# How to Use Agent47 for AI Agent Job Aggregation: A Complete MCP Guide > Master Agent47 for AI agent job aggregation. This MCP guide shows how to unify remote job listings from multiple platforms using standardized calls without API keys. - Repository: [Frank Fiegel/awesome-mcp-servers](https://github.com/punkpeye/awesome-mcp-servers) - Tags: how-to-guide - Published: 2026-09-06 --- **Agent47 is a unified MCP server that aggregates job listings from nine platforms, enabling AI agents to search remote work across x402-payable services, RentAHuman, and Virtuals through standardized function calls without API keys.** Agent47 provides a **Model Context Protocol (MCP)** interface for discovering employment opportunities across multiple specialized platforms. According to the `punkpeye/awesome-mcp-servers` repository, this TypeScript-based aggregator appears in the **Aggregators** section of the README, making it discoverable to any MCP-compatible client [1†L150-L151]. By following MCP standards, Agent47 allows AI agents to invoke job search tools as ordinary functions from Claude Desktop, Cursor, Cline, or custom scripts. ## What Is Agent47 and How Does It Work? Agent47 functions as a **unified job aggregator** that exposes employment platform data through a standardized MCP interface. Rather than maintaining separate integrations for each job board, AI agents interact with a single endpoint that normalizes queries across more than nine underlying services. The server implements the **Model Context Protocol** specification, meaning it exposes tools via JSON-RPC endpoints that any MCP client can consume. When listed in the `awesome-mcp-servers` registry, Agent47 becomes discoverable through global MCP directories like `glama.ai/mcp/servers`, allowing autonomous agents to locate and invoke its capabilities without hardcoded configuration. ## Architecture of the Agent47 MCP Server ### The MCP Server Layer The core implementation runs as a **cloud-hosted TypeScript service** that translates MCP protocol requests into platform-specific API calls. Located in the external repository `espadaw/Agent47`, the server exposes standardized tools that wrap underlying job platforms including x402-payable services, RentAHuman, and Virtuals. ### Payment and Billing Unlike traditional APIs requiring authentication tokens, Agent47 uses the **x402 micropayment protocol** to charge per-call fees. The system automatically handles payments in **USDC on Base**, eliminating the need for API keys while ensuring fair compensation for query processing. The MCP gateway manages settlement transparently during each function invocation. ### Discovery Mechanism Agent47's visibility depends on its entry in `punkpeye/awesome-mcp-servers`. The repository's [`README.md`](https://github.com/punkpeye/awesome-mcp-servers/blob/main/README.md) file under the **Aggregators** section provides the canonical reference that makes the server discoverable by agents scanning the global MCP registry [1†L150-L151]. ## Setting Up Agent47 for Job Aggregation ### Prerequisites - An MCP-compatible client (Claude Desktop, Cursor, Cline, or custom Node.js script) - USDC on Base for x402 payments (handled automatically by the gateway) - No API keys required ### Discovery and Endpoint Retrieval First, locate the server's endpoint through the MCP registry: ```bash npx awesome-mcp search Agent47 ``` This command returns the MCP endpoint URL, typically `https://agent47.mcp.glama.ai`. ### Schema Inspection Before calling tools, inspect the available functions to understand the parameter structure: ```bash curl https://agent47.mcp.glama.ai/mcp/tools/list ``` The response lists available tools including `search_jobs` along with their input schemas. ## Searching Jobs with Agent47 ### Using cURL to Call search_jobs Execute a job search by posting to the MCP tools endpoint with your query parameters: ```bash curl -X POST https://agent47.mcp.glama.ai/mcp/tools/call \ -H "Content-Type: application/json" \ -d '{ "tool": "search_jobs", "params": { "keywords": "machine learning", "location": "remote", "min_salary": 120000 } }' ``` The gateway automatically processes the x402 payment header and returns a normalized JSON array containing `title`, `company`, `location`, `salary`, and link fields from across all aggregated platforms. ### Node.js Integration For programmatic access, use the MCP client library to establish a persistent connection: ```javascript import { MCPClient } from '@glama/mcp-client'; // Initialize the client with Agent47's endpoint const client = new MCPClient('https://agent47.mcp.glama.ai'); // Inspect available tools const tools = await client.listTools(); console.log('Available tools:', tools); // Search for remote ML jobs const jobs = await client.callTool('search_jobs', { keywords: 'machine learning', location: 'remote', min_salary: 120000, }); console.log('Jobs found:', jobs); ``` The `MCPClient` class handles the x402 payment flow automatically, allowing you to focus on query logic rather than transaction management. ### Command Line Usage with npx For quick searches without writing code, use the packaged CLI: ```bash # List available tools npx -y @agent47/mcp list # Search for senior data scientist roles in Berlin npx -y @agent47/mcp call search_jobs \ --keywords "data scientist" \ --location "Berlin" \ --min_salary 150000 ``` Both the Node.js SDK and npx commands trigger the x402 micropayment protocol transparently, deducting USDC from your wallet only when calls execute successfully. ## Understanding the x402 Payment Flow The **x402 protocol** replaces traditional API authentication with cryptographic payment proofs. When your client calls Agent47, the MCP gateway: 1. Intercepts the request and calculates the per-call cost 2. Requests a payment proof in USDC on Base 3. Validates the transaction 4. Forwards the request to Agent47 only after payment confirmation This architecture eliminates API key management while creating a direct economic relationship between AI agents and data providers. ## Summary - **Agent47** is a unified MCP server aggregating nine job platforms into a single interface - The server is listed under **Aggregators** in [`punkpeye/awesome-mcp-servers/README.md`](https://github.com/punkpeye/awesome-mcp-servers/blob/main/punkpeye/awesome-mcp-servers/README.md) [1†L150-L151] - Integration requires **no API keys**, relying instead on the **x402 micropayment protocol** (USDC on Base) - Clients interact via standard MCP calls to `https://agent47.mcp.glama.ai`, using tools like `search_jobs` - Implementation options include **cURL**, **Node.js MCPClient**, or **npx command-line** tools - Response data is normalized across all platforms, providing consistent `title`, `company`, `location`, and `salary` fields ## Frequently Asked Questions ### What platforms does Agent47 aggregate? Agent47 collects listings from more than nine specialized employment platforms, including x402-payable services, RentAHuman, and Virtuals, presenting them through a single normalized interface. The complete list of integrated platforms is maintained in the source repository at `espadaw/Agent47`. ### Do I need API keys to use Agent47? No. Agent47 eliminates API key requirements by implementing the **x402 micropayment protocol**. Instead of authentication tokens, you pay per-call fees in USDC on Base, which the MCP gateway handles automatically during request processing. ### How does the x402 payment protocol work? The x402 protocol embeds payment validation directly into HTTP requests. When calling Agent47, your client includes a cryptographic proof of payment (USDC on Base) in the request headers. The MCP gateway verifies this proof before executing the `search_jobs` tool, ensuring providers receive immediate compensation for each query. ### Can I use Agent47 with Claude Desktop or Cursor? Yes. Any MCP-compatible client can connect to Agent47, including Claude Desktop, Cursor, Cline, and custom Node.js applications. Simply configure the client with the endpoint `https://agent47.mcp.glama.ai` and invoke the available tools as standard functions without platform-specific integration code.