🇺🇸 English   •   🇷🇺 Русский

Here you can find a Python app that allows you to monitor for new posts to instantly send a comment once published, using Telegram bot-like functionality. You can specify the list of channels to monitor, as well as the behavior (prompt) to generate comments according to.

Developed with Telethon, gpt4free and Tenacity. No OpenAI account required.

This is a public part of my freelance project. Need services? Check out my website 👈👀.

Notice

Absolutely no way you should use such software for spamming. Telegram moderation will ban your account after a few reports from other users. Use in your own channels for fun and to get familiar with the technology.

Structure

telegram-autocomment-bot/
│
├── config/
│   ├── settings.ini
│   ├── prompt.txt
│   └── config_loader.py
│
├── utils/
│   ├── gpt_utils.py
│   └── log_utils.py
│
├── bot/
│   ├── bot_logic.py (core functionality)
│   └── event_handlers.py
│
├── tests/
│   ├── test_g4f.py
│   └── test_bing_provider.py
│
├── requirements.txt
└── start.py

Installation

1. Clone and navigate to the repository:

   git clone git@github.com:avrtt/telegram-autocomment-bot.git
   cd telegram-autocomment-bot
   

2. Since using virtual environments is a good practice, I highly recommend you to create one:

   python3 -m venv venv
   

   Then activate the virtual environment:

   • On Linux/macOS:

     source venv/bin/activate
     

   • On Windows:

     venv\Scripts\activate
     

3. Install the required dependencies:

   pip install -r requirements.txt
   

4. Create a config/settings.ini file with the following content:

   [Telegram]
   api_id = <your_api_id*>
   api_hash = <your_api_hash*>
   auto_join = true
   channel_usernames = <channel1, channel2, ... **>
   
   [GPT]
   provider = Bing
   
   [Debug]
   verbose = false
   
   [Proxy] # optional
   proxy_login =
   proxy_password =
   proxy_ip =
   proxy_port =
   
   [Metadata] # optional
   device_model =
   system_version =
   

   * Get your Telegram API ID and hash here. You should log in using the phone number of the bot account, navigate to "API development tools" and create a new application. Keep them secret!
   ** Specify channels without the @ symbol. Your private channels won't work.

5. Run:

   python start.py
   

6. Once prompted, provide Telethon with the same phone number you entered when created the API. You'll receive a code again.

7. If your account has two-step verification enabled, Telethon will prompt you for a password as well.

Now, if you go into Active sessions in your bot account, you will see the new device you just registered. Have fun!

Configuration

Behavior

The first thing you'd probably want to configure is a prompt that sets the bot's behavior. You can write your prompt in the prompt.txt file. Here it's simple: use the bot as if it were ChatGPT.

It's recommended to describe the behavior in the language you want the bot to write in.

Add something like Comment on the following text: at the end of the prompt, since the publication text gets inserted after the prompt in the code.

Settings

Here's the detailed description of variables in the settings.ini file (default values are underscored):

• auto_join: Whether the bot should join channels at the beginning of the session (true/false)
• provider: Which model provider to use (Bing, Liaobots, Phind, etc.)
• verbose: Turn on the verbose debugging mode that shows connection trace, full prompt texts and other info in your console (true/false)

Testing & troubleshooting

Testing provider connection

Run the isolated script for g4f:

python tests/test_g4f.py

You will likely get Error: Bing is not working, so run the isolated script for Bing to ensure:

python tests/test_bing_provider.py

If Bing isn't working, try to change provider or use proxy.

My ISP has blocked Bing/OpenAI

If you can't access such services without a VPN, then you're probably from Russia you could try using a proxy. Or, maybe, use a system-wide VPN, DPI bypass utilities, etc. Try GreenTunnel.

Using different providers

The full list of providers and their compatibility with models are listed in the g4f repository.

To do

• Replace Bing with config['GPT']['verbose'] in gpt_utils.py
• Add more tweaks
• Add unit tests

Contrubition

Feel free to open PRs and issues.

License

MIT