Apify's official MCP server gives an AI client access to the Apify Store — thousands of ready-made "Actors" (scrapers and automations) for Google, Instagram, Amazon, maps, TikTok and more — that it can discover and run on demand. Here's the setup.
Step 1 — get an API token
Create an Apify account and copy your API token from account settings. Running Actors consumes Apify platform credits (there's a free monthly tier).
Step 2 — configure the client
Local via npx:
{
"mcpServers": {
"apify": {
"command": "npx",
"args": ["-y", "@apify/actors-mcp-server"],
"env": {
"APIFY_TOKEN": "your_apify_token"
}
}
}
}
Apify also runs a hosted endpoint (mcp.apify.com) you can add as a remote/OAuth connector. Restart the client; see MCP config file location.
What it gives your AI
Search the Apify Store for an Actor, then run it with parameters and get structured results — e.g. scrape search results, social profiles, or product listings, without writing a scraper.
Verify
Ask: "Find an Apify Actor that scrapes Google Maps and pull the top 10 coffee shops in Berlin." Structured data back means it works.
Common problems
unauthorized— wrong API token.- Actor run fails / credits — running Actors costs credits; check balance and the Actor's input schema.
- Too many tools — exposing every Actor floods the tool list; restrict to specific Actors via the server's config.
Going further
Apify pairs with Sequential Thinking and a filesystem/spreadsheet server for data pipelines. Browse the search category or curated loadouts. Scrape within sites' terms and keep your token private — see MCP security best practices.