controller 

@RestController
@RequestMapping("/api/langchain")
@SessionAttributes("memoryChatService")
public class LangChainController {

    private final MemoryChatService memoryChatService;

    public LangChainController(MemoryChatService memoryChatService) {
        this.memoryChatService = memoryChatService;
    }

    @PostMapping("/chat")
    public ResponseEntity<String> chat(@RequestBody Map<String, String> request) {
        String prompt = request.get("prompt");
        String response = memoryChatService.chat(prompt);
        return ResponseEntity.ok(response);
    }
}

5. Try It Out 🚀

Make a POST request:

curl -X POST http://localhost:8080/api/langchain/chat \
-H "Content-Type: application/json" \
-d '{"prompt": "Who is the CEO of Google?"}'

Then ask:

curl -X POST http://localhost:8080/api/langchain/chat \
-H "Content-Type: application/json" \
-d '{"prompt": "Where was he born?"}'

It’ll remember!

🔚 Wrapping Up

Now you have an intelligent chatbot that remembers your conversation! 🎉

You can also:

  • Store memory to DB or Redis
  • Customize personas
  • Add tools like calculators or file readers (LangChain4j supports them too)

🔜 Coming in Part 4...

  • File-based Q&A
  • Upload PDFs and ask questions
  • Embedding + vector search (e.g., using ChromaDB or Pinecone with Java!)

If you're enjoying the series, make sure to:
❤️ React

🗨️ Comment

📌 Follow for Part 4!

Stay sharp!

— RF 👨‍💻