🤖 GPT-Telegramus

	The best Telegram bot for ChatGPT, EdgeGPT (aka Bing AI), DALL-E, Bing Image generator, Bard and Gemini with stream writing, requests with images (for Bard and Gemini only), multiple languages, admin control, automatic proxy searcher, data logging and more!

😋 Support project

Support the project by buying and listening to my music 🎵

• 🟦 Bandcamp
• 🟧 SoundCloud

Or message me if you would like to donate 💰

🤗 Contributors

• 1️⃣ Sprav04ka - Tofii'skovyi' language, Testing, Super beautiful poster, Project Logo, Motivation
• 2️⃣ Hanssen - Markdown parsing, bard images, /chat command, caption fix, loading emoji, dynamic splitting, code block splitting, Gemini module, Docker fix, GitHub actions fix and much much more
• 3️⃣ Sergey Krashevich - Docker, GitHub Actions
• 4️⃣ Wahit Fitriyanto - Indonesian language
• 5️⃣ Alexander Fadeyev - EdgeGPT Fix
• 6️⃣ AnthroAsja - Author of the Belarusian translation
• 7️⃣ Anonymous Samurai - Author of the Ukrainian translation
• 8️⃣ Dunya Jafari - Author of the Persian translation

🏗️ Requirements

• Python 3.10
• GO lang (for ChatGPTProxy)
• Unblocked access to the telegram bot official API
• Other requirements specified in the requirements.txt file

📙 Project based on

• ChatGPT (ChatGPT by OpenAI): https://chat.openai.com/chat
• Bing chat (EdgeGPT by Microsoft): https://bing.com/chat
• DALL-E (DALLE by OpenAI): https://openai.com/dall-e-2/
• Bard (Bard by Google): https://bard.google.com/
• Gemini (Gemini by Google): https://makersuite.google.com/app/prompts/new_freeform
• acheong08/ChatGPT (API): https://github.com/acheong08/ChatGPT
• acheong08/EdgeGPT (API): https://github.com/acheong08/EdgeGPT
• jacobgelling/EdgeGPT (API): https://github.com/jacobgelling/EdgeGPT
• dsdanielpark/Bard-API (API): https://github.com/dsdanielpark/Bard-API
• acheong08/BingImageCreator (API): https://github.com/acheong08/BingImageCreator
• google/generative-ai-python (API): https://github.com/google/generative-ai-python
• python-telegram-bot (Telegram bot API): https://github.com/python-telegram-bot/python-telegram-bot

❓ Get started

1. Install Python 3.10, venv and pip
2. Download source code
3. Create venv python -m venv venv / python3 -m venv venv / python3.10 -m venv venv
4. Activate venv source venv/bin/activate / venv\Scripts\activate.bat
5. Check python version using python --version command. It should be 3.10.X
6. Install requirements pip install -r requirements.txt --upgrade
7. Access the modules you want to use (Generate an API key / access token for ChatGPT / DALL-E / Gemini, save cookies for EdgeGPT / Bard)
8. Start ChatGPTProxy for ChatGPT (see 🔗 Chat-GPT Base URL (proxy) section)
9. Carefully change all the settings (in the config.json file) to suit your needs. If you have questions regarding any setting, open an issue, I will add a more detailed description
10. Create bot at https://t.me/BotFather
11. Type Bot's token into api_key in telegram in config.json file
12. Run main script python main.py

• ChatGPT
  • Browser-like Chat-GPT. Currently, without extensions and image requests (text only) (because I don't have a paid account to test it)
  • Stream response support
  • Chat history support
  • Works better with API type 1 and an access_token for authentication
  • Note: Please refer to the 🔗 Chat-GPT Base URL (proxy) section for correct usage.
• DALL·E
  • Image generation tool from OpenAI
  • Requires an OpenAI account with unexpired credits
• EdgeGPT (aka Bing AI) (aka Sydney)
  • Supports conversation style /style
  • Stream response support
  • Chat history support
  • Web-browsing (probably) and sources support
  • Unfortunately, it can't accept images as input yet, nor can it generate them. Please use Bing ImageGen to generate images.
• Bing ImageGen
  • Bing Image Generator. Used as a separate module due to issues with the EdgeGPT module
  • Free and unlimited
• Bard
  • Google's Bard AI
  • Chat history support
  • Web-browsing support (probably)
  • Now supports requests with images (you can send an image with text to it) NEW
• Gemini
  • Google's AI using the Gemini Pro model
  • Chat history support
  • Requests with images (you can send an image with text to it)
  • Requests with images will not be recorded in the chat history since Google hasn't support this

💬 Bot messages

• You can edit telegram bot messages by editing file messages.json. You can add new lines by adding \\n
• You can add a new language simply by copying one of existing language (with {}) to the end of file (before the last ]) and translating all values

