AI agents | Personnel.com

We publish our job listings through a Model Context Protocol (MCP) server, an open standard that lets AI assistants read data from a website directly instead of guessing or scraping. Connect it once and you can ask your assistant to find openings in plain language, and it will search our live listings and link you to the real posting.

It is free, needs no account, and requires no API key.

Server address

https://personnel.com/mcp

Claude (web, desktop and mobile)

Go to Customize → Connectors on claude.ai, click +, choose Add custom connector and paste the server address above. Connectors are saved to your account, so once it is added on the web it also appears in the Claude mobile app.

Available on Free, Pro, Max, Team and Enterprise plans; Free accounts can add one custom connector. On Team and Enterprise an Owner adds it under Organization settings → Connectors first, then members enable it under Customize → Connectors.

ChatGPT

Custom MCP connectors sit behind developer mode. Turn it on in Settings → Connectors → Advanced → Developer mode, then add a new connector and paste the server address. Available on Plus and Pro; on Business, Enterprise and Edu plans a workspace admin can restrict which connectors are allowed.

Grok

Go to grok.com/connectors, choose New Connector → Custom and enter the server address. No authentication step is needed. The Grok Build CLI picks up MCP servers you have already configured for Claude Code.

Claude Code

claude mcp add --transport http personnel https://personnel.com/mcp

Gemini CLI

Add this to ~/.gemini/settings.json, then check it with gemini mcp list:

{
  "mcpServers": {
    "personnel": {
      "httpUrl": "https://personnel.com/mcp"
    }
  }
}

Cursor

Add this to .cursor/mcp.json in your project, or your global Cursor MCP settings:

{
  "mcpServers": {
    "personnel": {
      "url": "https://personnel.com/mcp"
    }
  }
}

VS Code

Requires GitHub Copilot Chat in agent mode. Add this to .vscode/mcp.json — note VS Code uses servers rather than mcpServers, and needs an explicit transport type:

{
  "servers": {
    "personnel": {
      "type": "http",
      "url": "https://personnel.com/mcp"
    }
  }
}

Other MCP clients

Anything that supports remote MCP servers over Streamable HTTP will work, including Windsurf and the many other editors and agents adding support. Point the client at the server address above, choose HTTP transport rather than stdio, and leave authentication empty. Config key names differ between clients, so check yours if the examples above do not match. Clients that read our server card can discover the endpoint and its tools automatically.

Meta AI does not currently support connecting to custom MCP servers, so there is no setup for it yet. Our JSON API is open to any client in the meantime.

What you can ask for

Tool	What it does
search-jobs	Search open positions by keyword, category, or employer
get-job	Read one posting in full, including pay and location
search-companies	Find employers hiring through Personnel.com
get-company	Read one employer profile
list-job-categories	List job categories and how many openings each has

For example: “Find part-time office jobs in Boise and show me the pay ranges.” Search matches the words in a posting, so concrete terms like “warehouse” or “dental assistant” work better than broad descriptions.

What it will not do

These tools are read-only. They can find and read job postings, and nothing else.

No applying on your behalf. There is deliberately no tool that submits an application. Applying means handing over your own personal details, and that should be you doing it, knowingly. Your assistant will link you to the posting so you can apply yourself.
No candidate or resume data. Nothing about job seekers is reachable through these tools — no resumes, contact details, applications or messages. That information stays private and behind a login, and no AI assistant can read it.
No account access. Connecting this server does not sign you in and gives an assistant no access to your Personnel.com account.

For developers

The same listings are available as a plain JSON API at https://personnel.com/api/v1/jobs, documented by an OpenAPI 3.1 spec. Both the API and the MCP server are rate limited per IP address; the API allows 30 requests a minute and the MCP server 60. Also published: llms.txt, agent skills and auth.md.

Building something that needs higher limits or bulk access? Get in touch.