Spring AI 2.0 GA: Native Java AI Development with Spring Boot 4 Integration

On June 12, 2026 the Spring team announced Spring AI 2.0.0 as GA, positioning it for AI Engineering scenarios and aligning it deeply with Spring Boot 4.0/4.1 and Spring Framework 7.0 to give Java developers a native, production‑grade AI application path without needing Python.

Core Highlight 1 – Deep Spring Boot 4 Integration

Adopts Jackson 3 for enhanced JSON serialization.

Annotates the codebase with JSpecify null‑safety annotations to reduce NPEs and improve Kotlin experience.

Reworks the Options system: unified defaults, builder‑style construction, immutable objects, and clearer property keys.

Core Highlight 2 – Agent and Tool‑Calling Upgrade

ToolCallingAdvisor : automatically registered by ChatClient to handle tool‑call round‑trips uniformly.

ToolSearchToolCallingAdvisor : supports progressive tool disclosure, scaling to hundreds of tools.

StructuredOutputValidationAdvisor : auto‑self‑corrects when structured‑output validation fails. ChatClient becomes the primary user‑facing API, while ChatModel serves as a lower‑level building block.

Core Highlight 3 – Native MCP Support

Spring maintains the official MCP Java SDK and bundles MCP SDK 2.0.0 compliant with the 2025‑11‑25 MCP specification.

Annotation‑driven exposure of services via @McpTool, @McpResource, and @McpPrompt.

Uses Streamable HTTP as the default transport (replacing deprecated SSE).

Integrates with enterprise observability and security stacks such as Micrometer, OpenTelemetry, and OAuth 2.0.

Core Highlight 4 – Model Provider Consolidation & Community Collaboration

OpenAI, Anthropic, and Google GenAI each collapse multiple implementations into a single SDK.

Support added for Ollama, OCI Generative AI, Azure Cosmos DB, and other community‑maintained providers.

A CONTRIBUTING.md guide is provided for human and AI Agent contributions.

Quick‑Start

Include the Spring AI BOM in a Maven dependencyManagement section or use the Gradle platform syntax. The required runtime is Spring Boot 4.0.x or 4.1.x, and projects can also be generated via start.spring.io.

Upgrade Considerations

Moving from Spring AI 1.1.x to 2.0.0 introduces breaking changes; users should consult the Upgrade Notes, reference documentation, and release notes. Major changes include the Options API refactor, the new Advisor‑based tool‑calling architecture, migration to Jackson 3, and the integration of the MCP module.

Community Extensions

spring-ai-session : event‑sourced conversation memory with tool‑call context and LLM summary compression.

spring-ai-agent-utils : bundles Agent skills and utilities (file, shell, web) wrapped as Advisors.

spring-ai-community/mcp-security : provides OAuth 2.0 and API‑Key security for MCP scenarios.