spring ai

It provides a unified API across AI providers including OpenAI, Anthropic, Azure OpenAI, Google Vertex AI, Ollama, and Hugging Face, so application code remains provider-agnostic and switching models requires only configuration changes rather than code rewrites.

The framework covers the full stack of AI application patterns: chat completions with streaming support, function calling and tool use, structured output generation with schema validation, text and multimodal embedding models, image generation, transcription, and retrieval-augmented generation (RAG) with vector store integrations for Chroma, Pinecone, Weaviate, Qdrant, and others.

Spring AI's AI Advisors pattern provides a middleware layer for adding RAG, conversation memory, and logging to chat interactions without modifying the core application logic.

Spring AI is maintained by the Spring team at VMware (now Broadcom) and receives updates aligned with the Spring Boot release cycle.

It targets the large community of Java enterprise developers who want to add AI capabilities to existing Spring Boot applications microservices, REST APIs, batch processing systems without adopting a Python-centric AI stack.

The Spring ecosystem's testing, monitoring (Actuator), and observability integrations apply to AI components, making it straightforward to add AI features to production Spring applications with the operational practices those teams already use.