DiscordBot with GPT Trainer API & AI Image Generation

A Discord bot that integrates with the GPT Trainer API for conversational AI and Google's Gemini for AI image generation. Features optional automated content scheduling that scrapes AI news and monitors YouTube channels.

Features

• Chat Command (/prof): Conversational AI powered by GPT Trainer with your custom knowledge base
• Image Generation (/image): Generate images using Google's Gemini model with aspect ratio options
• Session Management: Per-user conversation history with /reset and /sessioninfo commands
• Gallery Auto-Post: Generated images automatically posted to a forum channel (optional)
• Automated Content Scheduling (Optional):
  • News scraper: Posts AI news articles to a designated channel
  • YouTube monitor: Tracks AI-related YouTube channels and posts new videos
  • All content automatically added to your GPT Trainer knowledge base

Prerequisites

• Python 3.10+ - Download
• Discord Bot Token - Discord Developer Portal
• GPT Trainer API Token & Chatbot UUID - GPT Trainer
• Google API Key - Google AI Studio

Quick Start

1. Clone and Setup

git clone https://github.com/ProfSynapse/discordbot.git
cd discordbot
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Configure Environment

Create a .env file:

# Required
DISCORD_TOKEN=your-discord-bot-token
GPT_TRAINER_TOKEN=your-gpt-trainer-api-token
CHATBOT_UUID=your-gpt-trainer-chatbot-uuid
GOOGLE_API_KEY=your-google-api-key

# Optional - Image Gallery (auto-post generated images to a forum channel)
IMAGE_GALLERY_CHANNEL_ID=your-forum-channel-id

# Optional - Content Scheduling (news + YouTube)
CONTENT_CHANNEL_ID=your-discord-channel-id
YOUTUBE_API_KEY=your-youtube-api-key

3. Run

python main.py

Commands

/prof <prompt>

Chat with the AI. Uses your GPT Trainer knowledge base for context-aware responses.

/prof What are the latest developments in AI?

/image <prompt> [flags]

Generate images using Google's Gemini model.

Aspect Ratio Flags:

• --square (default, 1:1)
• --wide (16:9)
• --tall (9:16)
• --portrait (3:4)
• --landscape (4:3)
• --ultrawide (21:9)
• Also: --2:3, --3:2, --4:5, --5:4

Examples:

/image a cat reading a book
/image cyberpunk cityscape --wide
/image portrait of a robot --portrait
/image panoramic mountain view --ultrawide

Note: 60-second cooldown per user.

/reset

Clear your conversation history to start fresh.

/sessioninfo

View your current session details (session ID, message count, age).

Project Structure

discordbot/
├── main.py              # Bot initialization and core DiscordBot class
├── commands.py          # Slash command handlers (/prof, /image, /reset, /sessioninfo)
├── gallery.py           # Forum gallery auto-posting
├── api_client.py        # GPT Trainer API client
├── image_generator.py   # Google Gemini image generation
├── session_manager.py   # Per-user session persistence
├── config.py            # Environment configuration
├── citation_handler.py  # Citation link processing
├── health_check.py      # HTTP health endpoint for Railway/Docker
├── utils/
│   ├── constants.py     # Shared constants
│   ├── decorators.py    # Error handling decorators
│   └── text_formatting.py  # Message splitting and formatting
└── scraper/
    ├── content_scheduler.py  # News + YouTube scheduling
    ├── news_scraper.py       # RSS feed scraping
    └── content_scraper.py    # Article content extraction

Optional Features

Image Gallery Auto-Post

Set IMAGE_GALLERY_CHANNEL_ID to a Discord forum channel ID. Generated images will automatically be posted as new forum threads with the prompt as the title.

Content Scheduling

Set CONTENT_CHANNEL_ID and YOUTUBE_API_KEY to enable automated content posting. GPT Trainer URL/link sources from automated content are pruned after GPT_TRAINER_SOURCE_RETENTION_DAYS days by default. Failed URL/link sources with configured error statuses are also pruned because they are not usable for retrieval.

Pre-configured Sources:

News (RSS Feeds)	YouTube Channels
TechCrunch	AIExplained
VentureBeat	OpenAI
The Verge (AI)	Google DeepMind
MIT Tech Review	Anthropic
Wired (AI)	SynapticLabs
ArXiv AI papers	GodaGo
Various Substacks	WesRoth

Customize sources:

• Edit scraper/news_scraper.py → RSS_FEEDS dict
• Edit scraper/content_scheduler.py → YOUTUBE_CHANNELS dict

All posted content is automatically uploaded to your GPT Trainer knowledge base.

Deployment

Railway (Recommended)

1. Connect your GitHub repo to Railway
2. Add environment variables in the Variables tab
3. Deploy

See RAILWAY_SETUP.md for detailed instructions.

Docker

docker build -t discordbot .
docker run -d --env-file .env discordbot

Testing

# Configuration tests (free, no API calls)
python test_bot.py

# API integration tests (makes real API calls)
python test_apis.py

Environment Variables Reference

Variable	Required	Description
DISCORD_TOKEN	Yes	Discord bot token
GPT_TRAINER_TOKEN	Yes	GPT Trainer API token
CHATBOT_UUID	Yes	GPT Trainer chatbot UUID
GOOGLE_API_KEY	Yes	Google API key for Gemini
IMAGE_GALLERY_CHANNEL_ID	No	Forum channel for image gallery
CONTENT_CHANNEL_ID	No	Channel for news/YouTube posts
YOUTUBE_API_KEY	No	YouTube Data API key
SESSION_DB_PATH	No	SQLite path (default: /data/sessions.db)
GPT_TRAINER_SOURCE_RETENTION_DAYS	No	Delete old GPT Trainer URL sources after N days; 0 disables (default: 365)
GPT_TRAINER_SOURCE_CLEANUP_INTERVAL_HOURS	No	GPT Trainer source cleanup interval (default: 24)
GPT_TRAINER_SOURCE_CLEANUP_TYPES	No	Comma-separated source types eligible for cleanup (default: url,link)
GPT_TRAINER_SOURCE_CLEANUP_ERROR_STATUSES	No	Comma-separated failed source statuses eligible for cleanup (default: error,error:storage,error:token,fail)
LOG_LEVEL	No	Logging level (default: INFO)

Contributing

Contributions welcome! Open an issue or submit a pull request.

License

MIT License