This program is an AI-powered Telegram chatbot that utilizes various AI models to have conversations, answer questions, and generate responses. Some of the key features of this chatbot are:
- Handles private chats and group chats with Telegram users.
- Allows users to set a persona/profile for the bot, which changes its responses.
- Includes commands for generating custom responses, summarizing messages, outlining messages step-by-step, explaining meaning/significance, elaborating on messages, visualizing messages as graphs, telling jokes, and more.
- Persists user sessions and profiles to disk so conversations can continue across restarts.
- Uses AI models like GPT-4 (via EvilCorp SecondPilot) and Claude for message generation.
- Can visualize graphviz code via the
!vis
command (as reply, and verbatim). - Handles errors and messages that are too long for Telegram.
- Includes admin commands for banning users and debugging the bot.
To get started with this program, you will need to add the relevant tokens to the .env
file. The required tokens are:
- TELEGRAM_TOKEN
- GITHUB_PAT_TOKEN
- CLAUDE_API_TOKEN
Once you have added the tokens to the .env
file, you can run the program by executing the run.sh
script.
To use the chatbot, simply start a chat with the bot in Telegram. You can then use any of the available commands to interact with the chatbot and generate responses.
If you would like to contribute to this project, please fork the repository and submit a pull request. We welcome contributions of all kinds, including bug fixes, new features, and documentation improvements.
This program is licensed under the MIT License. See the LICENSE file for more information.