⚡ CareerLens MCP

Real-time career intelligence as an MCP server. Ask Claude if you're underpaid, find live job listings, benchmark your skills, and score your resume — all in one chat.

Live Demo →  |  Backend API Docs →

What It Does

CareerLens exposes 5 career intelligence tools via the Model Context Protocol (MCP). Claude decides which tools to call based on your question — you just chat.

Tool	What it answers
salary_benchmark	What should I be earning for this role?
job_search	What's actually hiring for my skills right now?
skill_demand	Is my tech stack trending up or dying?
resume_fit_score	How well does my resume match this job?
career_report	Full briefing: salary gap + live jobs + skill rankings

Bonus features:

• 📄 Upload your CV (PDF) → auto-extracts your profile
• 🔗 Paste your LinkedIn URL → auto-fills your profile

---

Architecture

React Frontend (Vercel)
        │
        │  POST /chat — SSE streaming
        ▼
FastAPI Bridge Server (Railway)
        │
        ├── Anthropic SDK (Claude Sonnet)
        │         │  tool_use (agentic loop)
        │         ▼
        └── tools/*.py ──► JSearch API (LinkedIn, Indeed, Glassdoor)
                         ──► Pure Python (resume ATS scoring)

Claude Desktop (optional)
        │  stdio / MCP protocol
        ▼
FastMCP Server ──► same tools/*.py

The MCP server and the API server share the same tool implementations in server/tools/. Zero duplication.

---

Tech Stack

Layer	Tech
MCP Server	Python + FastMCP
API Server	Python + FastAPI + SSE streaming
Job/Salary Data	JSearch API (LinkedIn, Indeed, Glassdoor)
AI	Claude Sonnet via Anthropic SDK
Frontend	React + Tailwind CSS
Backend Deploy	Railway
Frontend Deploy	Vercel

---

Quick Start

Prerequisites

• Python 3.11+
• Node.js 18+
• JSearch API key (free tier)
• Anthropic API key

1. Clone the repo

git clone https://github.com/yourusername/careerlens-mcp.git
cd careerlens-mcp

2. Backend setup

cd server
pip install -r requirements.txt
cp .env.example .env
# Edit .env and add your API keys
uvicorn api_server:app --reload --port 8000

3. Frontend setup

cd client
npm install
cp .env.example .env.local
# .env.local already points to localhost:8000
npm start

Open http://localhost:3000 — fill in your profile and start chatting.

4. Claude Desktop setup (optional)

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "careerlens": {
      "command": "python",
      "args": ["/absolute/path/to/careerlens-mcp/server/mcp_server.py"],
      "env": {
        "JSEARCH_API_KEY": "your_key",
        "ANTHROPIC_API_KEY": "your_key"
      }
    }
  }
}

---

Deploy Your Own

Backend → Railway

1. Create new project on railway.app
2. Connect your GitHub repo, set Root Directory to server
3. Add environment variables: JSEARCH_API_KEY, ANTHROPIC_API_KEY
4. Railway auto-deploys from Procfile

Frontend → Vercel

1. Import repo on vercel.com
2. Set Root Directory to client
3. Add environment variable: REACT_APP_API_URL=https://your-app.railway.app
4. Deploy

---

Project Structure

careerlens-mcp/
├── server/
│   ├── tools/
│   │   ├── __init__.py
│   │   ├── salary.py          # Tool 1: Salary benchmarks
│   │   ├── jobs.py            # Tool 2: Live job search
│   │   ├── skills.py          # Tool 3: Skill demand analysis
│   │   ├── resume.py          # Tool 4: ATS resume scoring
│   │   ├── report.py          # Tool 5: Full career report
│   │   ├── cv_parser.py       # CV PDF parsing
│   │   └── linkedin_parser.py # LinkedIn profile scraping
│   ├── api_server.py          # FastAPI + SSE bridge for frontend
│   ├── mcp_server.py          # FastMCP server for Claude Desktop
│   ├── config.py
│   ├── requirements.txt
│   ├── Procfile
│   └── railway.toml
├── client/
│   ├── src/
│   │   ├── App.js             # Full chat UI + result cards
│   │   └── App.css            # Dark bold theme
│   ├── vercel.json
│   └── package.json
├── .gitignore
└── README.md

---

Built With

Built in a weekend as a demonstration of the Model Context Protocol (MCP) — the open standard for connecting AI models to external tools and data sources.

---

License

MIT — do whatever you want with it.