Search Rolev.ai jobs from your AI assistant, autonomous agent, or browser.

Connect in Claude, your IDE, autonomous agent, or browser

Rolev.ai offers three open, standardized protocols so you can search live employer jobs from anywhere without API key friction or brittle web scraping.

What to ask

Ask in plain words for roles, cities, states, remote work, or specific employers. Follow up on any listing in the same conversation.

• Role + RemotePrompt
  “Find remote software developer jobs.”
  Ten of the newest, with pay where the posting states it, and a link to the full list of a few hundred.
• Role + StatePrompt
  “Registered nurse jobs in Texas, newest first.”
  The jobs on the page that are in Texas, plus the Texas cities with their own lists.
• By EmployerPrompt
  “What's open at [employer] right now?”
  That employer's newest openings on Rolev.ai, if it is one of the employers listed.
• Sector CatalogPrompt
  “Which sales roles does Rolev cover, and how many are open?”
  The sales catalog with live counts and a direct link for each role.
• Deep DivePrompt
  “Tell me more about the second one.”
  The posting's text, pay, place and type, with the Rolev.ai page and the employer's own posting.

What it does, and does not

Rolev.ai's MCP connector is scoped strictly to read-only job discovery. Your private career data stays inside Rolev.ai, and every application stays in your hands.

For developers

A Model Context Protocol server, Streamable HTTP, stateless, protected by OAuth 2.1 with PKCE and dynamic client registration (discovery at /.well-known/oauth-authorization-server). GET the address for a short description; POST JSON-RPC to it with a bearer token. Thirty requests a minute per account.

POST https://rolev.ai/a2a
Content-Type: application/json
Authorization: Bearer <token>

{
  "jsonrpc": "2.0",
  "id": "req-42",
  "method": "SendMessage",
  "params": {
    "message": {
      "parts": [
        { "text": "registered nurse jobs in Austin, TX" }
      ]
    }
  }
}

const tools = await document.modelContext.getTools();
const search = tools.find((t) => t.name === "search_jobs");
await document.modelContext.executeTool(search, {
  role: "registered nurse",
  place: "Austin, TX",
});

Frequently asked questions

Do I need a Rolev.ai account?

Yes. When you add the connector, your assistant sends you to Rolev.ai to sign in and allow it; a free account is enough. The connection is listed under Profile, where you can disconnect it at any time.

Does the assistant see my resume?

No. The connection covers job search only: the assistant reads public job postings as you. Your resume, saved answers and applications are not reachable through it.

Can it apply for me?

No. It finds jobs and links to their pages on Rolev.ai. There, you tailor your resume, the Chrome extension fills the employer's form, and you press Submit.

Which assistants work?

Claude, through Settings → Connectors → Add custom connector. Claude Code, Cursor, VS Code and other MCP clients that support OAuth connect to the same address. ChatGPT is not available yet; a listing is in progress.

Is there a limit?

Thirty requests a minute per account, shared by the MCP connector, the A2A agent and the WebMCP tools, which is plenty for a conversation. Searching is free; applying uses credits as usual.

What is the A2A (Agent-to-Agent) protocol?

A2A 1.0 (Agent2Agent) is an open standard allowing autonomous AI agents (such as CrewAI, LangGraph, AutoGPT, or custom agent systems) to delegate tasks to other agents. Rolev.ai provides an official A2A endpoint at https://rolev.ai/a2a with Agent Card discovery at https://rolev.ai/.well-known/agent-card.json.

How does A2A differ from MCP?

MCP (Model Context Protocol) is designed primarily for interactive human-in-the-loop chat clients (Claude, Cursor, VS Code) providing tools to the LLM context. A2A (Agent-to-Agent) is designed for agent-to-agent communication, where one autonomous agent sends JSON-RPC messages and gets each answer back at once as text and structured data.

What is WebMCP?

WebMCP is a draft browser standard that lets a web page hand tools to an AI agent working in the browser. Every Rolev.ai page registers the same three tools (search_jobs, get_job, list_roles), so a browser assistant can search jobs for you directly instead of reading and clicking through pages. It works with your own Rolev.ai sign-in, in browsers that support WebMCP; other browsers are unaffected.