Note: make sure you don't delete argumensts {0}, {1}, ... in message and please restart python script to apply changes

Currently available languages

• 🇺🇸 English
• 🇷🇺 Русский
• ‍☠️ Тофийсковый
• 🇮🇩 Bahasa Indonesia
• 🇨🇳 简体中文
• 🇧🇾 Беларуская
• 🇺🇦 Українська
• فارسی 🇮🇷

If you want to add a language, create a pull request 💜

🐧 Running as service on linux

1. Install Python 3.10, venv and pip

2. Download source code

3. Create venv python -m venv venv / python3 -m venv venv / python3.10 -m venv venv

4. Clone repo

   1. git clone https://github.com/F33RNI/GPT-Telegramus/
   2. cd GPT-Telegramus

5. Edit config.json

6. Install systemd

   1. sudo apt-get install -y systemd

7. Create new service file

   1. sudo nano /etc/systemd/system/gpt-telegramus.service

      [Unit]
      Description=GPT-Telegramus service
      After=multi-user.target
      
      [Service]
      Type=simple
      Restart=on-failure
      RestartSec=5
      
      WorkingDirectory=YOUR DIRECTORY HERE/GPT-Telegramus
      ExecStart=YOUR DIRECTORY HERE/GPT-Telegramus/run.sh
      
      [Install]
      WantedBy=multi-user.target
      

8. Reload systemctl daemon

   1. sudo systemctl daemon-reload

9. Enable and start service

   1. sudo systemctl enable gpt-telegramus
   2. sudo systemctl start gpt-telegramus

10. Note: Please use sudo systemctl kill gpt-telegramus and then sudo systemctl stop gpt-telegramus if only sudo systemctl stop gpt-telegramus not working

🍓 Running on Raspberry Pi (ARM)

1. Install Python 3.10 if not installed

   1. sudo apt-get update
      sudo apt-get install -y build-essential tk-dev libncurses5-dev libncursesw5-dev libreadline6-dev libdb5.3-dev libgdbm-dev libsqlite3-dev libssl-dev libbz2-dev libexpat1-dev liblzma-dev zlib1g-dev libffi-dev
      wget https://www.python.org/ftp/python/3.10.13/Python-3.10.13.tgz
      sudo tar zxf Python-3.10.13.tgz
      cd Python-3.10.13
      sudo ./configure --enable-optimizations
      sudo make -j 4
      sudo make altinstall

   2. Check version by typing python3.10 -V. After this, you should use python3.10 command instead of python or you can add it to the bashrc by typing echo "alias python=/usr/local/bin/python3.10" >> ~/.bashrc

2. For ChatGPTProxy, install GO and tls-client

   1. sudo apt-get install golang-go
      
      git clone https://github.com/bogdanfinn/tls-client.git
      cd tls-client/cffi_dist
      sudo GOOS=linux CGO_ENABLED=1 GOARCH=arm64 CC="gcc" /usr/local/go/bin/go build -buildmode=c-shared -o tls-client-arm64.so
      mv ./dist/tls-client-arm64.so /home/pi/.local/lib/python3.10/site-packages/tls_client/dependencies/
      mv ./dist/tls-client-arm64.h /home/pi/.local/lib/python3.10/site-packages/tls_client/dependencies/

3. Build ChatGPTProxy

   1. git clone https://github.com/acheong08/ChatGPTProxy
      /usr/local/go/bin/go build

   2. Create and start service as mention in the 🔗 Chat-GPT Base URL (proxy) section

4. Follow the 🐧 Running as service on linux guide

🐋 Running in Docker

From GitHub Package

1. Clone repo or download config.json and messages.json

2. Edit the config.json, set options in the files section to the path in the container (/app/config/<FILE_NAME>)

3. Run the container

   docker run -d -e TELEGRAMUS_CONFIG_FILE="/app/config/config.json" -v <YOUR_CONFIG_FOLDER>:/app/config --name gpt-telegramus --restart on-failure ghcr.io/f33rni/gpt-telegramus:latest

   If you want to try the preview version

   docker run -d -e TELEGRAMUS_CONFIG_FILE="/app/config/config.json" -v <YOUR_CONFIG_FOLDER>:/app/config --name gpt-telegramus --restart on-failure ghcr.io/f33rni/gpt-telegramus:edge

Build Manually

1. Install Docker

2. Clone repo

3. Build container

   docker buildx build -t telegramus --load -f Dockerfile .

