PaLM-Discord-Bot

Google PaLM API in Discord.

Google PaLM API only works in the United States.

Notice

Discord only can handle 2000 characters per message, so I manually set the PaLM can only outputting no more than 145 tokens. If your channel can handle 4000 or even more, make the output_token_limit= 145 in main.py value doubled or more.

Also, I dropped the temperature option, you can always add it back.

To-do

• Adapt the latest Google Vertex AI platform for Gemini(pro)
• Use .env instead
• Integrate this with any other AIGC APIs and make this more versatile?

Example:

Command

/palm to chat with PaLM.

Installation

Get your Discord API

1. Go Discord Dev
2. Go Application on the left part
3. Go New Application, create a new Bot
4. Go Bot on the left
5. Click Add bot
6. Turn on MESSAGE CONTENT INTENT
7. Save Change
8. View and save your API Key on the top of page View Token, or reset one if you don't remember
9. Click OAuth2 on the left part
10. Click on URL Generator
11. Choose bot in SCOPES, Administrator in BOT PERMISSIONS
12. Copy the link at bottom into your browser, add bot and authorize to you server.

Get your PaLM API

1. Join the waitlist here, after you passed the waitlist, you'll receive an email like this:
2. Go MakerSuite, and create an API key for this application.

Deploy on your machine

1. Star this project I guess
2. git clone this repository
3. pip install discord google-generativeai
4. cd PaLM-Discord-Bot
5. nano main.py
6. replace with your Discord API Key and PaLM API Key at following lines: palm.configure(api_key = "PaLM_API_KEY") client.run("DISCORD_TOKEN")
7. python3 main.py

If you want to run it background, you can try Screen.

Reference:

Huge thanks to following projects.

https://github.com/TheExplainthis/ChatGPT-Discord-Bot

https://discordpy.readthedocs.io/en/stable/#getting-started

https://developers.generativeai.google/api/python/google/generativeai

License:

This work is under MIT License.