Noureddine RAMDI / Apify MCP Server: Enabling autonomous AI agent payments for web automation tools

Created Mon, 04 May 2026 10:09:30 +0000 Modified Sat, 23 May 2026 20:41:27 +0000
991 Words
github-stars typescript mcp ai-agents web-scraping agentic-payments nodejs

apify/apify-mcp-server

Apify MCP Server exposes over 8,000 Apify Actors—web scrapers, crawlers, and automation tools—as MCP (Model Context Protocol) tools accessible to AI agents. What sets this server apart is its integration of agentic payments, enabling AI agents to autonomously pay for running Actors using crypto or managed payment protocols, without the need for individual API tokens. This opens a concrete example of machine-to-machine commerce for AI-driven workflows.

how apify mcp server exposes actor tools to ai agents

The core functionality of the Apify MCP Server is to transform Apify Actors into MCP tools. Each Actor is a small program designed for scraping, crawling, or automation tasks. By wrapping these in MCP tools, AI agents that understand the MCP protocol can invoke these Actors dynamically.

The server supports two primary connection modes. The first is a hosted HTTPS endpoint at https://mcp.apify.com, which offers OAuth authentication and a streamable HTTP transport mechanism. This setup is recommended for most use cases and allows seamless integration with MCP clients like Claude.ai or VS Code extensions.

The second mode is standard input/output (stdio), usable locally via the command npx @apify/actors-mcp-server. This mode is well suited for local environments and command-line clients such as Claude for Desktop.

Under the hood, the server implements smart tool selection strategies. It dynamically adapts its exposed toolset based on the client’s capabilities. For example, clients supporting dynamic tool discovery get access to an add-actor tool that can instantiate new tools on the fly, whereas static clients receive a fixed call-actor tool. This design balances flexibility with compatibility.

Technically, the project is written in TypeScript and requires Node.js version 20 or higher. It integrates tightly with the Apify platform APIs, managing Actor execution, status, and results seamlessly.

agentic payments: autonomous crypto transactions for actor runs

What distinguishes Apify MCP Server is the implementation of agentic payments. This system enables AI agents to pay for Actor runs without needing direct API tokens, which simplifies security and usability.

The server supports two payment protocols:

  • x402: A protocol for USDC payments on the Base mainnet blockchain. Agents can use a wallet with USDC on Base to prepay for Actor executions.

  • Skyfire: A managed payment infrastructure supporting PAY tokens with a prepaid balance model. Agents can top up their balance and the server handles automatic refunds for unused funds.

This integration means AI agents can autonomously request tool executions and settle payments on-chain or through managed wallets, enabling real machine-to-machine commerce. The server handles wallet creation or import, transaction signing, and balance reconciliation.

This approach is rare and valuable in AI tooling ecosystems because it removes friction around API tokens and billing, allowing agents to transact securely and transparently.

quick start with apify mcp server

You can use the Apify MCP Server in two ways:

HTTPS Endpoint (mcp.apify.com): Connect from your MCP client via OAuth or by including the Authorization: Bearer <token> header in your requests. This is the recommended method for most use cases. Because it supports OAuth, you can connect from clients like Claude.ai or Visual Studio Code using just the URL: https://mcp.apify.com.

  • https://mcp.apify.com streamable transport

Standard Input/Output (stdio): Ideal for local integrations and command-line tools like the Claude for Desktop client.

  • Set the MCP client server command to npx @apify/actors-mcp-server and the APIFY_TOKEN environment variable to your Apify API token.
  • See npx @apify/actors-mcp-server --help for more options.

Prerequisites

  • A wallet with USDC on Base mainnet.

Setup

Create or import a wallet:


### Prerequisites

- A Skyfire account with a funded wallet.
- An MCP client that supports multiple servers, such as Claude Desktop, OpenCode, or VS Code.

### Setup

Configure the Skyfire MCP server and the Apify MCP Server in your client. Add `payment=skyfire` to the Apify server URL:

```json
{
  "mcpServers": {
    "skyfire": {
      "url": "https://api.skyfire.xyz/mcp/sse",
      "headers": {
        "skyfire-api-key": "<YOUR_SKYFIRE_API_KEY>"
      }
    },
    "apify": {
      "url": "https://mcp.apify.com?payment=skyfire"
    }
  }
}

See the Skyfire integration documentation for setup details. The Agentic Payments with Skyfire post provides additional background.

Local environment

  • Node.js (v20 or higher)

Create an environment file, .env, with the following content:

APIFY_TOKEN="your-apify-token"

Build the actors-mcp-server package:

npm run build

The server is also available on Docker Hub, making containerized deployment straightforward.

verdict: who should consider apify mcp server

This server is highly relevant for developers and organizations building AI agents that require autonomous access to a broad range of web scraping and automation tools. The agentic payments integration is a rare feature that reduces friction in machine-to-machine commerce, making it easier for agents to pay for tool usage securely and transparently.

The tradeoff is the complexity introduced by managing blockchain wallets or managed payment accounts, which may not be suitable for all projects. Additionally, the reliance on Node.js v20+ narrows the runtime environments somewhat.

Overall, the Apify MCP Server is a solid choice if you need dynamic, large-scale actor tooling integrated with autonomous payment capabilities, especially in AI-driven automation scenarios. The codebase is mature, the architecture is thoughtfully designed, and the quickstart setup is clear and straightforward to follow.

→ GitHub Repo: apify/apify-mcp-server ⭐ 1,181 · TypeScript