Skip to content

README.md

Latest commit

 

History

History
260 lines (189 loc) · 7.02 KB

README.md

File metadata and controls

260 lines (189 loc) · 7.02 KB
Raw

JobPilot AI

Intelligent LinkedIn Easy Apply automation with AI-powered job matching

Python License LinkedIn Author

Automate Easy Apply, score job relevance with Gemini, and track every application — built by Md. Mahamudul Hasan.

✨ Key Features

  • 🤖 Easy Apply automation — Fills LinkedIn application forms automatically
  • 🎯 AI relevance scoring — Gemini scores each job 0–100; skips low-match listings
  • 📝 AI cover letter tailoring — Generates per-job cover letters from your resume text
  • 📊 Applied-jobs dashboard — Flask UI with relevance scores and external link tracking
  • 🌐 Multi-provider AI — Gemini (default), OpenAI, and DeepSeek supported
  • 🕵️ Stealth browser mode — Uses undetected-chromedriver to reduce bot detection
  • 📁 CSV history — Full audit trail of applied, skipped, and failed jobs

🛠 Tech Stack

Category Technologies
Language Python 3.11+
Browser automation Selenium, undetected-chromedriver
AI Google Gemini, OpenAI SDK, DeepSeek
Web dashboard Flask, Flask-CORS
Desktop automation PyAutoGUI
Data storage CSV (local files)

📋 Prerequisites

  • Windows 10/11 (primary platform; Linux/macOS partially supported)
  • Python 3.11+ — python.org/downloads
  • Google Chrome (latest stable)
  • LinkedIn account with Easy Apply access
  • Gemini API key (optional, for AI features) — Google AI Studio

🚀 Installation

1. Clone the repository

git clone https://github.com/mahamudul-hasan-cse/JobPilot-AI.git
cd JobPilot-AI

2. Install dependencies

pip install -r requirements.txt

3. Configure credentials

copy config\secrets.example.py config\secrets.py
copy config\personals.example.py config\personals.py
copy config\questions.example.py config\questions.py

Edit config/secrets.py:

  • Set your LinkedIn username and password
  • Add your Gemini llm_api_key
  • Set ai_provider = "gemini"

Edit config/personals.py with your real name, phone, and address.

4. Add your resume

Place your PDF resume at:

all resumes/default/resume.pdf

Paste your resume as plain text in config/questions.py → base_resume_text for AI tailoring.

5. (Optional) Install ChromeDriver

Run as Administrator on Windows:

setup\windows-setup.bat

⚙️ Configuration

File Purpose
config/secrets.py LinkedIn login, API keys, AI feature flags
config/personals.py Name, phone, address, EEO answers
config/questions.py Resume path, salary, cover letter, AI resume text
config/search.py Job titles, location, filters, blacklists
config/settings.py Bot behavior, browser timeouts, CSV paths

AI setup (config/secrets.py)

use_AI = True
ai_provider = "gemini"
llm_api_key = "YOUR_GEMINI_API_KEY"
llm_model = "gemini-2.0-flash"

use_relevance_scoring = True
ai_relevance_threshold = 65
strict_relevance_scoring = False

Job search (config/search.py)

search_terms = ["AI Engineer", "ML Engineer"]
location = "Bangladesh"
easy_apply_only = True
switch_number = 3   # Max successful applies per run

See SECURITY.md for what never to commit or share.

▶️ How to Run

Bot (main application)

Windows — double-click:

File Description
Start JobPilot AI.vbs Recommended — opens console and runs the bot
start_jobpilotai.bat Command Prompt launcher

Or from terminal:

python main.py

The launcher will: close Chrome → validate config → open guest Chrome → log into LinkedIn → apply to jobs.

Dashboard (optional)

python app.py

Open http://localhost:5000 to view applied jobs.

Preflight check

python scripts/preflight.py

📂 Project Structure

JobPilot AI/
├── main.py                      # Bot entry point
├── app.py                       # Flask dashboard entry point
├── requirements.txt
├── README.md
├── CHANGELOG.md
├── COPYRIGHT
├── LICENSE
├── SECURITY.md
├── .gitignore
├── .env.example
├── Start JobPilot AI.vbs        # Windows double-click launcher
├── start_jobpilotai.bat         # Windows batch launcher
│
├── config/                      # User settings (*.py gitignored except *.example.py)
│   ├── secrets.example.py
│   ├── personals.example.py
│   ├── questions.example.py
│   ├── search.py
│   └── settings.py
│
├── modules/                     # Core application code
│   ├── auth.py                  # LinkedIn login
│   ├── helpers.py               # Utilities and logging
│   ├── validator.py             # Config validation
│   ├── browser/                 # Chrome session + Selenium helpers
│   ├── ai/                      # Gemini, scoring, tailoring, factory
│   └── jobs/                    # Applier, form filler, filters, logger
│
├── scripts/
│   └── preflight.py             # Pre-run config/deps check
│
├── setup/                       # One-time environment setup
│   ├── windows-setup.bat
│   ├── windows-setup.ps1
│   ├── setup.sh
│   └── start_jobpilotai.ps1
│
├── web/
│   └── templates/
│       └── index.html           # Dashboard UI
│
├── all excels/                  # CSV history (gitignored, created at runtime)
├── all resumes/                 # Resume files (gitignored)
└── logs/                        # Debug logs (gitignored)

🗺 Roadmap

  • Multi-account support with profile switching
  • PostgreSQL / SQLite backend for application history
  • Email notifications on apply completion
  • Docker container for one-command deployment
  • Improved Easy Apply button detection (Next / Review / Submit)
  • Scheduled runs via Windows Task Scheduler integration
  • Application analytics dashboard (success rate, company stats)
  • Support for Indeed and Glassdoor

👤 Author

Md. Mahamudul Hasan

📄 License

This project is licensed under the GNU Affero General Public License v3.0.

See LICENSE for the full text.

Copyright (C) 2026 Md. Mahamudul Hasan

⭐ Star this repo if JobPilot AI helps your job search!