🌐 AnySite MCP Server

Agent-First Web Scraping Infrastructure via Model Context Protocol

Connect your AI agents to real-time data from LinkedIn, Instagram, Reddit, Twitter, and any website through a single MCP server.

🚀 Quick Start • 📖 Documentation • 🎮 Playground • 💬 Discord • 🐦 Twitter

AnySite MCP Server is a Model Context Protocol (MCP) implementation that gives AI agents direct access to web data through platform-specific APIs. Unlike traditional web scrapers, AnySite provides:

• 🔒 OAuth Authentication - Secure, one-click connection for Claude Desktop and other MCP clients
• 🌐 Multi-Platform Support - LinkedIn, Instagram, Reddit, Twitter, and custom web parsing
• 🤖 Agent-First Design - Built specifically for AI agents with structured data formats
• 🔄 Self-Healing APIs - Auto-recovery from platform changes and rate limits
• ⚡ Real-Time Data - Fresh data extraction without stale caches

Perfect for: AI research, lead generation, market intelligence, content monitoring, competitive analysis

• Advanced Search & Filtering - Find people by title, company, location, education, skills
• Bulk Data Extraction - Extract thousands of profiles, posts, or comments in one request
• Network Analysis - Map connections, followers, engagement patterns
• Content Monitoring - Track posts, comments, reactions in real-time
• Account Management - Send messages, connection requests, post comments (LinkedIn)
• Smart Web Parsing - Extract structured data from any website using CSS selectors

• Rate Limit Management - Automatic backoff and retry with exponential delays
• Proxy Rotation - Built-in proxy support for high-volume requests
• Error Recovery - Self-healing mechanisms for platform changes
• Usage Analytics - Track API consumption and costs
• Team Management - Multi-user accounts with role-based access

Perfect for Claude Desktop, Cline, Cursor, Windsurf, and other MCP clients that support OAuth.

1. Sign up at app.anysite.io (100 free credits included)
2. Navigate to MCP Server Integration
3. Copy your OAuth URL: https://api.anysite.io/mcp/sse

{
  "mcpServers": {
    "anysite": {
      "command": "npx",
      "args": ["-y", "@anysite/mcp"],
      "env": {
        "ANYSITE_OAUTH_URL": "https://api.anysite.io/mcp/sse"
      }
    }
  }
}

Ask your AI agent:

What MCP tools do you have access to?

Expected response should include:

• search_linkedin_users
• get_linkedin_profile
• get_instagram_user
• search_reddit_posts
• google_search
• parse_webpage
• ... and 50+ more tools

Perfect for testing, development, custom integrations.

# Clone the repository
git clone https://github.com/anysiteio/anysite-mcp-server.git
cd anysite-mcp-server

# Install dependencies
npm install

# Build the project
npm run build

Create .env file:

ANYSITE_ACCESS_TOKEN=your_access_token
ANYSITE_ACCOUNT_ID=your_account_id

Get your credentials from app.anysite.io

npm start

Add to your MCP configuration:

{
  "mcpServers": {
    "anysite-local": {
      "command": "node",
      "args": ["/path/to/anysite-mcp-server/build/index.js"],
      "env": {
        "ANYSITE_ACCESS_TOKEN": "your_token",
        "ANYSITE_ACCOUNT_ID": "your_account_id"
      }
    }
  }
}

Find me 10 CTOs at AI companies in San Francisco

The MCP server will:

1. Search LinkedIn users with title="CTO", company_keywords="AI", location="San Francisco"
2. Return structured profiles with name, headline, company, location
3. Provide direct LinkedIn URLs for each profile

Get the latest 20 Instagram posts mentioning @yourbrand

Search Reddit for posts about "LLM agents" in the last week,
sorted by top engagement

1. Find the LinkedIn profile of John Doe at Company X
2. Get his recent posts and engagement metrics
3. Find his Twitter profile and latest tweets
4. Cross-reference with Instagram presence

• AnySite MCP Server Docs
• API Reference
• All Available Tools

AnySite MCP Server works with any MCP-compatible client:

• ✅ Claude Desktop - Native OAuth support
• ✅ Cline - MCP configuration
• ✅ Cursor - Custom MCP server
• ✅ Windsurf - MCP integration
• ✅ Any MCP Client - Standard protocol

• 🔧 n8n - AnySite n8n nodes
• 🔧 LangChain - Custom tool integration
• 🔧 AutoGen - Agent tool registration
• 🔧 REST API - Direct API access

Recommendation: Use Remote MCP for production and Local MCP for development/testing.

