But if you're planning to use Indeed for searching jobs without auto-applying - it's a great site. You can use it for searching suitable jobs and then apply for them manually.

LinkedIn AI Job Applier Chat 👇

📋 Table of Contents

• ✨ Features
• 🚀 Getting Started
  • Prerequisites
  • Installation
• 🔧 Configuration
• ▶️ Usage
  • LinkedIn Messages Manager
  • Dashboard
  • Dashboard Routes And APIs
• 💵 Vacancy application cost
• ✅ Running Tests
• 🐞 Troubleshooting
• 📨 Telegram Instruction
• 🤝 Contributing
• 📜 License
• 🙏 Acknowledgements

✨ Features

This project enhances the original codebase with several powerful new features:

• 🌐 Multi-Platform Support: Applies to jobs on LinkedIn and Indeed. Switch between platforms with a single JOB_SITE setting in config/app_config.py.
• 🌐 Universal Job Application: Applies to ALL job vacancies on LinkedIn (not just Easy Apply) thanks to the browser-use library.
• 🔒 Data Anonymization: Protects your privacy by replacing personal data with mock information before sending it to the LLM provider, ensuring your sensitive information remains secure.
  • Note: Auto resume parsing and applying of Non-Easy Apply vacancies don't use anonymization. Additionally, country, city, and birth date are not anonymized to maintain the quality of LLM responses.
• 🎯 Improved Intelligent Resume Generation: Uses AI to tailor every generated resume to the current vacancy for maximum match, adapting skills, experience, projects and achivements to the job description.
• 🎭 Patchright Integration: Now uses Patchright (a patched build of Playwright) instead of Selenium/Playwright for faster, more reliable and undetectable browser automation. Patchright patches Chromium's automation detection signals to bypass Cloudflare and other bot-detection systems.
• 🖥️ Headless Mode: Run the bot in headless mode if you want to use the bot in server environment or while working with your computer. This allows the bot to operate without a visible browser window while maintaining full functionality.
• ⏸️ Pause/Resume Control: Pause the bot at any time by pressing Ctrl+X and continue when ready, giving you full control over execution without stopping the entire process.
• 📊 Skill Statistics: Analyzes job descriptions to identify the most in-demand skills, helping you tailor your resume effectively.
• 🧠 Intelligent Error Handling: If LinkedIn's Easy Apply feature encounters errors (e.g., incorrectly filled fields), the bot will attempt to fix them automatically.
• 📈 Local Monitoring Dashboard: Includes a local web dashboard for live monitoring, run history, screenshot review, config editing, bot controls, and JSON export of individual runs.
• 💬 LinkedIn Messages Manager: Includes a separate inbox triage workflow for classifying LinkedIn conversations, drafting replies, archiving spam, starring recruiter outreach, applying the Jobs label, and recording all message decisions to a persistent ledger for dashboard review.
• 🤝 Automated Networking: Includes a powerful tool to search for and connect with "Open Networkers" (LIONs) automatically, expanding your professional network with people likely to accept your requests.
• ☑️ Smart Checkbox Handling: Automatically detects and answers checkbox questions in LinkedIn Easy Apply forms with intelligent context-aware responses.
• 🔗 Contextual Question Processing: Considers previous answers when responding to follow-up questions like "If yes/no, who/when/where?" for more accurate and relevant responses.
• 🤖 AI-Powered Resume Parsing: Automatically parses your resume from a text file into a structured format using an LLM (Large Language Model).
• 📄 Auto Resume Text Generation: If data/resumes/resume_text.txt is missing but a PDF resume is present in data/resumes/, the bot automatically extracts the text from the PDF and generates a properly formatted resume_text.txt using the LLM — no manual transcription needed.
• 📄 New Resume Style: Includes the modern "FAANGPath" resume style for generating professional-looking resumes.
• 📲 Telegram Integration: Delivers comprehensive reports and error notifications directly to your Telegram chat after each run.
• 💡 Resume Recommendations: Provides AI-generated suggestions to improve your resume based on job market trends.
• 🌐 Proxy Support: Supports using proxies for both Gemini, OpenAI and OpenRouter models.
• 🏗️ Robust Configuration: Uses Pydantic models for validating configuration, resume, and other data structures, reducing runtime errors.
• 🚀 Improved LLM Logic:
  • Utilizes improved system instructions and prompts for higher-quality LLM responses.
  • Employs Gemini or OpenRouter as the default LLM, which is often more cost-effective (sometimes even free!) than OpenAI.
  • Enhances question-answering logic to avoid LLM "hallucinations" by skipping questions where it lacks sufficient information.
• 🔐 Secure Secrets Management: Stores sensitive keys and credentials in a .env file for better security.
• 🕒 Automated Scheduling: A built-in timer allows the bot to run automatically every 24 hours.
• 💻 Numerous improvements that simplify development and debugging: advanced logging, pre-commit hooks, fast and easy installation using uv, etc.

