convez.ai
MCP ServerDesenvolvimento WebIntermediário
forbiddenlink

MCP Server Studio

forbiddenlink/mcp-server-studio

Visual builder e IDE para criar servidores MCP com drag-and-drop

Ver no GitHub
InicialPluginScore · de 0 a 100

IDE visual para construir, testar e gerenciar servidores MCP (Model Context Protocol) sem escrever código boilerplate. Oferece canvas de drag-and-drop, simulador de testes interativo e geração automática de código TypeScript production-ready com suporte a Docker e Railway.

Instalação

$npx convez-ai forbiddenlink/mcp-server-studio

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
2
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 2 dias
Versão
v1.1.3
Issues abertas
2
Contribuidores
3

Casos de uso

  • Criar servidores MCP rapidamente sem escrever boilerplate
  • Aprender a especificação MCP de forma visual e interativa
  • Prototipar integrações de ferramentas de IA
  • Converter OpenAPI specs automaticamente em MCP tools
  • Gerar código TypeScript production-ready e fazer deploy

Veredito

MCP Server Studio é uma ferramenta madura e bem arquitetada que elimina atrito na criação de servidores MCP. Ideal para quem quer prototipar rapidamente ou aprender MCP sem mergulhar em código boilerplate; suporte a OpenAPI import e geração de código com IA agregam valor real.

README

A tradução para o português deste README ainda está na fila.

⚡ MCP Server Studio

Visual builder for Model Context Protocol servers

Design MCP tools, resources, and prompts on a drag-and-drop canvas. Test them in a built-in simulator. Export production-ready TypeScript server code with Docker and Railway deployment bundles.

CI Next.js TypeScript Tests License


What is this?

MCP Server Studio makes it easy to create Model Context Protocol servers without writing boilerplate code. Design your tools visually, test them in a playground, and export ready-to-use TypeScript code.

Perfect for:

  • Building MCP servers quickly
  • Learning the MCP specification
  • Prototyping AI tool integrations
  • Teaching MCP concepts visually

Features

Visual Canvas — Drag-and-drop tool, resource, and prompt creation with 29 templates across 7 categories. Intuitive node-based interface with real-time manifest generation.

Test Simulator — Interactive test inspector with parameter validation, batch schema validation, and structured test results for tools, resources, and prompts.

Code Generation — Production-ready TypeScript with Zod schemas, full MCP SDK integration, proper error handling, and support for both stdio and HTTP/SSE transports.

Export & Deploy — Export as TypeScript, Docker bundle, or Railway bundle. Includes Dockerfile, docker-compose, package.json, tsconfig, and deployment instructions.

AI Tool Generator — Describe a tool in natural language and the parser generates a full MCP tool definition with parameters, types, and icons.

OpenAPI Import — Paste an OpenAPI/Swagger spec (JSON or YAML) and automatically convert API operations to MCP tools.

Advanced Capabilities — Configure sampling (LLM inference), elicitation (user input), and tasks (async operations) on a per-tool basis.


Quick Start

Prerequisites: Node.js 18+ and pnpm 8+

git clone https://github.com/lizstein/mcp-server-studio.git
cd mcp-server-studio
pnpm install
pnpm dev

Visit http://localhost:3000


Development

pnpm dev          # Start dev server (localhost:3000)
pnpm build        # Production build (includes TypeScript checking)
pnpm lint         # ESLint (flat config, strict no-explicit-any)
pnpm test:run     # Vitest single run (466 tests)
pnpm test         # Vitest watch mode

Run a single test file:

pnpm vitest run lib/generators/__tests__/constraints.test.ts

Tech Stack

  • Next.js 16 — React 19, App Router, Turbopack
  • TypeScript 5 — Strict mode
  • React Flow (@xyflow/react) — Visual canvas and node system
  • Zustand — Global state management with localStorage persistence
  • Monaco Editor — Code preview and export
  • shadcn/ui + Radix — UI component library
  • Tailwind CSS v4 — Styling with custom dark theme
  • Framer Motion — Animations
  • Vitest — Testing with jsdom environment