┌─────────────────┐
│   MCP Client    │ (Claude Desktop, Cline, etc.)
│   (AI Agent)    │
└────────┬────────┘
         │ MCP Protocol
         │
┌────────▼────────┐
│  AnySite MCP    │
│     Server      │
└────────┬────────┘
         │ REST API
         │
┌────────▼────────┐
│  AnySite API    │ ← OAuth Authentication
│   Platform      │ ← Rate Limiting
└────────┬────────┘ ← Proxy Rotation
         │
    ┌────┴────┬────────┬─────────┬──────────┐
    ▼         ▼        ▼         ▼          ▼
LinkedIn  Instagram  Reddit  Twitter  Any Website

Key Components:

• MCP Protocol Layer - Standardized tool interface for AI agents
• API Abstraction - Platform-specific API adapters with error handling
• Authentication - OAuth 2.0 for secure credential management
• Self-Healing - Auto-retry and fallback mechanisms
• Rate Limiting - Smart backoff to prevent API bans

• Find decision makers by title, company, location
• Enrich CRM data with LinkedIn profiles
• Monitor competitor hiring and expansion

• Track brand mentions across platforms
• Analyze competitor social media strategy
• Monitor industry trends and discussions

• Collect datasets for AI training
• Social network analysis
• Content performance tracking

• Multi-platform data correlation
• Automated outreach campaigns
• Cross-platform identity resolution

# Clone repository
git clone https://github.com/anysiteio/anysite-mcp-server.git
cd anysite-mcp-server

# Install dependencies
npm install

# Build TypeScript
npm run build

# Run in development mode
npm run dev

# Run tests
npm test

anysite-mcp-server/
├── src/                      # Source code
│   ├── index.ts              # MCP server entry point (Smithery runtime)
│   ├── server.ts             # Server implementation (Traditional MCP)
│   └── types.ts              # TypeScript type definitions
├── build/                    # Compiled JavaScript
│   ├── index.js              # Main entry point
│   ├── server.js             # Server executable
│   ├── types.js              # Type definitions
│   ├── remote-server.js      # Remote server support
│   ├── streamable-server.js  # Streaming support
│   └── smithery.js           # Smithery integration
├── .claude/                  # Claude Code settings
│   └── settings.local.json
├── .cursor/                  # Cursor IDE settings
│   └── rules/
├── .smithery/                # Smithery CLI files
│   └── index.cjs
├── package.json              # npm package configuration
├── package-lock.json
├── tsconfig.json             # TypeScript configuration
├── smithery.yaml             # Smithery config
├── glama.json                # Glama integration
├── .env                      # Environment variables (local)
├── .npmrc                    # npm configuration
├── .gitignore
├── README.md                 # Documentation
├── CLAUDE.md                 # Claude Code instructions
├── LICENSE.md                # MIT License
└── LICENSE

We welcome contributions! See CONTRIBUTING.md for guidelines.

Areas we need help with:

• 📝 Documentation improvements
• 🐛 Bug fixes and testing
• ✨ New platform integrations
• 🌍 Translations

• ✅ 100 free credits on signup
• ✅ All 57 tools available
• ✅ OAuth authentication
• ⚠️ Rate limits apply

• 🚀 Higher rate limits
• 🚀 Priority support
• 🚀 Dedicated proxies
• 🚀 Team collaboration

See pricing details

• OAuth 2.0 - Industry-standard authentication
• No Credential Storage - Your API keys stay with AnySite
• Encrypted Transport - All data transmitted over HTTPS
• GDPR Compliant - Data processing follows EU regulations
• Revocable Access - One-click disconnect in dashboard

⚠️ Important: Always comply with platform terms of service and local regulations when scraping data.

• 📖 Documentation
• 💬 Discord Community - Ask questions, share workflows
• 🐦 Twitter - Latest updates
• 📧 Email Support
• 🐛 GitHub Issues

This project is licensed under the MIT License.

Built with:

• Model Context Protocol by Anthropic
• TypeScript
• Node.js

Special thanks to the MCP community for feedback and contributions.