MCP / AI Integration

Connect AI assistants to AWSYS.CO via the Model Context Protocol

The AWSYS.CO MCP server implements the Model Context Protocol (version 2024-11-05) over stateless Streamable HTTP. It exposes 54 tools that AI assistants such as Claude can call to create links, view analytics, manage folders, configure webhooks, and more — all through natural-language conversation.

Protocol Details

Authentication

The AWSYS MCP server accepts two authentication methods:

• OAuth 2.1 (recommended for Claude.ai connector) — Connect via the Claude.ai UI without copying an API key. Full PKCE flow, refresh tokens, and automatic re-auth. See OAuth 2.1 docs.
• API Key (existing integrations) — Pass your awsys_... key as a Bearer token. Works alongside OAuth with no migration required.

Authorization: Bearer YOUR_API_KEY

The key is hashed and matched against your account. Tier access is evaluated live on every call — upgrading or downgrading takes effect immediately.

Quick Start with Claude

If you are using Claude via the Claude Code CLI or the Claude.ai web interface with MCP support, add the AWSYS.CO server to your configuration:

{
  "mcpServers": {
    "awsys": {
      "type": "http",
      "url": "https://awsys.co/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Once connected, you can ask Claude things like:

• "Shorten https://example.com/launch with the slug launch"
• "Show me clicks for the last 7 days on my launch link"
• "Create a folder called Marketing and move my last 3 links into it"
• "Set up a webhook to notify https://my-server.com/hook on link clicks"

Available Tools (54)

Links

Analytics

QR Codes

Folders

Tags, UTM Templates, Saved Views

Domains, Namespace, Trust Score, Export

Webhooks

Affiliate Programs

Account

Example MCP Request

MCP uses JSON-RPC 2.0. Here is a raw request to call the shorten_url tool:

curl -X POST https://awsys.co/mcp \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "shorten_url",
      "arguments": {
        "url": "https://example.com/campaign",
        "slug": "launch26"
      }
    }
  }'

In practice, you do not call MCP endpoints directly — AI clients handle the JSON-RPC layer automatically when you configure the server URL and credentials.

Tier Access

Tool availability mirrors the REST API tier rules. The MCP server reads your current tier from Firestore on every request — there is no caching. Tools that require a higher tier return an error with a clear upgrade message if your account does not qualify.