Spring AI has added support for DeepSeek AI

Integrating Spring AI with DeepSeek R1 locally using Ollama

You’ve probably all heard about the DeepSeek R1 model, which is famed for its reasoning, performance, and cost-efficiency these days. In this article, we will discuss how Spring AI can easily support new models like the DeepSeek R1 because of its modular and extensible API, and at the end, we will change the Employee Assistance chatbot source code in the tutorial and use the DeepSeek R1 model instead of OpenAI or Llama.

Integrating Spring AI with DeepSeek R1 locally using Ollama

· How can we use DeepSeek?
· How does Spring AI support DeepSeek?
· Without introduction, let’s code!
∘ Migrating AI model provider from Llama 3.1 to DeepSeek R1
∘ Testing the result
· Final Thoughts

How can we use DeepSeek?

Similar to Llama, DeepSeek is also Open Source, so we have two options to use it using our chatbot application:

• Run DeepSeek locally using Ollama
• Online DeepSeek Chat

How does Spring AI support DeepSeek?

Fortunately, Spring AI supports both approaches, whether you want to run DeepSeek locally using Ollama or use the online DeepSeek Chat…