home / mcp / splitmind mcp server

SplitMind MCP Server

A2AMCP is a Agent2Agent MCP communication Server taking the concept from Google's Agent2Agent Protocol (A2A)

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "webdevtodayjason-a2amcp": {
      "command": "docker",
      "args": [
        "exec",
        "-i",
        "splitmind-mcp-server",
        "python",
        "/app/mcp-server-redis.py"
      ],
      "env": {
        "REDIS_URL": "redis://localhost:6379"
      }
    }
  }
}

You deploy and run a dedicated MCP server to enable real-time collaboration among AI agents working on the same codebase. This server uses a Redis-backed MCP channel to coordinate access, share interfaces, and track tasks, helping your agents avoid conflicts and work as a cohesive team.

How to use

You connect your MCP-enabled agents to the server and start coordinating across projects. The server exposes a standard, multi-agent coordination workflow where agents can query each other, broadcast updates, lock files to prevent conflicts, share API interfaces, and track task progress. Use this flow to orchestrate parallel development tasks with clear ownership, visibility, and conflict prevention.

How to install

Prerequisites: ensure Docker and Docker Compose are installed on your machine. You also need Python installed if you plan to use Python tooling for verification and SDKs.

# 1) Install Docker and Docker Compose if not already installed
# 2) Clone the MCP server project and navigate into it
# 3) Bring up the services in detached mode

docker-compose up -d

# 4) Verify the MCP server container is running
docker ps | grep splitmind

# 5) Run a connectivity check using the provided verification script
python verify_mcp.py

Configure your agents to connect to the MCP server. You can use either the CLI integration or the desktop configuration shown in the examples.

# Claude Code CLI example
claude mcp add splitmind-a2amcp \
  -e REDIS_URL=redis://localhost:6379 \
  -- docker exec -i splitmind-mcp-server python /app/mcp-server-redis.py

# Claude Desktop JSON configuration example (macOS)
{
  "mcpServers": {
    "splitmind-a2amcp": {
      "command": "docker",
      "args": ["exec", "-i", "splitmind-mcp-server", "python", "/app/mcp-server-redis.py"],
      "env": {
        "REDIS_URL": "redis://redis:6379"
      }
    }
  }
}

Additional setup and notes

Status and capabilities are designed around a modern MCP SDK. The server currently supports real-time agent communication, file conflict prevention, shared context management, task transparency, multi-project support, and full MCP integration with the 17 tools implemented.

Health and verification are straightforward: you can check the health endpoint or run the provided verification script to ensure connectivity and proper operation.

Troubleshooting

If your agents cannot see the MCP server tools or face connectivity issues, restart the desktop client or verify that the server is running and healthy. Use the health check and verification steps to diagnose issues.

Restart Claude Desktop to re-establish MCP connections

Common issues include port conflicts, Redis connection errors, or missing configuration. Ensure Docker is running, the required ports are free, and the Redis service is healthy.

Available tools

real_time_communication

Enables direct queries between agents, broadcast messaging, and asynchronous queues for coordinating tasks in real time.

file_conflict_prevention

Implements automatic file locking, conflict detection, and negotiation strategies to prevent merge issues.

shared_context_management

Provides interface/type registries, API contract sharing, and dependency tracking so agents can reuse and coordinate across modules.

task_transparency

Manages todo lists, tracks progress, signals completion, and makes task status visible to all agents.

multi_project_support

Supports isolated project namespaces with Redis-backed persistence and automatic cleanup.

modern_mcp_integration

Leverages MCP SDK 1.9.3 with decorators and STDIO protocol for full A2AMCP compatibility.