yasuhiroinoue/gemini_discordbot_vertexai

A Discord bot powered by Google Gemini Pro, capable of text generation, image analysis, audio transcription, and more.

Gemini Discord Bot Using Google VertexAI

Introduction

Want to supercharge your Discord server with the mind-blowing power of Google's Gemini Pro? This bot is your portal to cutting-edge AI, right at your fingertips. Imagine being able to:

• Generate awesome stories, poems, or even scripts for your next Discord roleplay.
• Analyze images shared in your server, instantly extracting text or getting cool image descriptions.
• Transcribe audio messages, so you can easily keep up with every conversation.
• Get AI insights and assistance on your code snippets, no matter how long they are!

Features

• Text generation using Gemini Pro: Unleash your creativity and write stories, poems, scripts, musical pieces, emails, letters, and anything else you can dream up!
• Handles various input formats (using Gemini 1.5 Pro): Let the bot analyze images, extract text from PDFs, transcribe audio, and more.
• Accepts file input for processing text beyond Discord's character limit: Have a super long code snippet you need help with? No problem! This bot can handle it.
• Image generation from text prompts: Create stunning visuals based on descriptive text.

Supported Input Formats

This bot welcomes a variety of formats, including:

• Text-based documents
• Data files
• Source code
• PDFs
• Images
• Audio files
• Video files

Prerequisites

Before you start, you'll need:

• Python 3.8 or newer
• Discord Bot Token
• Google Cloud Platform account
• VertexAI enabled on your GCP project

Setup Instructions

Ready to get started? It's easy!

1. Clone this repository to your local machine or server.
2. Install the required dependencies by running pip install -r requirements.txt.
3. Create a .env file in the root directory of the project and populate it with your environment variables as described below.
4. For image generation, ensure that the appropriate libraries and APIs (Imagen3) are enabled and configured in your GCP project.

Technical Details

This bot is built for flexibility! It uses the gemini-1.5-pro-001 version of Google's Gemini Pro model by default, but you can easily switch to a different version (check out the MODEL_ID in GeminiDiscordBot.py). Oh, and by the way, this README.md? Yeah, Gemini 1.5 Pro helped write this too!

Environment Variables

To get this bot up and running, you'll need to add these environment variables to your .env file:

• GCP_REGION - Your Google Cloud Platform region
• GCP_PROJECT_ID - Your Google Cloud Platform project ID
• DISCORD_BOT_TOKEN - Your Discord Bot Token
• IMG_COMMANDS_ENABLED # True to enable, or False to disable !img and !fimg commands

Other Customization Options

Want to tweak things even further? Go for it! You can customize the bot's behavior by exploring the source code. For example:

• Adjust the safety settings: Control the level of content filtering applied to the bot's responses.
• Modify the text generation parameters: Play around with settings like temperature and top_p to change how creative and coherent the generated text is.
• Integrate additional commands: Add your own custom commands to make the bot even more powerful!

Using Image Generation

To generate images based on text prompts, use the following command in your Discord server:

For standard image generation:

!img <prompt> | <negative_prompt> | <aspect_ratio>

For faster image generation:

!fimg <prompt> | <negative_prompt> | <aspect_ratio>

Replace <prompt> with a description of the image you want to create, <negative prompt> with a description of things you don't want in the image, and <aspect_ratio> with the desired aspect ratio (e.g., 16:9, 4:3). The bot will respond with the generated image based on the provided text prompts and aspect ratio.

Note: The !fimg command uses a faster model that may have slightly lower image quality compared to the !img command.

Running the Bot

1. Start the Bot (On your server)

To start the bot on your server, open your terminal and run this command:

python GeminiDiscordBot.py

2. Reset Conversation History (Within Discord)

Want to start a fresh conversation? Send this command in a Discord channel where the bot is active:

reset

Contributions

Got some awesome ideas to make this bot even better? We'd love to hear them! Feel free to open an issue or submit a pull request.

License

This project is licensed under the MIT License - see the LICENSE file for details.