Skip to content
MCP Shelf

Find and run ready-made Apify scrapers for maps, social media, shops and search.

  • Needs your setup first
  • ★ 13.3K GitHub stars
  • Last commit 26 Oct 2026

At a glance

Apify's server turns Apify Store into a tool box. The assistant searches the Store for an Actor (Apify's name for a packaged scraper or automation), reads its input schema and runs it, then pages through the results. Out of the box it also has a web search and a page fetcher.

Runs are billed to your Apify account like any other run, and some Actors charge per result, so set a spending limit before you let an agent loose. The hosted server at mcp.apify.com signs in with OAuth and gets new features first.

Made by
Apify
Licence
MIT
Latest version
0.19.3 (21 Oct 2026)
GitHub stars
13.3K
Last commit
26 Oct 2026
Tools
10
Runs on
Your computer or hosted

Our check: It needs something our Linux test machine does not have (an account, a desktop app, a Mac or a running service), so on 27 Oct 2026 we listed its tools from its package instead of running this config.

What's in the pod: its tools

We found 10 tools when we ran its package, before any sign-in or app was connected, on 27 Oct 2026. 7 only read. The rest can change things, so keep your app's "ask before each tool call" setting on for them.

How we tested it: Started @apify/actors-mcp-server 0.19.3 with a placeholder token. It lists its helper tools; the two Actor tools it loads by default (RAG Web Browser and Web Fetch) need a real Apify token, so they are not in this list.

Changes things3 tools · Can create, edit, send or delete.

ToolWhat it is for
abort-actor-runStops a run that is still going.
call-actorRuns an Actor with the given input and returns run details plus a short summary.
report-problemSends a problem report about a tool or Actor to Apify.

Only reads7 tools · Looks things up and changes nothing.

ToolWhat it is for
fetch-actor-detailsReturns an Actor's description, input schema, pricing and output schema, so the model can fill in the input correctly.
fetch-apify-docsFetches one documentation page in full.
get-actor-runChecks the status and storage IDs of a run.
get-dataset-itemsPages through the rows an Actor run produced.
get-key-value-store-recordReads one stored record, such as a screenshot or an output file, by key.
search-actorsSearches Apify Store for scrapers and automations that fit a site or a job.
search-apify-docsSearches the Apify and Crawlee documentation.

Tool groups

actors
search-actors, fetch-actor-details and call-actor: find and run any Actor. On by default.
docs
Search and read the Apify and Crawlee docs. On by default.
apify/rag-web-browser, apify/web-fetch
Two Actors loaded as direct tools by default: a web search that reads the top results, and a single-page fetcher.
runs
List runs, read their details and logs, and abort them.
storage
Datasets and key-value stores: list them, read schemas and records.
tasks
Create, update and publish saved Actor tasks.
schedules
Create, change and delete schedules that run Actors on a timer.
builds, versions
Build Actors, read build logs, and create or change an Actor's source files.
api
Search the Apify API reference and send read or write requests to any API path.

Grouped and described by us. The README on GitHub lists the inputs each tool takes.

How clearly it describes its tools to your assistant

Tools with a description
10 of 10
Say if they change things
10 of 10
Context it takes, at least
~2,700 tokens

Every tool says whether it only reads or can change things, so a client can ask before the risky ones.

From the tool list it sent us. The token figure counts names, descriptions and input names; the full schemas your client loads add more.

Set it up

Pick your app for the exact snippet and where it goes.

Where it goes: claude_desktop_config.json (Settings → Developer → Edit Config)

Claude Desktop
{
  "mcpServers": {
    "apify": {
      "command": "npx",
      "args": [
        "-y",
        "@apify/actors-mcp-server"
      ],
      "env": {
        "APIFY_TOKEN": "YOUR_APIFY_TOKEN"
      }
    }
  }
}

Add "--tools", "actors,docs" to pin the tool set, since Apify says the defaults can change. Set "TELEMETRY_ENABLED": "false" in env to switch off its telemetry.

Remote option

Apify recommends the hosted server: it signs in with OAuth, needs no install and can use rental Actors, which the local server cannot.

Server URL
https://mcp.apify.com

New to this? See the setup guides for Claude Code, Cursor, Claude Desktop and ChatGPT.

Settings

Names are exactly as the project spells them. Required ones are marked.

SettingWhat it does
APIFY_TOKENEnvironment variable · requiredYour Apify API token, from Apify Console → Settings → API & Integrations. Runs are billed to this account.apify_api_...
--toolsCommand line flagPicks tool groups, single tools or Actors to load. Without it you get the defaults, which Apify says may change, so pin it in anything you rely on.--tools actors,docs,apify/rag-web-browser
?tools=ConfigThe same choice for the hosted server, as a query string on the URL.https://mcp.apify.com?tools=actors,docs
--uiCommand line flagReturns interactive MCP App widgets for some tools in clients that can show them. Also UI_MODE in env.--ui true
TELEMETRY_ENABLEDEnvironment variableSet to false to turn off usage telemetry and Sentry error reporting. --telemetry-enabled=false does the same.false

What it can touch

It can read

  • Apify Store listings and Actor details
  • Your runs, datasets and key-value stores
  • Any public website an Actor scrapes

It can change

  • Your Apify balance: each Actor run costs platform credits, and some Actors charge per result
  • Actor tasks, schedules, builds and source files, when those tool groups are loaded

Safer setup: Load only --tools actors,docs, set a monthly spending limit in Apify Console, and leave the api, versions and schedules groups off unless you need them.

Before you start

  • An Apify account and API token (free plan available)
  • Node.js 22 or newer for the local server

If it does not work

A client still uses https://mcp.apify.com/sse
The SSE endpoint was removed. Point the client at https://mcp.apify.com with no /sse on the end.
The model cannot find a rental Actor
Rental Actors only work through the hosted server. Switch from npx to https://mcp.apify.com.
A run finishes but the model shows no data
call-actor returns a summary, not the rows. Ask it to fetch the results with get-dataset-items.
Unexpected charges
Check the Actor's pricing with fetch-actor-details before running it, and keep a spending limit on the account.

Peas in a pod: similar servers

How Apify lines up against other browser and web data servers on the shelf.

ServerToolsRuns onLicenceStars
Apify10Your computer or hostedMIT13.3K
BrowserTools MCP16Your computerMIT7.3K
ExecuteAutomation Playwright MCP33Your computerMIT5.7K
Wigolo10Your computerAGPL-3.0-only5.5K
Perplexity MCP4Your computer or hostedMIT2.6K

Questions

Is the Apify MCP server free?

The server is free and open source (MIT). Running Actors uses your Apify credits; the free plan includes a small monthly amount.

Hosted or local?

Apify recommends the hosted server: OAuth sign-in, no install, newer features and rental Actors. The local npx server suits clients that only start commands.

Can it run without an account?

The hosted server lets you search Actors and the docs without a token. Running anything needs a token, or one of the agent payment options Apify documents.

Which sites can it scrape?

Whatever the Actors in Apify Store cover, which includes Google Maps and Search, Instagram, TikTok, Amazon and general websites.

Full docs on GitHubapify/apify-mcp-server

Something wrong on this page? Tell us.

More in Browser and web data