Remote Jobs MCP Server

What the Jobicy/remote-jobs-mcp-server MCP server does

Jobicy/remote-jobs-mcp-server MCP server exposes Jobicy’s remote job data to MCP-compatible AI clients. It is intended for searches across remote vacancies from companies worldwide and covers more than 20 professional industries, according to the project documentation.

The server returns structured job listings rather than requiring a client to parse a web page. An agent can request a chosen number of results and narrow the search with a geographic region, an industry, or a keyword tag. This makes it suitable for job discovery workflows where the client needs current listings and machine-readable filters.

How it works

The project supports three connection methods. The recommended option is the hosted Streamable HTTP endpoint at https://jobicy.com/mcp. Older MCP clients that require server-sent events can use https://jobicy.com/mcp/sse with SSE transport enabled.

A local mode is also available. After cloning the repository and installing its npm dependencies, a client starts server.js with the --stdio argument. The local configuration must point to the actual path of that file on the machine running the client.

The hosted and local modes expose the same documented MCP tools. No API key or other credential is specified in the supplied setup instructions.

Setup and configuration

For a hosted connection, add the endpoint URL to the MCP client configuration. A modern client can use:

{
  "mcpServers": {
    "jobicy-jobs": {
      "url": "https://jobicy.com/mcp"
    }
  }
}

For a client that only supports legacy SSE, configure the /mcp/sse URL and set the transport type to sse.

To run Jobicy/remote-jobs-mcp-server MCP server locally, clone the repository, run npm install, and configure the client to execute node with the path to server.js followed by --stdio. The example path in the project documentation is a placeholder, so it must be replaced with the real cloned-directory path.

Tools and capabilities

• get_jobs fetches a structured list of remote jobs.
• get_jobs accepts a result count and supports the geo, industry, and tag filters.
• get_taxonomies returns valid slugs for location and industry filtering.
• The locations taxonomy supplies region filter values.
• The industries taxonomy supplies category filter values.

Using get_taxonomies before get_jobs can help an agent select valid filter slugs instead of guessing them. The available interface is focused on reading and filtering job listings; the supplied material does not describe job applications, saved searches, alerts, or employer-side actions.

Limitations and notes

The server is specific to Jobicy’s remote job database. Results and taxonomy values come from that service, and the documentation does not describe an offline dataset or a way to supply another job source. The project documents real-time access to listings, but it does not specify freshness guarantees, pagination behavior, ranking rules, or rate limits.

The repository is licensed under MIT. Claude, Cursor, and Cline are listed as compatible client examples. Other MCP-compatible clients may work when they support one of the documented transports, but the supplied material does not provide a broader client compatibility matrix.