Prysm

The Prysm MCP Server lets you deploy a flexible web-scraping agent that AI assistants can use to extract and format content from the web. It supports multiple scraping modes, output formats, image handling, and configurable scrolling to adapt to diverse sites, all exposed through MCP-compatible endpoints or local runtimes.

How to use

Set up Prysm MCP Server as a local (stdio) runtime or connect to it via an MCP-enabled client. Choose a scraping mode according to your needs: focused for speed, balanced for general use, or deep for thorough extraction. Format results as markdown, HTML, or JSON, and optionally include and download images. You can tailor scrolling for single-page apps and let Prysm adapt to different site structures. Use the server to produce ready-to-consume results for your AI workflows.

How to install

Prerequisites: ensure you have Node.js and npm installed on your system.

# Recommended: Install the LLM-optimized version
npm install -g @pinkpixel/prysm-mcp

# Or install the standard version
npm install -g prysm-mcp

# Or clone and build
git clone https://github.com/pinkpixel-dev/prysm-mcp.git
cd prysm-mcp
npm install
npm run build

Using npx for quick start

You can run the server directly with npx without installing the package globally. Use default settings or customize output directories as needed.

# Run with default settings
npx @pinkpixel/prysm-mcp

# Run with custom output directories
PRYSM_OUTPUT_DIR=./my-output PRYSM_IMAGE_OUTPUT_DIR=./my-output/images npx @pinkpixel/prysm-mcp

Configuration for MCP integration

Create a configuration file to register Prysm MCP Server as an MCP endpoint. The following example registers a local stdio-based server that you can start from an MCP client. It assigns dedicated output directories for scraped data and downloaded images.

{
  "mcpServers": {
    "prysm-scraper": {
      "description": "Prysm web scraper with custom output directories",
      "command": "npx",
      "args": [
        "-y",
        "@pinkpixel/prysm-mcp"
      ],
      "env": {
        "PRYSM_OUTPUT_DIR": "${workspaceFolder}/scrape_results",
        "PRYSM_IMAGE_OUTPUT_DIR": "${workspaceFolder}/scrape_results/images"
      }
    }
  }
}

Available tools

scrapeFocused

Fast web scraping optimized for speed with main content extraction only. Parameters include url, maxScrolls, scrollDelay, scrapeImages, downloadImages, maxImages, and output.

scrapeBalanced

Balanced web scraping with good coverage and reasonable speed. Includes options for timeouts and scroll control.

scrapeDeep

Maximum extraction mode offering thorough scraping with higher maxScrolls and image handling limits for comprehensive results.

formatResult

Format scraped data into markdown, HTML, or JSON, with optional image inclusion and an output file path.