Job Search Meets AI: How to Use MCP Servers to Find Remote Gigs

You're coding in Cursor. You've got Claude helping you debug. And suddenly you think: "I wonder if there are any Rust jobs posted this week."

Normally, you'd alt-tab to a browser, search a job board, scroll, open tabs. Five minutes later, you've lost your flow.

But what if you could just ask your AI assistant — right in your editor — and get structured job results back?

That's what MCP servers enable. And we're building one for HireQL.

What is MCP (Model Context Protocol)?

MCP is Anthropic's open protocol for connecting AI assistants to external tools and data sources. Think of it like a plugin system for Claude (and other LLMs).

Instead of teaching an AI how to browse the web or query a database, you give it access to a server that knows how to do those things. The AI can then call that server's tools as needed.

How It Works

1. You install an MCP server (a lightweight process that exposes "tools")
2. Your AI client connects to it (Claude Desktop, Cursor, etc.)
3. The AI can now call those tools when answering your questions

For example, an MCP server for job search might expose tools like:

• search_jobs(tech, location, vibe)
• get_job_details(job_id)
• get_new_jobs_since(date)

When you ask Claude "find me remote TypeScript jobs with async culture," it knows to call search_jobs with the right parameters.

Why MCP for Job Search?

Job hunting is context-switching hell. You're in the zone, coding, learning, building — and then you have to stop, open a browser, and manually search.

With an MCP server, job search becomes ambient. You can:

• Ask your AI "what remote jobs match my skills?" without leaving your editor
• Get daily job summaries in your coding chat
• Have Claude compare job descriptions while you're reviewing them
• Automate application tracking workflows

It's not about replacing your job search — it's about making it frictionless.

Setting Up the HireQL MCP Server

(Note: This is in development. Instructions will be updated when the server is public. For now, here's how it'll work.)

1. Install the Server

npm install -g @hireql/mcp-server
# or
npx @hireql/mcp-server

2. Configure Your API Key

Grab your API key from the developer portal, then set it in your environment:

export HIREQL_API_KEY="your_key_here"

3. Add to Claude Desktop

In Claude Desktop's settings, add the MCP server:

{
  "mcpServers": {
    "hireql": {
      "command": "npx",
      "args": ["@hireql/mcp-server"],
      "env": {
        "HIREQL_API_KEY": "your_key_here"
      }
    }
  }
}

4. Restart Claude Desktop

That's it. Now Claude can search jobs.

Using the MCP Server

Once connected, you can ask Claude things like:

"Find me remote senior engineer jobs posted in the last 3 days"

Claude will call the search_jobs tool behind the scenes and return results:

Found 12 jobs matching your criteria:

1. Senior Full-Stack Engineer at Vibe Corp
   - Remote (US)
   - $160k-$200k
   - Tech: TypeScript, React, PostgreSQL
   - Vibe: Async-first, no meetings culture
   - https://remotevibecodingjobs.com/jobs/abc123

2. Principal Engineer at RemoteFirst Inc
   - Remote (Worldwide)
   - €140k-€180k
   - Tech: Rust, Kubernetes, gRPC
   - Vibe: 4-day week, transparent salary
   - https://remotevibecodingjobs.com/jobs/def456

...

You can drill down:

"Tell me more about the Vibe Corp role"

And Claude will fetch the full job description, benefits, interview process — all without you leaving the chat.

Advanced Workflows

Daily Job Digest

Set up a cron job that asks Claude to summarize new jobs every morning:

#!/bin/bash
# daily-jobs.sh

echo "What remote jobs were posted yesterday? Focus on Rust, Go, or TypeScript roles with transparent salaries." | claude-cli --mcp hireql

Application Tracking

Combine HireQL with a filesystem MCP server to auto-log applications:

"I just applied to the Vibe Corp job. Log it in my applications.md file."

Claude uses HireQL to get job details, then writes them to your tracking doc.

Skill Gap Analysis

"Find 10 remote senior engineer jobs and tell me what skills I'm missing based on my resume."

Claude searches jobs, compares requirements to your resume (via another MCP tool or context), and gives you a gap analysis.

Building Your Own MCP Server

Want to build an MCP server for something else? The protocol is open. Here's a minimal example in TypeScript:

import { McpServer } from '@anthropic-ai/mcp';

const server = new McpServer({
  name: 'hireql',
  version: '1.0.0'
});

server.tool({
  name: 'search_jobs',
  description: 'Search remote coding jobs by tech stack and vibe',
  parameters: {
    tech: { type: 'array', items: { type: 'string' } },
    vibe_tags: { type: 'array', items: { type: 'string' } },
    limit: { type: 'number', default: 10 }
  },
  async execute({ tech, vibe_tags, limit }) {
    const response = await fetch('https://remotevibecodingjobs.com/api/v1/jobs', {
      method: 'POST',
      headers: {
        'Authorization': `Bearer ${process.env.HIREQL_API_KEY}`,
        'Content-Type': 'application/json'
      },
      body: JSON.stringify({ tech, vibe_tags, limit })
    });
    
    return await response.json();
  }
});

server.start();

That's it. The server exposes a search_jobs tool. Claude can now call it.

MCP Beyond Job Search

MCP isn't just for jobs. People are building servers for:

• Filesystems — let Claude read/write files
• Databases — query Postgres, MongoDB, etc.
• APIs — GitHub, Slack, Linear, you name it
• Search engines — Brave, Google, custom indexes

It's a plugin ecosystem for AI. And job search is a perfect fit.

When Will the HireQL MCP Server Launch?

Soon. We're testing it internally now. If you want early access, sign up for an API key and we'll ping you when it's ready.

Final Thoughts

AI assistants are getting good at code. But they're still blind to the outside world unless you connect them to it.

MCP is the bridge. And job search — especially for remote, vibe-aligned work — is one of the best use cases.

Imagine never alt-tabbing to Indeed again. Imagine your AI just knowing what jobs fit you.

That's where we're headed.

Get your API key and stay tuned →