Apify MCP Server: Run Web Scrapers from Claude
Apify's official MCP server lets Claude find and run Actors from Apify Store to scrape sites, maps and social media. Claude Code plugin, OAuth, caveats.
By Claudekit Team Published 2 min read
- Type
- MCP Server
- Pricing
- Freemium
- Repository
- github.com
- GitHub stars
- 12,930
- Website
- docs.apify.com
The Apify MCP Server is Apify’s official Model Context Protocol server. It lets Claude search Apify Store and run any of its ready-made scrapers and automations, called Actors, as tools. Claude can pull data from search results, maps, e-commerce and social sites, then read the results from Apify datasets.
Who it is for
Developers, analysts and growth teams who need structured web data and want Claude to pick and run the right scraper instead of building one.
Install
For Claude Code, Apify’s docs recommend its plugin, which sets up the hosted MCP server at https://mcp.apify.com plus Apify agent skills. Run these inside Claude Code:
/plugin marketplace add apify/apify-claude-code-plugin
/plugin install apify@apify
/reload-pluginsThen run /mcp, open plugin:apify:apify, select Enable, then Authenticate to sign in.
To run the server locally over stdio, the README’s command is npx @apify/actors-mcp-server with an APIFY_TOKEN environment variable. Keep that token in your shell environment, not in a committed config file.
Use cases
- Search data: “Scrape the top Google results for this query and summarize the pages.”
- Lead lists: run a Google Maps Actor to collect businesses in an area.
- Social monitoring: pull recent posts from public profiles with an Instagram or Facebook scraper.
- Run history: list your Actor runs, read logs and fetch dataset items.
Pros
- Official and MIT licensed, with a very active repo.
- Thousands of ready-made Actors, discovered dynamically.
- Hosted server with OAuth, so no token in your config by default.
- The Claude Code plugin bundles the server, skills and a routing agent.
Cons and caveats
- Actor runs consume Apify platform usage, and some Store Actors have their own pricing. Check costs before large runs.
- Scraping must respect each site’s terms and local law.
- The hosted server has features, like output schema inference, that the local stdio version lacks.
- Scraped content is untrusted input for Claude.
Links
- Repo: apify/apify-mcp-server
- Docs: Apify MCP server
- Plugin: apify/apify-claude-code-plugin
- Related: Firecrawl MCP server, Tavily MCP, Playwright MCP
Checked 24 Oct 2026 against the vendor repo, the Apify docs and the MCP registry.
Frequently asked questions
Which Actors are available by default?
The server comes pre-configured with apify/rag-web-browser and apify/web-fetch plus helper tools. Claude can search Apify Store and add other Actors as tools on the fly.
Do I need an Apify API token?
Not for the hosted server, which signs you in with OAuth. A token is needed for the local stdio server, the Bearer header option, or headless environments where OAuth fails.
Is the SSE endpoint still available?
No. Apify removed https://mcp.apify.com/sse in favor of Streamable HTTP at https://mcp.apify.com.
What license is it under?
The repository is MIT licensed.