Apify Store holds thousands of maintained scrapers — Google Maps, Instagram, TikTok, Amazon, LinkedIn, search engines and much more. The official MCP server turns that catalogue into tools an agent can discover and call at runtime, instead of you wiring up each scraper by hand.
https://mcp.apify.com
Claude Sonnet 4.5
Chat with 60+ AI models on the same workflow — switch to a different model mid-conversation and re-run the same prompt, or use Compare mode to put several side-by-side and balance quality vs. cost.
OAuth in the browser on first connection, or an Apify API token as a bearer header. Actor runs consume Apify platform credit.
How models use it and what it is built for.
The interesting design decision here is dynamic tool discovery. Rather than shipping a fixed tool list, the server lets an agent search Apify Store for an Actor that fits the task, read its input schema, and then call it — so a request the server was never configured for can still be satisfied. Results land in an Apify dataset, which the agent reads back with pagination and filtering, so a scrape returning fifty thousand rows does not have to pass through the context window in one piece. Actors are billed per run on the Apify platform, which is worth understanding before an agent starts calling them in a loop. The server also exposes Apify documentation search, which is genuinely useful when the agent needs to work out an Actor’s input shape. It is hosted at mcp.apify.com over streamable HTTP with OAuth, and also available as a local stdio server.
Typical tools an AI model can call. Exact names vary by version.
Taken from the official Apify documentation — see Apify MCP server — official documentation for the full reference.
Environment variables
APIFY_TOKENAPI token from Apify Console under API & Integrations. Only needed if your client cannot do OAuth, or for the local stdio server.
Client configuration
Local stdio server
Same tools, running on your machine, with the token in your own environment rather than exchanged over OAuth.
{
"mcpServers": {
"apify": {
"command": "npx",
"args": ["-y", "@apify/actors-mcp-server"],
"env": { "APIFY_TOKEN": "YOUR_APIFY_TOKEN" }
}
}
}Copy any of these into MCP Agent Studio after connecting.
Find every coffee shop in Lisbon on Google Maps with their rating and review count.
Scrape the last 50 posts from this Instagram profile and summarise the themes.
Find an Actor that can scrape Amazon product listings and show me its inputs.
Collect the top 100 Google results for this query with titles and snippets.
This is not a single-model product: you get the same MCP connection with 60+ models (Claude, GPT, Gemini, DeepSeek, open-weight, and more), you can switch mid-conversation, and you can open Compare mode to run the same prompt against multiple models at once. The card above is a suggested starting point for this server — not the only choice.
Default pick for Apify
Claude Sonnet 4.5
Dynamic tool discovery means the model must read an Actor’s input schema and construct a valid payload. Sonnet 4.5 is reliable at that; smaller models guess the shape.
Listing tools proves the server is reachable, not that a model can work with it. Evals go further: they read every tool on the server, write a test suite from its real schemas, and run it — code decides pass/fail on the responses (schema conformance, error codes, pagination, result caps) while a scoring model grades plain-English tasks driven through the tools.
Get a pass/fail report per tool with the evidence behind each verdict — and replay the same suite after every schema change. Destructive tools are excluded from the run.
Run evalsOpen MCP Agent Studio with the connection pre-filled. Add your token, pick any of 60+ models, and start chatting — no install required.
Open Agent StudioCommon questions about connecting, scoping and using it safely.
It is Apify’s official MCP server, hosted at mcp.apify.com. It lets an AI agent search Apify Store, run any of the thousands of Actors published there, and read the resulting datasets — covering social media, maps, search engines, e-commerce and general web scraping.
It does not expose them all at once. The agent searches the store, fetches the input schema for the Actor it picked, and calls it — or adds it to the session as a dedicated tool. That keeps the tool list small while leaving the whole catalogue reachable.
Actor runs consume Apify platform credit, priced per run or per result depending on the Actor. The MCP server itself adds nothing, but an agent left in a loop can spend real money — check an Actor’s pricing with fetch-actor-details before letting an agent call it repeatedly.
Firecrawl is one general-purpose engine that fetches and cleans any URL. Apify is a catalogue of purpose-built scrapers that already know the structure of a specific site. For arbitrary pages use Firecrawl; for Instagram profiles or Google Maps listings, a dedicated Actor returns structured fields Firecrawl would make you extract.
That depends on the target site’s terms, the data involved and your jurisdiction — the tool does not settle it. Apify publishes guidance on ethical scraping, and the responsibility for what an Actor is pointed at sits with whoever runs it.
Firecrawl
Scrape, crawl and search the live web as structured markdown.
Playwright
Give AI models real browser control — navigate, click, fill forms and screenshot any page.
Exa
Give an agent live web search and clean page reads — no API key to get started.
Brave Search
Give an AI agent live web and local search via the Brave Search API.