🚀 Getting Started

Prerequisites

• Python 3.12
• Git
• uv (optional, for faster installation)

Installation

Option 1: Local Installation

1. Clone repository and create virtual environment

   # Clone the repository
   git clone https://github.com/beatwad/LinkedIn-AI-Job-Applier-Ultimate.git
   cd LinkedIn-AI-Job-Applier-Ultimate
   

2. Install dependencies

   uv sync
   

3. Install additional software

   # Install Chromium browser for Playwright
   playwright install chromium
   

Option 2: Docker Installation

1. Clone repository

   git clone https://github.com/beatwad/LinkedIn-AI-Job-Applier-Ultimate.git
   cd LinkedIn-AI-Job-Applier-Ultimate
   

2. Build Docker image

   docker build -t linkedin .
   

3. Run Docker container

   docker run -v $(pwd)/data:/app/data \
              -v $(pwd)/browser_session:/app/browser_session \
              -v $(pwd)/logs:/app/logs \
              linkedin
   

   ⚠️ IMPORTANT: When running from Docker:

   • You MUST set HEADLESS_MODE = True in config/app_config.py. Docker containers don't support graphical interfaces, so the browser must run in headless mode.
   • The Ctrl+X pause/resume feature will not work in Docker (keyboard control requires an X server). The bot will run continuously without the ability to pause.
   • Resume style selection is automatic - the bot will use the default style (FAANGPath) instead of prompting for selection, since Docker doesn't support interactive prompts.

🔧 Configuration

1. Secrets (.env file): Create a .env file in the root directory by copying the example file:

   cp .env_example .env
   

   Now, fill in the required values in your .env file:

   # Your LinkedIn credentials (used when JOB_SITE="linkedin" in app_config.py)
   linkedin_email="your_linkedin_email@example.com"
   linkedin_password="your_linkedin_password"
   
   # Your Indeed credentials (used when JOB_SITE="indeed" in app_config.py)
   indeed_email="your_indeed_email@example.com"
   
   # Your LLM API Key (e.g., Gemini)
   llm_api_key="your_llm_api_key"
   
   # Optional proxy for LLM requests
   llm_proxy="http://your_proxy_url:port"
   
   # LLM endpoint URL (required for "openai_compatible", optional for "ollama")
   llm_api_url="http://localhost:8000/v1"
   
   # Your Telegram Bot Token for sending error messages and reports
   tg_token="your_telegram_bot_token"
   
   # Address of your chat in format "@name_of_your_chat"
   tg_chat_id="@name_of_your_chat"
   
   # ID of topic where bot will send error messages
   # You can find out ID of the topic by copying the link to the topic in Telegram app and extracting the number before the last slash
   # For example, if the link to the topic is https://t.me/channel_name/123/321, the ID of the topic is 123
   tg_err_topic_id="[ID of error topic]"
   
   # ID of topic where bot will send everyday report on the job applies done
   tg_report_topic_id="[ID of report topic]"
   

   Telegram options are optional. You can remove them from .env file and application will still be able to work but won't be able to set reports and error messages via Telegram.

2. Job Search Parameters (config/search_config.yaml): Copy this file from `examples/config/search_confi