LinkedIn Networking

Python code to automatically expand your LinkedIn network based on your interests with personalized messages using AI (Gemini or GPT-4o).

Features

• Search LinkedIn for people based on keywords and location
• Filter connections by country
• Send personalized connection requests
• Generate AI-powered connection messages using either Google's Gemini AI or GPT-4o
• Log successful connections
• Exclude specific profiles
• Rate limiting and connection limits for account safety
• Automated browser interaction with anti-detection measures

Prerequisites

• Python 3.8 or higher
• Chrome browser installed
• Either a Google Gemini API key or GPT-4o API key
• LinkedIn account

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/LinkedIn_networking.git
   cd LinkedIn_networking

2. Install required packages:

   pip install -r requirements.txt

Setup

1. Create a .env file in the root directory with the following variables:

   LINKEDIN_USERNAME=your_email@example.com
   LINKEDIN_PASSWORD=your_password_here
   GEMINI_API_KEY=your_gemini_api_key_here  # Optional if using GPT-4o
   gpt_4o=your_gpt4o_api_key_here  # Optional if using Gemini
   

2. Get your preferred AI API key:

   • For Gemini: Visit https://makersuite.google.com/app/apikey
   • For GPT-4o: Use your existing GPT-4o API key

3. Configure settings in config/parameters.py:

   # Search parameters
   search_keywords = 'Your, Keywords, Here'  # Comma-separated
   search_country = 'Your Country Here'  # For reference only
   search_geo_urn = '103644278'  # LinkedIn's geoUrn ID for location filtering
   exclude_connections = ''  # Comma-separated list of names to skip
   max_pages_to_search = 1  # Number of search pages to process
   
   # Connection limits
   max_connections_per_day = 20  # LinkedIn typically allows ~100 per week
   max_connections_per_session = 10  # Stop after this many in a single run
   connection_delay_seconds = 5  # Time between requests
   
   # AI Configuration
   ai_provider = "gemini"  # Options: "gemini" or "gpt4o"
   enable_ai_analysis = True  # Toggle AI message generation
   gpt4o_endpoint = "https://models.inference.ai.azure.com"  # If using GPT-4o
   gpt4o_model = "gpt-4o"  # Model name for GPT-4o

Usage

1. Test the AI integration (recommended before first use):

   python tests/test_gemini.py

   This will test both AI providers if you have configured them, validating:

   • API key configuration
   • Direct API connections
   • Message generation
   • Message processing and formatting

2. Run the main script:

   python src/linkedin/linkedIn.py

The script will:

• Log in to your LinkedIn account
• Search for profiles matching your keywords
• Generate personalized connection messages using your chosen AI provider
• Send connection requests while respecting rate limits
• Log successful connections in CSV format

Command-Line Arguments

The script supports the following command-line arguments when running directly from the connector:

python src/linkedin/linkedin_connector.py --max-connections 15 --test-mode

Available arguments:

• --max-connections - Set the maximum number of connection requests to send in this session (overrides the default in parameters.py)
• --test-mode - Run in test mode with additional logging output

When running the main script (linkedIn.py), these arguments are passed through configuration in parameters.py rather than command-line arguments.

Configuration Options

In config/parameters.py, you can customize:

• search_keywords: Comma-separated keywords for finding connections
• search_country: Country name (for reference only)
• search_geo_urn: LinkedIn's geoUrn ID for location filtering (e.g., "103644278" for United States)
• exclude_connections: Comma-separated list of profile names to skip
• max_pages_to_search: Number of search result pages to process
• max_connections_per_day: Daily connection request limit (default: 20)
• max_connections_per_session: Connection limit per script run (default: 10)
• connection_delay_seconds: Delay between requests (default: 5)
• ai_provider: Choose between "gemini" or "gpt4o"
• enable_ai_analysis: Toggle AI message generation (default: True)
• gpt4o_endpoint: API endpoint for GPT-4o (if using)
• gpt4o_model: Model name for GPT-4o requests
• connections_file: CSV file to log successful connections (default: 'connections.csv')

Supported Countries

The script includes built-in geoUrn IDs for the following countries:

• United States (103644278)
• United Kingdom (101165590)
• Canada (101174742)
• Australia (101452733)
• India (102713980)
• Germany (101282230)
• France (105015875)
• Italy (103350119)
• Spain (105646813)
• Netherlands (102890719)

Safety Features

• Connection Limits: Built-in daily and per-session limits
• Rate Limiting: Configurable delays between requests
• Anti-Detection: Browser automation masking
• Safe Fallbacks: Default messages if AI generation fails
• Error Recovery: Handles connection failures gracefully

Important Notes

• LinkedIn Limits: LinkedIn restricts connection requests (~100/week)
• CAPTCHA Handling: Manual intervention needed if CAPTCHA appears
• Account Safety: Start with conservative limits and adjust based on your account age
• API Usage: Monitor your Gemini API quota usage
• Terms of Service: Use responsibly and in compliance with LinkedIn's terms

Troubleshooting

1. If AI messages fail:

   • Verify your API key in .env
   • Run tests/test_gemini.py to diagnose issues
   • Check API quota and billing status
   • Verify the GPT-4o endpoint URL if using GPT-4o

2. If LinkedIn automation fails:

   • Check your credentials in .env
   • Verify internet connection
   • Ensure Chrome browser is installed
   • Try reducing connection frequency
   • Check if LinkedIn has updated their HTML structure
   • Clear browser cookies/cache if login issues persist

Contributing

Pull requests are welcome. For major changes:

1. Open an issue first to discuss proposed changes
2. Update tests as needed
3. Follow the existing code style
4. Add comments for complex logic

License

GNU General Public License v3.0

Find me