Scrapi

You can use this MCP server to fetch web pages and receive clean, AI-friendly content in Markdown or plain text. By feeding it a URL, you enable large language models to read, summarize, or analyze web content without exposing extraneous page chrome, ads, or scripts.

How to use

You connect your MCP client to one of the supported transport methods and then request content from a URL. Choose Markdown output to preserve headings, lists, and links, or choose plain text when you need a simple reading strip. Use the server to fetch single pages or multiple pages in parallel and incorporate the results into your AI workflows. The server handles JavaScript rendering for dynamic pages, auto-retries on rate limits, and provides metadata like title, author, and date.

How to install

Prerequisites you need before installation:

# Option A: use via npx (no local installation)
npx -y @scrapi.ai/mcp-server
# Then configure your MCP client to point at the npx-based server

Option B: install from source and run locally. Follow these steps in order:

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

# Install dependencies and build
npm install
npm run build

# Start the server (the exact start may depend on your environment)
node dist/index.js
```)

Configuration and usage notes

You have multiple ways to connect to the MCP server. The supported options shown here include using npx for a quick start, running a local server from source, or connecting over HTTP with a remote endpoint.

API key management is required for authenticated access. When you configure a client, ensure the API key is provided via the SCRAPI_API_KEY environment variable in the chosen configuration method.

Step-by-step client setup examples

{
  "mcpServers": {
    "scrapi": {
      "command": "npx",
      "args": ["-y", "@scrapi.ai/mcp-server"],
      "env": {
        "SCRAPI_API_KEY": "your-api-key"
      }
    }
  }
}

Step 3: Restart your client

After you change the MCP server configuration, completely restart your AI client to re-establish the connection. For example, quit and reopen your client application or reload the session as appropriate for your environment.

What you can do with the server

The server exposes several endpoints you can leverage through your MCP client to fetch and process web content. Typical actions include scraping a single URL into Markdown or text, scraping multiple URLs in parallel, checking server health, and reviewing usage and billing information.

Example workflow quick start

User: Summarize this article: https://news.example.com/article/12345
Claude: [calls scrape_url]

Here's a summary of the article:

## Key Points
- Point 1: ...
- Point 2: ...
- Point 3: ...

Available tools

scrape_url

Fetches a single webpage and returns AI-ready content in Markdown or text.

scrape_urls

Fetches multiple webpages in parallel and returns AI-ready content for each URL.

scraper_server_status

Checks the status and health of MCP server instances.

get_usage

Returns API usage and remaining credits for your account.

get_billing

Provides detailed billing information including subscription, plans, daily usage, and spending limits.