Nyko MCP Server
Provides production-ready implementation patterns for features like authentication, payments, and deployments to AI coding assistants. It enables developers to search and retrieve battle-tested code snippets and setup steps for full-stack tasks.
README
Nyko MCP Server
MCP (Model Context Protocol) server that provides battle-tested implementation patterns to AI coding assistants like Claude Code, Cursor, and Windsurf.
What is Nyko?
Nyko provides production-ready code patterns for common features:
- Auth: Supabase OAuth (Google, GitHub), Magic Links, Protected Routes
- Payments: Stripe Checkout, Webhooks, Customer Portal
- Database: RLS Policies, Migrations
- Deploy: Docker Compose, GitHub Actions + Vercel
- API: Rate Limiting with Upstash
Usage
With Claude Code
Add to ~/.claude/settings.json:
{
"mcpServers": {
"nyko": {
"url": "https://nyko-mcp.nyko-ai.workers.dev/mcp"
}
}
}
Then ask Claude:
- "Set up Google OAuth with Supabase"
- "Add Stripe payments to my app"
- "Create protected routes"
Available Tools
| Tool | Description |
|---|---|
nyko_search |
Search for patterns by keyword |
nyko_get |
Get complete pattern with code, env vars, and setup steps |
nyko_sequence |
Get ordered list of patterns for a complete feature |
nyko_check |
Check if a pattern is compatible with your project |
Development
Prerequisites
- Node.js 18+
- Wrangler CLI (
npm install -g wrangler)
Setup
# Clone the repo
git clone https://github.com/nyko-ai/mcp.git
cd mcp
# Install dependencies
npm install
# Copy env vars example
cp .dev.vars.example .dev.vars
# Edit .dev.vars with your Upstash credentials (optional for local dev)
# Start dev server
npm run dev
Testing
# Test initialize
curl -X POST http://localhost:8787/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"test","version":"1.0.0"}}}'
# Test tools/list
curl -X POST http://localhost:8787/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'
# Test nyko_search
curl -X POST http://localhost:8787/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"nyko_search","arguments":{"query":"google oauth"}}}'
Deploy
# Login to Cloudflare
wrangler login
# Set secrets
wrangler secret put UPSTASH_REDIS_REST_URL
wrangler secret put UPSTASH_REDIS_REST_TOKEN
# Deploy
npm run deploy
Architecture
┌─────────────────────────────────────────────────────────┐
│ MCP Client (Claude Code, Cursor) │
│ │ │
│ ▼ │
│ ┌─────────────────────────────────────────────────┐ │
│ │ Cloudflare Worker (MCP Server) │ │
│ │ │ │
│ │ /mcp (POST) - MCP protocol endpoint │ │
│ │ /sse (GET) - Server-sent events (future) │ │
│ │ /health - Health check │ │
│ └─────────────────────────────────────────────────┘ │
│ │ │
│ ┌──────────────┼──────────────┐ │
│ ▼ ▼ ▼ │
│ ┌───────────┐ ┌───────────┐ ┌───────────┐ │
│ │ Upstash │ │ GitHub │ │ (Future) │ │
│ │ Redis │ │ Raw │ │ Supabase │ │
│ │ (cache) │ │ (patterns)│ │ DB │ │
│ └───────────┘ └───────────┘ └───────────┘ │
└─────────────────────────────────────────────────────────┘
API Reference
Endpoints
| Endpoint | Method | Description |
|---|---|---|
/mcp |
POST | Main MCP protocol endpoint (JSON-RPC 2.0) |
/health |
GET | Health check |
Tools Reference
nyko_search
Search for patterns by keyword.
Input:
{
"query": "google oauth",
"category": "auth" // optional: auth, payments, database, deploy, api
}
Output:
{
"patterns": [
{
"id": "supabase-google-oauth",
"name": "Google OAuth with Supabase",
"description": "Complete Google OAuth implementation",
"difficulty": "intermediate",
"time_estimate": "20-30 min",
"status": "beta"
}
],
"total": 1
}
nyko_get
Get complete implementation details for a pattern.
Input:
{
"pattern_id": "supabase-google-oauth",
"has_src_dir": false // optional: adjust paths for src/ directory
}
Output:
{
"pattern": {
"id": "supabase-google-oauth",
"name": "Google OAuth with Supabase",
"install": "npm install @supabase/supabase-js@^2.49.0 @supabase/ssr@^0.5.2",
"env": [{"key": "NEXT_PUBLIC_SUPABASE_URL", "required": true}],
"files": [{"path": "lib/supabase/client.ts", "code": "..."}],
"external_setup": [{"provider": "supabase", "steps": [...]}],
"edge_cases": [{"symptom": "...", "solution": "..."}],
"validation": ["Navigate to /login", "Click Google sign in"]
}
}
nyko_sequence
Get ordered sequence of patterns for a complete feature.
Input:
{
"goal": "google auth with protected routes",
"already_implemented": [] // optional: skip patterns already done
}
Output:
{
"sequence": [
{"order": 1, "id": "supabase-client-nextjs", "reason": "Base setup"},
{"order": 2, "id": "supabase-google-oauth", "reason": "Authentication"},
{"order": 3, "id": "supabase-protected-routes", "reason": "Route protection"}
],
"total_time": "30-45 min"
}
nyko_check
Check pattern compatibility with project dependencies.
Input:
{
"pattern_id": "supabase-google-oauth",
"dependencies": {"next": "15.1.0", "@supabase/supabase-js": "2.49.0"}
}
Output:
{
"compatible": true,
"issues": [],
"missing": ["@supabase/ssr"],
"install_command": "npm install @supabase/ssr@^0.5.2"
}
Environment Variables
# Required for caching (optional for local dev)
UPSTASH_REDIS_REST_URL=https://xxx.upstash.io
UPSTASH_REDIS_REST_TOKEN=xxx
Contributing
Patterns are stored in the nyko-ai/patterns repository. See the contributing guide there for adding new patterns.
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。