Puppeteer Vision MCP Server

What does “Puppeteer Vision” MCP Server do?

The Puppeteer Vision MCP Server enables AI assistants to scrape and convert web pages into Markdown format using Puppeteer, Readability, and Turndown. It offers advanced AI-driven interaction to automatically handle web elements like cookie banners, CAPTCHAs, paywalls, and more, ensuring robust content extraction even from interactive or protected sites. The server exposes this capability via the Model Context Protocol (MCP), making it easy to integrate into AI development workflows. This allows tasks such as automated web scraping, content summarization, and data ingestion to be performed seamlessly by LLMs. The server is easily deployable via npx, requires minimal configuration, and supports both stdio and SSE communication for flexible integration.

List of Prompts

No prompt templates are mentioned in the repository or documentation.

List of Resources

No specific MCP resources are listed or described in the repository or documentation.

List of Tools

• scrape-webpage: Scrapes a webpage at a specified URL, using AI to interact with and bypass interactive elements (like cookie banners or CAPTCHAs), extracts main content using Readability, and converts the result to Markdown. Parameters include:
  • url (string, required): The webpage to scrape.
  • autoInteract (boolean, optional, default: true): Whether to automatically handle interactive elements.
  • maxInteractionAttempts (number, optional, default: 3): Maximum AI interaction attempts.
  • waitForNetworkIdle (boolean, optional, default: true): Wait for network to be idle before scraping.

Use Cases of this MCP Server

• Automated Web Scraping for Knowledge Ingestion
  Developers can extract readable, well-formatted Markdown from arbitrary web pages, making it easy to ingest up-to-date content into AI workflows, databases, or knowledge bases.
• Bypassing Interactive Barriers
  The AI-driven interaction can automatically bypass CAPTCHAs, cookie banners, and other interactive obstacles, enabling seamless content extraction from sites that would otherwise block automation.
• Summarization and Content Analysis
  Extracted Markdown can be fed into LLMs for summarization, sentiment analysis, or classification, streamlining research and data processing.
• Real-time Browser Automation
  Developers can run the tool in visible (non-headless) mode for debugging, demonstrations, or when visual confirmation of browser activity is needed.
• Integration into LLM Orchestration Pipelines
  As an MCP server, it can be used as a component in orchestrators like Windsurf, Claude, Cursor, and Cline, expanding AI agents’ abilities to interact with the live web.

How to set it up

Windsurf

1. Prerequisites: Install Node.js and npm.

2. Environment Setup: Create a .env file or export the required environment variables, including OPENAI_API_KEY.

3. Edit Configuration: Locate Windsurf’s configuration file.

4. Add Puppeteer Vision MCP: Insert the following JSON snippet:

   {
     "mcpServers": {
       "web-scraper": {
         "command": "npx",
         "args": ["-y", "puppeteer-vision-mcp-server"],
         "env": {
           "OPENAI_API_KEY": "YOUR_OPENAI_API_KEY_HERE"
         }
       }
     }
   }
   

5. Save/Restart: Save the file and restart Windsurf.

6. Verify: Check logs or UI to confirm the MCP server is running.

Securing API Keys:
Store secrets in environment variables (e.g., .env):

"env": {
  "OPENAI_API_KEY": "${OPENAI_API_KEY}"
}

Claude

1. Prerequisites: Ensure Node.js and npm are installed.

2. Set Environment: Prepare .env or export OPENAI_API_KEY and other variables.

3. Edit Configuration: Open Claude’s MCP configuration.

4. Add the MCP Server:

   {
     "mcpServers": {
       "web-scraper": {
         "command": "npx",
         "args": ["-y", "puppeteer-vision-mcp-server"],
         "env": {
           "OPENAI_API_KEY": "YOUR_OPENAI_API_KEY_HERE"
         }
       }
     }
   }
   

5. Restart Claude: Apply changes and restart the platform.

6. Verify: Confirm successful startup.

Cursor

1. Prerequisites: Install Node.js and npm.

2. Environment: Set up .env with the OpenAI API key.

3. Edit Cursor Config: Add the MCP server as below:

   {
     "mcpServers": {
       "web-scraper": {
         "command": "npx",
         "args": ["-y", "puppeteer-vision-mcp-server"],
         "env": {
           "OPENAI_API_KEY": "YOUR_OPENAI_API_KEY_HERE"
         }
       }
     }
   }
   

4. Save & Restart: Save changes and restart Cursor.

5. Check Logs: Ensure the server is running.

Cline

1. Prerequisites: Install Node.js and npm.

2. Environment: Set or export OPENAI_API_KEY.

3. Configuration: Add to Cline’s MCP config:

   {
     "mcpServers": {
       "web-scraper": {
         "command": "npx",
         "args": ["-y", "puppeteer-vision-mcp-server"],
         "env": {
           "OPENAI_API_KEY": "YOUR_OPENAI_API_KEY_HERE"
         }
       }
     }
   }
   

4. Restart Cline: Apply and restart.

5. Confirm: Validate that the server is accessible.

Note: Secure API keys via environment variables and never hard-code secrets in config files.

How to use this MCP inside flows

Using MCP in FlowHunt

To integrate MCP servers into your FlowHunt workflow, start by adding the MCP component to your flow and connecting it to your AI agent:

Click on the MCP component to open the configuration panel. In the system MCP configuration section, insert your MCP server details using this JSON format:

{
  "puppeteer-vision": {
    "transport": "streamable_http",
    "url": "https://yourmcpserver.example/pathtothemcp/url"
  }
}

Once configured, the AI agent is now able to use this MCP as a tool with access to all its functions and capabilities. Remember to change “puppeteer-vision” to whatever the actual name of your MCP server is and replace the URL with your own MCP server URL.

Overview

| Roots Support | ⛔ | No mention of Roots. |

Based on the above, the Puppeteer Vision MCP Server offers a robust and focused web scraping tool with strong documentation and security guidance, but lacks multiple tools, prompt templates, resources, and advanced MCP features like roots or sampling. Its one-tool, one-purpose design gives it high reliability for its use case, but limits extensibility.

Our opinion

MCP Score: 5/10
This MCP server is well-documented, useful for its specific purpose, and easy to set up, but its lack of prompt templates, explicit resources, and advanced MCP features (roots, sampling) limits its versatility and ecosystem integration.

MCP Score