AI Job Agent

A Chrome extension + self-hosted backend that captures job postings, scores them against your resumes, finds the recruiter's email, and writes a personalised cold outreach email — in under 30 seconds.

1. Visit any job posting on LinkedIn, Greenhouse, Lever, Workday, Indeed, or any job board
2. Click the extension — it reads the page automatically
3. Click "Capture & Process" — in 15–30 seconds you get:
   • Your best-matching resume selected from all your uploaded resumes
   • A match score (0–100%) with matched and missing skills
   • The recruiter or hiring manager's email via Hunter.io
   • A personalised cold outreach email ready to copy and send
   • Day-7 and Day-14 follow-up emails scheduled automatically
4. Review and send — nothing is sent without your approval

Recommended: set both GROQ_API_KEY (free, no card) and OPENAI_API_KEY (embeddings only). Costs essentially nothing.

Deploy the backend to Render's free tier in about 10 minutes.

1. Fork this repository to your GitHub account
2. Go to render.com and sign up (free)
3. Click New → Web Service → connect your forked repo
4. Set Runtime to Python, leave build/start commands as-is
5. Add a free PostgreSQL database: New → PostgreSQL → copy the Internal Database URL
6. Set environment variables (see Step 2)
7. Click Deploy

Go to your web service → Environment tab:

Once deployed, open https://your-app.onrender.com/dashboard and click ⬆ Upload Resume in the top-right. Upload all your resume versions (general, backend-focused, AI-focused, etc.).

1. Download or clone this repo to your computer
2. Open Chrome → chrome://extensions
3. Enable Developer mode (top-right toggle)
4. Click Load unpacked → select the extension/ folder
5. Pin the extension to your toolbar
6. Click the extension icon → paste your Render URL into the Backend field → Save

You're ready. Visit any job posting and click Capture & Process.

Used to embed resumes and job descriptions for similarity matching. Required even if you use Groq for the LLM — Groq has no embedding models.

1. Go to platform.openai.com/api-keys → create a key
2. Add $5 credit — will last years at 30 jobs/day (embeddings cost ~$0.01/month)
3. Set OPENAI_API_KEY

Runs Llama-3 for job parsing and outreach generation at no cost, no credit card required. When set, OpenAI is used only for embeddings.

1. Go to console.groq.com → sign up → create a key
2. Set GROQ_API_KEY

Routing logic:

• Both keys set → Groq for LLM, OpenAI for embeddings (recommended)
• Only OpenAI → OpenAI handles everything (~$6.80/month at 30 jobs/day)
• Only Groq → won't work (no embeddings)

# 1. Clone
git clone https://github.com/YOUR_USERNAME/ai-job-agent.git
cd ai-job-agent

# 2. Virtual environment
python -m venv .venv
source .venv/bin/activate   # Windows: .venv\Scripts\activate

# 3. Install
pip install -r requirements.txt

# 4. Configure
cp .env.example .env
# Edit .env — add OPENAI_API_KEY and optionally GROQ_API_KEY

# 5. Run
uvicorn backend.main:app --host 0.0.0.0 --port 8000

Then open http://localhost:8000/dashboard to upload resumes, and load the extension pointing at http://localhost:8000.

1. Navigate to any job posting
2. Click the AI Job Agent icon in your toolbar
3. Wait for "Ready to process" (green status)
4. Click Capture & Process
5. In 15–30 seconds you'll see:
   • Match score + which resume was selected
   • Matched and missing skills
   • Outreach email — click "View email" to expand and copy
   • Contacts found with LinkedIn links
6. Click Mark as Applied after submitting — starts the Day-7 / Day-14 follow-up countdown

Click ⬆ Upload resume PDF at the bottom of the extension popup to upload directly without opening the dashboard.

If you try to process a job you've already captured, the extension shows "Already captured — Application #N" instead of running the pipeline again.

Open https://your-app.onrender.com/dashboard to track all applications:

• Upload resumes via the ⬆ Upload Resume button
• Update status via dropdown (captured → applied → interviewing → offer / rejected)
• Filter by status with the chip strip
• Export full history as CSV
• View skill gaps — skills appearing most in jobs you're missing

Chrome Extension (popup.js + content.js)
         │
         │  POST /api/jobs/process
         ▼
   FastAPI Backend
         │
         ├─ job_parser.py        → Groq Llama-3.3-70B or GPT-4o
         ├─ resume_matcher.py    → Hybrid score: 0.4×embedding + 0.6×keyword overlap
         ├─ embeddings.py        → OpenAI text-embedding-3-small
         ├─ outreach_generator   → Groq Llama-3.1-8B or GPT-4o-mini
         ├─ recruiter_finder     → Hunter.io: find + verify recruiter email
         └─ database.py          → Postgres (Render) / SQLite (local)

Hybrid scoring formula:

hybrid = 0.4 × norm_embedding + 0.6 × keyword_overlap
norm_embedding = clamp((cosine − 0.20) / 0.50, 0, 1)

Full interactive docs at https://your-app.onrender.com/docs.

LinkedIn · Greenhouse · Lever · Workday · Indeed · Ashby · any site with <main> or <article> content

See CONTRIBUTING.md for setup instructions.

Good first issues:

• Add more job board selectors to extension/content.js
• Add more skills to the vocabulary in backend/agents/resume_matcher.py
• Write integration tests for the /api/jobs/process pipeline

• Resume text and embeddings are stored in your own database — not shared with anyone
• Job posting text is sent to OpenAI or Groq for parsing (subject to their data policies)
• Only the company domain is sent to Hunter.io — not your personal data

MIT — free to use, modify, and distribute.

Built with FastAPI · OpenAI · Groq · Hunter.io · Chrome Extensions API