What is dual-protocol support in mcp-use?

It implements the SEP-1865 standard, allowing the same widget code to work natively in MCP clients like Claude and via Skybridge in ChatGPT.

What is the fastest way to start a new MCP project?

Use the command 'npx create-mcp-use-app my-server' to bootstrap a new project with templates for full-featured, ChatGPT-optimized, or minimal implementations.

How does automatic widget registration work?

The server scans the 'resources/' folder for .tsx files or widget.tsx entry points, automatically extracting metadata and registering them as both MCP Tools and Resources.

How do I deploy my MCP server?

You can deploy directly to mcp-use Cloud using the 'yarn deploy' command after logging in with 'npx mcp-use login' for auto-scaled hosting.

Can I test my MCP server with ChatGPT before deploying?

Yes, you can use the built-in tunneling feature with 'mcp-use start --tunnel' to generate a public URL for testing with external clients like ChatGPT.

MCP Server Builder

Name: MCP Server Builder
Author: mcp-use

bymcp-use

•

9,036

API 开发

Streamlines the development of production-ready Model Context Protocol (MCP) servers with the mcp-use framework and automatic widget registration.

关于

This skill empowers developers to build, test, and deploy robust MCP servers using the mcp-use framework. It provides a standardized workflow for defining tools, resources, and prompts, while offering unique features like automatic React widget registration for interactive UI. With built-in support for dual-protocol standards (SEP-1865), servers created with this skill work seamlessly across MCP-native clients like Claude and AI platforms like ChatGPT, making it an essential toolkit for extending LLM capabilities with domain-specific functionality.

主要功能

Automated project bootstrapping with create-mcp-use-app templates
Dual-protocol support (SEP-1865) for Claude, Goose, and ChatGPT
Zero-downtime deployment to mcp-use Cloud with built-in scaling
Integrated local testing with built-in tunneling and Inspector UI
Automatic React widget discovery and registration from the resources folder
9,036 GitHub stars

使用场景

Exposing internal databases and proprietary APIs as secure MCP resources
Rapidly prototyping and deploying production-grade MCP servers for team use
Creating interactive UI tools for AI models using React widgets

关于

主要功能

Automated project bootstrapping with create-mcp-use-app templates
Dual-protocol support (SEP-1865) for Claude, Goose, and ChatGPT
Zero-downtime deployment to mcp-use Cloud with built-in scaling
Integrated local testing with built-in tunneling and Inspector UI
Automatic React widget discovery and registration from the resources folder
9,036 GitHub stars

使用场景

Exposing internal databases and proprietary APIs as secure MCP resources
Rapidly prototyping and deploying production-grade MCP servers for team use
Creating interactive UI tools for AI models using React widgets