4. Run the container

   docker run -d --name gpt-telegramus --restart on-failure telegramus

   or if you want to use a custom config

   docker run -d -e TELEGRAMUS_CONFIG_FILE="/app/config/config.json" -v <YOUR_CONFIG_FOLDER>:/app/config --name gpt-telegramus --restart on-failure telegramus

🔗 Chat-GPT Base URL (proxy)

Default base URL by acheong08/ChatGPT stopped working. Below is an instruction on how to launch your own ChatGPT proxy server (on Linux)

1. Install GO

   1. sudo apt-get update
   2. sudo apt-get install golang-go

2. Clone acheong08/ChatGPTProxy repo

   1. git clone https://github.com/acheong08/ChatGPTProxy
   2. cd ChatGPTProxy

3. Build it

   1. go build

4. Install systemd

   1. sudo apt-get install systemd

5. Create service

   1. sudo nano /etc/systemd/system/chatgpt-proxy.service

      [Unit]
      Description=ChatGPTProxy service
      After=multi-user.target
      
      [Service]
      Type=simple
      Restart=on-failure
      RestartSec=5
      
      # Proxy (if needed, or remove this and next line)
      Environment="http_proxy=http://USERNAME:PASSWORD@IP:PORT"
      
      # ChatGPT login
      Environment="OPENAI_EMAIL=YOUR_EMAIL"
      Environment="OPENAI_PASSWORD=YOUR_PASSWORD"
      
      WorkingDirectory=PATH_TO_ChatGPTProxy_DIRECTORY
      ExecStart=PATH_TO_ChatGPTProxy_DIRECTORY/ChatGPTProxy
      
      [Install]
      WantedBy=multi-user.target
      

6. Reload systemctl daemon

   1. sudo systemctl daemon-reload

7. Enable and start service

   1. sudo systemctl enable chatgpt-proxy
   2. sudo systemctl start chatgpt-proxy

8. See logs to make sure it's running and see current port

   1. systemctl status chatgpt-proxy

9. Add it's IP to config.json

   1. ex. "base_url": "http://127.0.0.0:9090/api/",

🌎 Proxy to bypass geo-blocking

NOT TESTED AND MOST LIKELY WILL NOT WORK WITH BARD AND BING AI

I won't respond to the issue if it's related to auto proxy

It is possible to bypass geo-blocking of ChatGPT, EdgeGPT, DALL-E or Bard. GPT-Telegramus includes automatic proxy-list downloading with periodic checks

1. Set enabled in proxy_automation in config.json to true
2. Set auto in proxy for modules you want to connect automatic proxy search
3. Restart app and hope for the best.

GPT-Telegramus will download the proxy list itself and start trying various proxies (see console for logs). Sometimes, searching for a proxy can take a long time. If you think that the found proxy is not suitable, you can restart the bot using the /restart command

• To use manual proxy, specify tham in http://IP:Port or http://user:password@IP:Port format (even if they are HTTPS proxies) in proxy for modules you need

🤖 Telegram bot commands

• 📄 /start - Welcome message and bot version
• ❓ /help - Show this message
• ↕️ /module - Change module to chat with
• 🧹 /clear - Clear chat history
• 🌎 /lang - Change the language
• 🆔 /chatid - Show your chat_id
• /style - Bing AI conversation style
• /chat <Text request> - Request to the default module
• /chatgpt <Text request> - Request to the ChatGPT module
• /dalle <Text request> - Request to the DALL-E module
• /bing <Text request> - Request to the Bing AI module
• /bingigen <Text request> - Request to the Bing Image Generator
• /bard <Text request> - Request to the Bard module
• /gemini <Text request> - Request to the Gemini module

Admin commands

• 💬 /queue - Show requests queue
• 🔃 /restart - Restart proxies and telegram bot
• 👤 /users - Show list of all users
• 🔨 /ban <id> [reason] - Ban a user by their id with reason (optional)
• 🔓 /unban <id> - Unban a user by their id
• 📢 /broadcast <message> - Send text message to everyone except banned users

📄 Chat history

GPT-Telegramus saves chat history for some modules locally ("conversations_dir": "conversations") to keep conversation with users

📜 Data collecting

GPT-Telegramus has a built-in data collecting function (request and response)

• For ChatGPT, EdgeGPT (aka Bing AI), Bard and Gemini response is saved as plain text and Base64-encoded images
• For DALL-E and Bing Image generator response is saved as Base64-encoded image (in the same text file)

You can enable and configure data collection in config in data_collecting block

NOTE: Please make sure you notify your bot users that you are collecting data

📝 TODO

• Add some free GPT-4 model
• Add image input and generation for EdgeGPT (and ChatGPT?)
• Add Bard's text-to-speech
• Add an API, so you can use GPT-Telegramus from outside the bot

✨ Contribution

• Anyone can contribute! Just create a pull request

🚧 P.S. This project is still under development

• If you find a bug in GPT-Telegramus, please create an Issue