Apify MCP Server

Content

Apify MCP Server is the official Model Context Protocol server for Apify. It lets Claude and other MCP clients discover Apify Actors, run scraping and web automation workflows, inspect Actor runs, and retrieve dataset or key-value store data from the Apify platform.

The server supports a hosted Streamable HTTP endpoint with OAuth or bearer-token authentication, plus a local stdio package for clients that run MCP servers locally. The README recommends explicitly configuring tools for production use because default tool loading may change across versions.

Source Review

• https://github.com/apify/apify-mcp-server
• https://docs.apify.com/platform/integrations/mcp
• https://registry.npmjs.org/%40apify%2Factors-mcp-server
• https://github.com/apify/apify-mcp-server/blob/master/README.md
• https://github.com/apify/apify-mcp-server/blob/master/server.json
• https://github.com/apify/apify-mcp-server/blob/master/package.json
• https://github.com/apify/apify-mcp-server/blob/master/LICENSE.md

These sources were reviewed on 2026-06-06. Prefer the live Apify docs, repository, MCP server metadata, npm package metadata, and license for current setup commands, supported transports, tool configuration, payment behavior, and telemetry controls.

Features

• Official Apify MCP server from apify/apify-mcp-server.
• Hosted Streamable HTTP server with OAuth or bearer-token authentication.
• Local stdio server via npx @apify/actors-mcp-server.
• Actor discovery with search-actors and fetch-actor-details.
• Actor execution through call-actor and configured Actor-specific tools.
• Default access to Apify documentation search and fetch tools.
• Run inspection, run-log retrieval, and optional run abort tools.
• Dataset metadata, schema, pagination, and item retrieval tools.
• Key-value store metadata, key listing, and record retrieval tools.
• Optional add-actor workflow for adding Actors as tools.
• Tool-selection controls for limiting the exposed Actor, docs, storage, run, or experimental tools.
• Hosted support for output schema inference, MCP Apps UI mode, and agentic payments.
• Telemetry opt-out through CLI flag, environment variable, or hosted URL parameter.

Installation

For local stdio usage, configure an MCP client to run the official npm package and provide an Apify API token in the client environment:

{
  "mcpServers": {
    "apify": {
      "command": "npx",
      "args": [
        "-y",
        "@apify/actors-mcp-server",
        "--tools",
        "actors,docs,apify/rag-web-browser"
      ],
      "env": {
        "APIFY_TOKEN": "<your-apify-token>"
      }
    }
  }
}

For hosted usage, configure the Streamable HTTP endpoint through the Apify MCP configuration page and authenticate with OAuth or an Apify bearer token. Avoid SSE configurations for new setups because Apify documents a transition to Streamable HTTP.

Use Cases

• Discover a suitable Apify Actor for a scraping or automation task.
• Inspect an Actor's input schema before allowing Claude to run it.
• Run the RAG Web Browser Actor for research workflows.
• Retrieve dataset items after an Actor run finishes.
• Inspect Actor run metadata and logs during debugging.
• Fetch Apify documentation into the conversation for setup or troubleshooting.
• Limit a production client to a single approved Actor tool.
• Use hosted output schema inference for structured Actor results.
• Abort an Actor run when the model launched the wrong workflow or parameters.

Safety and Privacy

Treat Apify MCP as a live automation and data-extraction control surface. Actor runs can cost money, access third-party websites, trigger anti-abuse systems, and collect regulated or personal data. Require human approval before launching Actors, adding new Actor tools, enabling payment modes, or retrieving sensitive datasets.

Use the narrowest tools configuration that fits the workflow. Keep Apify API tokens, OAuth credentials, payment credentials, dataset IDs, and run identifiers out of prompts and logs. Review telemetry settings before using local stdio with sensitive prompts or Actor inputs, and confirm that scraping targets and data retention practices are allowed for the project.

Duplicate Check

Existing content includes a general Apify platform listing in content/tools, but no Apify-specific MCP entry was found in content/mcp. This submission is scoped to apify/apify-mcp-server, the official MCP server, and does not change the existing tool listing.

• Source repositorygithub.com 2026-07-21T07:17:01+00:00
• Documentationdocs.apify.com
• Submitted by oktofeesh12026-06-06