SignalGPT

Chat with AI on Signal

Definition

SignalGPT is a Python bot that connects Signal messaging with DeepSeek's AI API, allowing you to get AI responses directly in your Signal conversations.

Features ✨

• Receive Signal messages and respond with AI-generated answers
• Automatic timestamp tracking to avoid duplicate responses
• Profile picture changes to indicate online/offline status
• Conversation logging for debugging and record-keeping
• Simple setup with Signal's command-line interface
• Linux / Windows support in one file (changing OSCHOICE variable)

Prerequisites 📋

• Python 3.7+
• Java Runtime Environment (JRE) 17+ (required for signal-cli)
• A Signal account with phone number verification
• DeepSeek API key

Installation Guide 🛠️

1. Clone the repository

git clone https://github.com/rainerigius/SignalGPT.git
cd SignalGPT

2. Install Python dependencies

pip install requests

3. Set up signal-cli

1. Download signal-cli 0.13.13 from the official releases:

   wget https://github.com/AsamK/signal-cli/releases/download/v0.13.13/signal-cli-0.13.13.tar.gz

   Or download manually from:
   https://github.com/AsamK/signal-cli/releases/tag/v0.13.13

2. Extract the archive:

   tar -xzf signal-cli-0.13.13.tar.gz

3. Rename the extracted folder:

   mv signal-cli-0.13.13 signal-cli

4. Configure the bot

Edit the following variables in the script:

MY_NUMBER = "+1234567890"  # Your Signal-registered phone number
DEEPSEEK_API_KEY = "your_deepseek_api_key"  # Your DeepSeek API key
DEEPSEEK_MODEL = "deepseek-chat" # or "deepseek-reasoner" for better performance but slower response time
OSCHOICE = "Windows" # Change to "Windows" if running on Windows

You have to create your own red.png and green.png files.

How It Works ⚙️

1. Message Reception:

   • The bot continuously checks for new Signal messages using signal-cli
   • Messages are parsed and filtered to avoid duplicates

2. AI Processing:

   • Valid messages are sent to DeepSeek's API
   • Responses are sanitized to work with Signal's formatting

3. Message Sending:

   • AI responses are sent back to the original sender
   • All interactions are logged for reference

4. Status Indicators:

   • Profile picture changes to green.png when online
   • Changes to red.png when shutting down

First Run Setup 🚀

1. Run the script:

   python signalgpt.py

2. If this is your first time:

   • The bot will attempt to register your phone number
   • You'll receive an SMS verification code
   • Enter the code when prompted

3. The bot will then:

   • Start listening for messages
   • Change its profile picture to indicate it's online

Usage 💬

1. Simply send a message to your bot's number from any Signal client
2. The bot will respond with AI-generated content
3. All conversations are logged in response_log.txt

Complete File Structure 📂

SignalGPT/
├── signal-cli/            # Signal CLI binaries
├── signalgpt.py           # Main bot script
├── last_message.txt       # Stores last processed timestamp
├── response_log.txt       # Conversation history
├── green.png              # Online status avatar
└── red.png                # Offline status avatar

Customization 🎨

You can modify:

• AVATAR_ONLINE/AVATAR_OFFLINE - Change status indicator images
• System prompt in get_deepseek_response() - Adjust AI behavior
• sanitize_text() - Modify how messages are cleaned before sending

Troubleshooting 🐛

Problem: "Error on bin/signal-cli.bat or bin/signal-cli"
Solution: Check the OSCHOICE variable and restart the bot

Problem: "Signal number not registered"
Solution: Delete the signal-cli/data folder and restart the bot

Problem: Messages not being received
Solution: Check your internet connection and verify signal-cli is working:

Windows

./signal-cli/bin/signal-cli.bat -a YOUR_NUMBER receive

Linux

./signal-cli/bin/signal-cli -a YOUR_NUMBER receive

Problem: API errors
Solution: Verify your DeepSeek API key and quota

Security Note 🔒

• Your DeepSeek API key is stored in plaintext - keep it secure
• Signal credentials are stored in signal-cli/data - protect this directory
• Consider using environment variables for sensitive data

License 📜

MIT License - Free for personal and educational use