MCP Uplink

MCP Uplink

A CLI and cloud platform providing managed hosting for MCP servers with intelligent tool filtering to reduce token costs and improve LLM accuracy. It serves as a secure proxy that connects AI clients to cloud tools while keeping sensitive credentials on the local machine.

Category
访问服务器

README

MCP Uplink 🚀

npm version License: MIT Node.js

The official CLI for MCP Uplink - Cloud MCP Hosting for AI Agents

Your AI Agents: Faster, Cheaper, Safer

Connect Claude Desktop, Cursor, VS Code, Windsurf and any MCP-compatible AI client to the MCP Uplink cloud platform.


⚡ Why MCP Uplink?

Problem Solution
🔥 Too many tokens Only load the tools you need. Reduce context by 60%.
💸 High API costs Fewer tokens = lower costs. Calculate savings at mcp-calculator
🤖 LLM hallucinations Less noise = more accurate responses. Specialized agents perform better.
🔐 Security concerns Zero Trust: Your credentials NEVER leave your machine.
⚙️ Infrastructure headaches We manage MCP servers. You focus on building.

🎯 3 Pillars for Optimal AI Agents

1. 🧠 Intelligent Tool Filtering

Reduce your context window by only exposing the tools your agent actually needs.

❌ Native MCP: Sends ALL 50+ tool definitions → Heavy load, confused LLM
✅ MCP Uplink: Only sends 5 tools you need → Fast, focused, accurate

Result: 60% less tokens, fewer hallucinations, better responses.

2. 🔐 Dual-Layer Security

Layer 1: Zero Trust Architecture

  • Your API keys and tokens stay in YOUR local .env file
  • Secrets are transmitted in-memory via encrypted HTTPS headers
  • We NEVER store your credentials - we're just a secure proxy

Layer 2: Tool Blocking

  • Block dangerous tools like delete, drop, modify
  • Prevent catastrophic accidents where an agent deletes your database

3. 💰 Maximum Performance, Minimum Cost

  • Less noise = faster and more accurate LLM responses
  • Agents respond instantly with higher quality answers
  • Managed infrastructure = no DevOps overhead for you

📦 Installation

No installation required! Just use npx:

npx mcp-uplink connect --url https://mcp-uplink.com/api/mcp/slack

Or install globally for faster startup:

npm install -g mcp-uplink

🚀 Quick Start

Claude Desktop Configuration

Add this to your claude_desktop_config.json:

OS Path
Mac ~/Library/Application Support/Claude/claude_desktop_config.json
Windows %APPDATA%/Claude/claude_desktop_config.json
Linux ~/.config/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "slack": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-uplink",
        "connect",
        "--url", "https://mcp-uplink.com/api/mcp/slack"
      ],
      "env": {
        "MCP_API_KEY": "your-mcp-uplink-api-key",
        "MCP_ENABLED_TOOLS": "send_message,list_channels",
        "SLACK_BOT_TOKEN": "xoxb-your-slack-token",
        "SLACK_TEAM_ID": "T0123456789"
      }
    }
  }
}

Cursor / VS Code / Windsurf

Same configuration format - just add to your MCP settings.


⚙️ CLI Options

Option Description
-u, --url <url> Required. MCP Uplink endpoint URL
-k, --api-key <key> Platform API key (or set MCP_API_KEY env var)
--enabled-tools <list> Comma-separated list of tools to enable
--no-forward-env Disable environment variable forwarding
--env-prefix <prefix> Only forward env vars with this prefix

Environment Variables

Variable Description
MCP_API_KEY Your MCP Uplink API key
MCP_ENABLED_TOOLS Tools to enable (e.g., send_message,list_users)
MCP_SERVER_URL Default server URL
MCP_ENV_PREFIX Prefix filter for env vars

🔒 Security Architecture

┌─────────────────────────────────────────────────────────────┐
│  YOUR MACHINE (Secure)                                      │
│  ┌──────────────────────────────────────────────────────┐   │
│  │  .env file                                           │   │
│  │  SLACK_BOT_TOKEN=xoxb-xxx (never leaves here)        │   │
│  └──────────────────────────────────────────────────────┘   │
│                           │                                 │
│                           ▼                                 │
│  ┌──────────────────────────────────────────────────────┐   │
│  │  mcp-uplink CLI (local process)                      │   │
│  │  - Captures env vars in-memory                       │   │
│  │  - Encrypts via HTTPS header (X-Mcp-Env)             │   │
│  └──────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────┘
                            │
                            ▼ HTTPS (encrypted)
┌─────────────────────────────────────────────────────────────┐
│  MCP UPLINK CLOUD                                           │
│  - Receives request                                         │
│  - Decrypts env vars IN-MEMORY ONLY                         │
│  - Executes MCP tool                                        │
│  - Immediately forgets credentials (no storage)             │
└─────────────────────────────────────────────────────────────┘

Key Security Features:

  • Ephemeral Secrets: Never stored, only used in-memory
  • HTTPS Transport: All data encrypted in transit
  • Blacklist Filtering: System vars (PATH, SHELL) are never forwarded
  • Tool Blocking: Prevent dangerous operations

📊 How It Works

sequenceDiagram
    participant Claude as Claude Desktop
    participant Bridge as mcp-uplink (local)
    participant Cloud as MCP Uplink Cloud
    participant MCP as MCP Server

    Claude->>Bridge: tools/list
    Bridge->>Cloud: POST /api/mcp/slack (+ X-Mcp-Env header)
    Cloud->>MCP: Execute with filtered tools
    MCP-->>Cloud: Tool definitions
    Cloud-->>Bridge: Filtered response
    Bridge-->>Claude: Only enabled tools

🔗 Links


📄 License

MIT © MCP Uplink

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选