MCP Server

Use Averra Extract directly from AI agents like Claude, Cursor, and ChatGPT — no code needed.

The MCP (Model Context Protocol) server wraps the full Extract API as native agent tools. Your AI assistant can extract webpages, check usage, and manage API keys on your behalf.


Install (Claude Desktop)

Add this to your Claude Desktop config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "averra-extract": {
      "command": "npx",
      "args": ["-y", "@averra/extract-mcp"],
      "env": {
        "AVERRA_EXTRACT_API_KEY": "sk_live_your_key_here"
      }
    }
  }
}

Restart Claude Desktop. You can now ask Claude things like "extract https://example.com and summarize it".

Install (Cursor)

Add to ~/.cursor/mcp.json (or use Cursor's MCP settings UI):

{
  "mcpServers": {
    "averra-extract": {
      "command": "npx",
      "args": ["-y", "@averra/extract-mcp"],
      "env": {
        "AVERRA_EXTRACT_API_KEY": "sk_live_your_key_here"
      }
    }
  }
}

Other MCP Clients

Any MCP-compatible client works. Point it at npx -y @averra/extract-mcp with AVERRA_EXTRACT_API_KEY set in the environment.


Available Tools

Once connected, your AI agent has access to these tools:

ToolWhat it does
averra_extract_urlConvert any URL to clean Markdown with metadata
averra_check_usageCheck your monthly quota and remaining requests
averra_create_api_keyCreate a new API key (inherits your current plan)
averra_list_api_keysList all API keys on your account
averra_revoke_api_keyRevoke an API key by ID

All tools support both human-readable (Markdown) and machine-readable (JSON) output formats.


Example Prompts

Once installed, try asking your AI agent:


Configuration

VariableRequiredDefaultDescription
AVERRA_EXTRACT_API_KEYYesYour API key (starts with sk_live_)
AVERRA_EXTRACT_BASE_URLNohttps://averra.devOverride API host

How It Works

The MCP server runs locally on your machine (via npx). When your AI agent needs to extract a webpage, it calls the averra_extract_url tool, which makes an API request to Averra Extract using your API key. The result comes back as clean Markdown that the agent can use directly.

Requests made through the MCP server count against your normal monthly quota, just like direct API calls. Cached pages return instantly.