MCP server

Chuger ships a first-class Model Context Protocol server on every plan. Once connected, your AI client can scrape pages, extract clean markdown, search the web, map sites, and check your credit balance — as native tools, with no glue code.

https://api.chuger.com/mcp/chuger

The server uses the streamable HTTP transport and authenticates with the same Bearer API token as the REST API (see Authentication).

Fastest setup: paste a prompt

Copy the agent install prompt, replace <CHUGER_API_TOKEN> with your token, and paste it into your AI chat. The agent detects its own client, adds the MCP server, installs the Chuger skill, and verifies the connection — no manual configuration.

Manual setup

claude mcp add --transport http chuger https://api.chuger.com/mcp/chuger \
  --header "Authorization: Bearer YOUR_API_TOKEN"

{
  "mcpServers": {
    "chuger": {
      "type": "http",
      "url": "https://api.chuger.com/mcp/chuger",
      "headers": { "Authorization": "Bearer YOUR_API_TOKEN" }
    }
  }
}

{
  "mcpServers": {
    "chuger": {
      "type": "http",
      "url": "https://api.chuger.com/mcp/chuger",
      "headers": { "Authorization": "Bearer ${CHUGER_API_TOKEN}" }
    }
  }
}

Available tools

Credits, rate limits, and billing work exactly like the REST API — see Credits.

Built to save tokens

• extract-content returns markdown only by default. Request formats: ["html", "metadata"] only when you need them.
• Content responses include an estimated_tokens field so agents can budget context before reading further.
• map-domain returns a URL list without fetching any page content — map first, then extract only the pages you need.

Troubleshooting