Spring AI MCP Client Duplicate Tools Issue Demonstration

This repository demonstrates the issue with duplicate SSE connections and MCP tools creation in Spring AI when both WebFlux and HTTP Client implementations are available.

Issue Description

When both WebFlux and HTTP Client transport implementations are available, the following issues occur:

1. Duplicate SSE Connections Problem

   • McpClientAutoConfiguration loads both implementations
   • SseWebFluxTransportAutoConfiguration and SseHttpClientTransportAutoConfiguration create duplicate connections
   • Results in unnecessary resource consumption and potential performance issues

2. Duplicate MCP Tools Creation

   • Multiple identical MCP tools are created due to duplicate transport implementations
   • This affects system efficiency and resource utilization
   • May cause confusion in tool management and callbacks

How to Reproduce

This repository contains a minimal project that reproduces the issue:

Scenario 1: SYNC Mode (Application Fails to Start)

1. Open spring-ai-mcp-clinet/src/main/resources/application.yml
2. Set spring.ai.mcp.client.type to SYNC
3. Run the application
4. Observe the error: Multiple tools with the same name (get_addr_date)

Scenario 2: ASYNC Mode (Application Starts but with Duplicate Tools)

1. Open spring-ai-mcp-clinet/src/main/resources/application.yml
2. Set spring.ai.mcp.client.type to ASYNC
3. Run the application
4. Access the endpoint http://localhost:8080/getServerList
5. Observe in logs that duplicate tools with the same name are registered

Dependencies Used

<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-mcp-client-webflux-spring-boot-starter</artifactId>
</dependency>

Current Workarounds

Temporary workarounds (until PR #2493 is merged):

1. Exclude one of the auto-configurations:

   spring.autoconfigure.exclude=org.springframework.ai.autoconfigure.mcp.client.SseHttpClientTransportAutoConfiguration

2. Or in application class:

   @SpringBootApplication(exclude = {
       org.springframework.ai.autoconfigure.mcp.client.SseHttpClientTransportAutoConfiguration.class
   })

Proposed Solution

The proposed solution in PR #2493 adds explicit transport mode configuration:

spring.ai.mcp.client.sse.transport-mode=WEBFLUX  # or HTTP_CLIENT

This ensures only one transport implementation is active at a time.