What is MCP?
The Model Context Protocol (MCP) is an open standard for communication between AI agents and infrastructure services. Created by Anthropic and adopted across the agent ecosystem, MCP enables:
- Resources: Expose data and content to agent browsers
- Tools: Declare executable actions agents can invoke
- Prompts: Define interaction patterns for agent systems
This site provides reference documentation, validation tools, and integration examples for developers building MCP-compliant applications.
Tools & Resources
📚 Specification
Complete MCP v2025-03-26 specification, integration patterns, and best practices.
View Documentation →🔍 Validator
Validate your MCP manifests against the official schema. Ensure compatibility with agent browsers.
Validate Manifest →💻 Examples
Real-world MCP implementations, code samples, and integration templates.
Browse Examples →Quick Start
Add MCP support to your application in three steps:
1. Create Your Manifest
{
"@context": "https://modelcontextprotocol.io/schema/2025-03-26",
"protocol_version": "2025-03-26",
"name": "Your Service Name",
"version": "1.0.0",
"capabilities": {
"resources": true,
"tools": true,
"prompts": false
}
}2. Expose Your Endpoints
Place your manifest at /.well-known/mcp.json or reference it in your agent.json:
{
"@context": "https://modelcontextprotocol.io/schema/2025-03-26",
"mcp_manifest": "https://yourdomain.com/mcp.json"
}3. Validate
Use our manifest validator to ensure schema compliance before deployment.
Agent Browser Integration
MCP is supported by major agent platforms:
- ChatGPT Atlas: Native MCP support for resource discovery
- Windows Copilot: Copilot Actions built on MCP foundation
- Claude Workspaces: MCP-compatible tool invocation
- Custom Implementations: Any MCP v2025-03-26 compliant client
See our integration guide for platform-specific implementation details.