Skip to content

Sarcastic image roasting Telegram bot, powered by the GPT-4 Vision API. Delivers sick burns, balancing on the edge of OpenAI's policy.

License

Notifications You must be signed in to change notification settings

atwoodhouse/roastie

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Roastie - Sarcastic image roasting Telegram bot

Roastie is a unique Telegram bot, crafted in Typescript, that harnesses the power of OpenAI's GPT-4 Vision API. It delivers snappy, borderline roasts in response to images sent by users. Perfectly balancing on the edge of OpenAI's policy, Roastie offers humorously sarcastic comments, making it a hilarious addition to any chat.

Prerequisites

  • Node.js v18. Use nvm for optimal installation.
  • pnpm for package management. Install globally via npm i -g pnpm.

Setup Guide

  1. Clone the repository.
  2. Create a .env file as per the given template.
  3. Install dependencies using pnpm i.
  4. Initiate the bot with pnpm start.

Configuration

Prepare a .env file in the project's root directory. Include the following keys and your respective values:

Key Description
BOT_TOKEN Your Telegram bot token, obtainable from BotFather.
OPENAI_API_KEY API key for OpenAI.
ALLOWED_CHAT_ID The ID of the chat where the bot is allowed to operate (optional).

Ensure to replace the placeholders with actual data.

The current roasting prompt is in Swedish but can easily be translated to any language to suit your audience.

Installation

pnpm install

Usage

pnpm start

Contributing

Contributions are welcome! Submit pull requests or open issues for any bugs or feature suggestions.

Support

Need help or support? Create an issue in the repository for assistance.

About

Sarcastic image roasting Telegram bot, powered by the GPT-4 Vision API. Delivers sick burns, balancing on the edge of OpenAI's policy.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published