Dont-Read-GPT is my Discord bot designed to help me save time by summarizing web content from many knowledge sources like Arxiv. Instead of saving content that I may never read, let the bot generate a concise summary so I can quickly grasp the key points and insights. It currently supports:
- Arxiv
- Huggingface model
- Github
- Youtube
- WeChat Article
- other blogs or article page
On the roadmap: (contribution welcomed!)
- PDF link
- Medium member blog
- Hackernews
- X/Twitter thread
- Supports summarization of web content from variety of knowledge sources
- Uses OpenAI GPT model for high-quality summaries tailored for each source
- Easy integration with Discord, just send a link and AI does everything else.
- Clone the repository:
git clone https://github.com/phunterlau/dont-read-gpt
- Change to the project directory:
cd dont-read-gpt
- Install the required dependencies:
pip install -r requirements.txt
- Set up your OpenAI API key:
export OPENAI_KEY="your_openai_api_key"
- Set up your Discord bot token and other app tokens:
export DISCORD_BOT_TOKEN="your_discord_bot_token"
# get Reddit tokens as https://www.reddit.com/prefs/apps
export REDDIT_APP_ID="your reddit app ID"
export REDDIT_APP_SECRET="your reddit secret"
- Run the bot:
python my_bot.py
Invite the Dont-Read-GPT bot to your Discord server and use the following commands:
!wget <URL>: Summarizes the content from the given URL (webpage, ArXiv paper, or GitHub repository)
- Summarize a tech blog post:
!wget https://toooold.com/2023/04/08/magnificient_underdogs.html
- Summarize an ArXiv paper:
!wget https://arxiv.org/abs/1706.03762
- Summarize a GitHub repository:
!wget https://github.com/phunterlau/dont-read-gpt
Contributions are welcome! Feel free to open issues or submit pull requests to improve the Dont-Read-GPT bot.
This project is licensed under the MIT License. See the LICENSE file for details.