Implementation of an MCP server for the RAG Web Browser Actor. This Actor serves as a web browser for large language models (LLMs) and RAG pipelines, similar to a web search in ChatGPT.
This MCP server is deprecated in favor of mcp.apify.com
For the same functionality and much more, please use one of these alternatives:
The easiest way to get the same web browsing capabilities is to use mcp.apify.com with default settings.
Benefits:
- ✅ No local setup required
- ✅ Always up-to-date
- ✅ Access to 6,000+ Apify Actors (including RAG Web Browser)
- ✅ OAuth support for easy connection
- ✅ Dynamic tool discovery
Quick Setup:
- Go to https://mcp.apify.com
- Authorize the client (Claude, VS Code, etc.)
- Copy the generated MCP server configuration (or use OAuth flow if supported)
- Start using browsing & other tools immediately
You can also call the RAG Web Browser Actor directly via its HTTP/SSE interface.
Benefits:
- ✅ Direct integration without mcp.apify.com
- ✅ Real-time streaming via Server-Sent Events
- ✅ Full control over the integration
- ✅ No additional dependencies
Docs: Actor Documentation
This server is specifically designed to provide fast responses to AI agents and LLMs, allowing them to interact with the web and extract information from web pages. It runs locally and communicates with the RAG Web Browser Actor in Standby mode, sending search queries and receiving extracted web content in response.
- Web Search: Query Google Search, scrape top N URLs, and return cleaned content as Markdown
- Single URL Fetching: Fetch a specific URL and return its content as Markdown
- Local MCP Integration: Standard input/output (stdio) communication with AI clients
- name:
search
description: Query Google Search OR fetch a direct URL and return cleaned page contents. arguments:query
(string, required): Search keywords or a full URL. Advanced Google operators supported.maxResults
(number, optional, default: 1): Max organic results to fetch (ignored whenquery
is a URL).scrapingTool
(string, optional, default:raw-http
): One ofbrowser-playwright
|raw-http
.raw-http
: Fast (no JS execution) – good for static pages.browser-playwright
: Handles JS-heavy sites – slower, more robust.
outputFormats
(array of strings, optional, default: [markdown
]): One or more oftext
,markdown
,html
.requestTimeoutSecs
(number, optional, default: 40, min 1 max 300): Total server-side AND client wait budget. A local abort is enforced.
Before (Deprecated local server):
{
"mcpServers": {
"rag-web-browser": {
"command": "npx",
"args": ["@apify/mcp-server-rag-web-browser"],
"env": {
"APIFY_TOKEN": "your-apify-api-token"
}
}
}
}
After (Recommended Apify server):
{
"mcpServers": {
"apify": {
"command": "npx",
"args": ["@apify/actors-mcp-server"],
"env": {
"APIFY_TOKEN": "your-apify-api-token"
}
}
}
}
Or use the hosted endpoint: https://mcp.apify.com
(when your client supports HTTP transport / remote MCP).
- Claude Desktop: https://claude.ai/download
- Visual Studio Code
- Apify Tester MCP Client: https://apify.com/jiri.spilka/tester-mcp-client
- Node.js (v18 or higher)
- Apify API Token (
APIFY_TOKEN
)
Clone & install:
git clone https://github.com/apify/mcp-server-rag-web-browser.git
cd mcp-server-rag-web-browser
npm install
npm install
npm run build
Since MCP servers operate over standard input/output (stdio), debugging can be challenging. For the best debugging experience, use the MCP Inspector.
You can launch the MCP Inspector via npm
with this command:
export APIFY_TOKEN=your-apify-api-token
npx @modelcontextprotocol/inspector node dist/index.js
Upon launching, the Inspector will display a URL that you can access in your browser to begin debugging.
- Model Context Protocol
- RAG Web Browser Actor
- What are AI Agents?
- What is MCP and why does it matter?
- How to use MCP with Apify Actors
- Tester MCP Client
- Webinar: Building and Monetizing MCP Servers on Apify
- How to build and monetize an AI agent on Apify
- Build and deploy MCP servers in minutes with a TypeScript template
This repository is maintained for archival purposes only. Please use the recommended alternatives above for active development.