Project Structure

mcp-server-studio/
├── app/                         # Next.js App Router
│   ├── layout.tsx               # Root layout with metadata
│   └── page.tsx                 # Main app page (client component)
├── components/
│   ├── canvas/                  # React Flow canvas
│   │   ├── CanvasPanel.tsx      # Main canvas with toolbar
│   │   ├── ToolNode.tsx         # Tool node component
│   │   ├── ResourceNode.tsx     # Resource node component
│   │   ├── PromptNode.tsx       # Prompt node component
│   │   └── DataFlowEdge.tsx     # Custom edge component
│   ├── config/                  # Slide-in configuration panels
│   │   ├── ToolConfigPanel.tsx
│   │   ├── ResourceConfigPanel.tsx
│   │   ├── PromptConfigPanel.tsx
│   │   ├── ServerConfigPanel.tsx
│   │   ├── CapabilitiesSection.tsx
│   │   ├── SamplingConfigPanel.tsx
│   │   └── ElicitationConfigPanel.tsx
│   ├── preview/                 # Right sidebar tabs
│   │   ├── PreviewPanel.tsx     # Tab container
│   │   ├── StructureTab.tsx     # JSON manifest viewer
│   │   ├── TestTab.tsx          # Interactive test simulator
│   │   └── CodeTab.tsx          # Monaco code editor
│   └── ui/                      # shadcn/ui + custom components
│       ├── ai-generator.tsx     # Natural language tool generator
│       ├── command-palette.tsx  # Cmd+K command palette
│       ├── import-dialog.tsx    # OpenAPI import dialog
│       ├── template-gallery.tsx # Template browser
│       └── error-boundary.tsx   # React error boundary
├── lib/
│   ├── generators/              # Code generation (pure functions)
│   │   ├── mcpServerGenerator.ts
│   │   ├── manifestGenerator.ts
│   │   ├── dockerGenerator.ts
│   │   ├── railwayGenerator.ts
│   │   ├── readmeGenerator.ts
│   │   ├── exportBundler.ts
│   │   ├── zipCreator.ts
│   │   ├── aiToolGenerator.ts
│   │   └── __tests__/           # Generator tests
│   ├── importers/
│   │   ├── openApiImporter.ts
│   │   └── __tests__/           # Importer tests
│   ├── simulators/
│   │   ├── mcpTestSimulator.ts
│   │   └── __tests__/           # Simulator tests
│   ├── store/
│   │   └── useStore.ts          # Zustand store with undo/redo
│   ├── utils/
│   │   ├── sanitize.ts          # Code generation sanitization
│   │   └── __tests__/           # Utility tests
│   ├── templates/
│   │   └── toolTemplates.ts     # 29 pre-built tool templates
│   └── types.ts                 # TypeScript type definitions
└── .github/workflows/ci.yml    # CI pipeline (lint, test, build)

Testing

183 tests across 10 test files covering:

  • Zod schema generation with all constraint types
  • JSON Schema manifest generation
  • Runtime parameter validation (all types and formats)
  • MCP server code generation (stdio + HTTP transports)
  • Docker and Railway config generation
  • OpenAPI/Swagger import parsing
  • AI tool description parsing
  • Test simulator execution and batch validation
  • Export bundle composition
  • Input sanitization for safe code generation

Roadmap

Done

  • Visual tool, resource, and prompt builder
  • 29 templates across 7 categories
  • Interactive test simulator
  • TypeScript code generation (stdio + HTTP/SSE)
  • Docker and Railway deployment exports
  • AI tool generator (natural language → MCP tool)
  • OpenAPI/Swagger import
  • Sampling, elicitation, and tasks configuration
  • Undo/redo with history stack
  • Command palette (Cmd+K)
  • localStorage persistence
  • CI pipeline

Next

  • Custom tool icons
  • npm package scaffolding
  • Team sharing
  • Workflow connections between tools

Contributing

Contributions welcome!

  1. Fork the repo
  2. Create your feature branch (git checkout -b feature/amazing)
  3. Commit your changes (git commit -m 'Add amazing feature')
  4. Push to the branch (git push origin feature/amazing)
  5. Open a Pull Request

License

MIT License - see LICENSE for details


Acknowledgments


Built by Liz Stein.

⚡ Start building MCP servers visually!

Ver o original no GitHub

Bom para

Relacionados