⚡ MCP Forge
Visual MCP Server Builder & Registry
Build, test, and ship Model Context Protocol servers without writing boilerplate. MCP Forge provides a visual interface for creating MCP servers with tools, resources, and prompts — generating production-ready TypeScript code that follows the official MCP SDK patterns.
What is MCP?
The Model Context Protocol is an open standard created by Anthropic that enables AI applications (like Claude Desktop, Cursor, VS Code Copilot) to connect to external data sources and tools through a unified interface. Think of it as USB-C for AI integrations — one protocol to connect any tool to any AI client.
MCP servers expose three primitives:
- Tools — Functions the LLM can call (e.g., get_weather, query_database)
- Resources — Data the application can read (e.g., file contents, API responses)
- Prompts — Pre-written templates for common tasks
MCP Forge makes building these servers visual and fast.
Features
🔧 Visual Server Builder
Design MCP servers by clicking — not coding. Add tools with typed parameters, resources with URI patterns, and prompt templates. The three-panel layout shows your primitive list, property editor, and live code output simultaneously.
📦 Server Registry
Browse and search existing MCP servers from the Anthropic ecosystem and community. Filter by official/community, search by tags, and get instant install commands for any server.
▶️ Testing Playground
Simulate JSON-RPC calls against your server definition. See formatted request/response logs, test tools/list and tools/call methods, and verify your server behavior before deployment.
⚙️ Config Generator
Auto-generates your claude_desktop_config.json, package.json, and complete TypeScript server code. Copy and deploy — everything follows the official @modelcontextprotocol/sdk patterns.
Tech Stack
| Layer | Technology |
|---|---|
| Framework | Next.js 15 (App Router) |
| Language | TypeScript 5.3 |
| Styling | Tailwind CSS 3.4 + CSS Variables |
| State | React hooks (useState, useCallback, useRef) |
| Code Gen | AST-based TypeScript generation |
| Font | IBM Plex Mono |
| Deploy | Vercel / Netlify / GitHub Pages |
Quick Start```bash
Clone the repo
git clone https://github.com/YOUR_USERNAME/mcp-forge.git
cd mcp-forge
Install dependencies
npm install
Run development server
npm run dev
Open http://localhost:3000 to see the app.
---
## Project Structure```
mcp-forge/
├── src/
│ ├── app/
│ │ ├── page.tsx # Landing page
│ │ ├── builder/
│ │ │ └── page.tsx # Main builder application
│ │ ├── layout.tsx # Root layout
│ │ └── globals.css # Global styles
│ ├── components/
│ │ ├── landing/
│ │ │ ├── Hero.tsx # Landing hero section
│ │ │ ├── Features.tsx # Feature showcase
│ │ │ └── CTA.tsx # Call-to-action
│ │ ├── builder/
│ │ │ ├── BuilderTab.tsx # Server builder interface
│ │ │ ├── RegistryTab.tsx # Server registry/marketplace
│ │ │ ├── PlaygroundTab.tsx # JSON-RPC testing
│ │ │ ├── ConfigTab.tsx # Config file generator
│ │ │ ├── CodePreview.tsx # Live TypeScript output
│ │ │ └── PrimitiveEditor.tsx # Tool/Resource/Prompt editor
│ │ └── ui/
│ │ ├── Badge.tsx
│ │ ├── TextInput.tsx
│ │ └── SelectInput.tsx
│ ├── lib/
│ │ ├── codegen.ts # TypeScript code generation
│ │ ├── registry.ts # Server registry data
│ │ └── types.ts # MCP primitive types
│ └── types/
│ └── index.ts # TypeScript interfaces
├── public/
│ └── og-image.png # Social preview image
├── next.config.ts
├── tailwind.config.ts
├── tsconfig.json
├── package.json
└── README.md
How It Works
1. Define Your Server
Name your server, choose a transport type (stdio, HTTP, or SSE), and start adding primitives.
2. Add Primitives
Click to add tools, resources, or prompts. Each gets a full property editor:
- Tools: Name, description, typed parameters (string/number/boolean/array/object), required flags
- Resources: URI patterns, MIME types
- Prompts: Template strings with variable interpolation
3. Preview Code
The right panel generates valid TypeScript in real-time using @modelcontextprotocol/sdk:```typescript
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
import { z } from "zod";
const server = new McpServer({
name: "my-server",
version: "1.0.0",
});
server.tool(
"get_inventory",
"Fetch current vehicle inventory",
{
make: { type: z.string(), description: "Vehicle make" },
year: { type: z.number(), description: "Model year" },
},
async ({ make, year }) => {
// TODO: implement
return { content: [{ type: "text", text: "result" }] };
}
);
```
4. Test & Deploy
Use the Playground to simulate JSON-RPC calls, then grab your generated claude_desktop_config.json and deploy.
MCP Specification Compliance
MCP Forge generates code targeting the 2025-11-25 MCP specification, including:
- ✅ JSON-RPC 2.0 message format
- ✅
tools/listandtools/callmethods - ✅
resources/readwith URI patterns - ✅
prompts/getwith argument schemas - ✅ Zod-based parameter validation
- ✅ Stdio and HTTP transport support
- ✅ Proper error handling patterns
Roadmap
- [ ] Import from OpenAPI — Auto-generate MCP servers from OpenAPI/Swagger specs
- [ ] Python code generation — Output Python SDK code alongside TypeScript
- [ ] MCP Inspector integration — Connect directly to the official MCP Inspector for live testing
- [ ] One-click publish — Push to npm registry with proper package configuration
- [ ] Server templates — Pre-built starter templates for common integrations (database, API wrapper, file system)
- [ ] Auth configuration — OAuth 2.0 Resource Server setup per June 2025 spec update
Why MCP Forge?
The MCP ecosystem is growing fast — adopted by OpenAI, Google DeepMind, Microsoft, and thousands of developers. But building MCP servers still requires manually writing boilerplate code, understanding the JSON-RPC protocol, and configuring transport layers.
MCP Forge bridges the gap between understanding what MCP can do and shipping a working server. It's the tool I wished existed when I started building MCP integrations for production automotive AI systems.
Related Projects
- AgentForge — TypeScript framework for AI agents
- FlowForge — Visual AI workflow builder
- QUIRK AI Kiosk — Full-stack showroom AI assistant
Contributing
Contributions are welcome! Please read our Contributing Guide for details.
- Fork the repo
- Create your feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
License
MIT — see LICENSE for details.
Built with ⚡ by Michael
Targeting MCP Spec 2025-11-25 · Powered by Next.js + TypeScript
