Based on the codebase, here's a detailed specification of SmolBot:
## Project Overview
SmolBot is a sophisticated Discord bot designed to engage in natural conversations with a unique blockchain/crypto-focused personality. It features advanced image processing capabilities and maintains contextual awareness across conversations.
## Core Features
### Response Triggers
#### Mention or direct reply:
- Users can @mention the bot, or reply to the bot's previous message
- Bot processes the message and any attached images
- Generates contextual response considering previous messages (Conversation History Management)
#### Image Interactions:
- When mentioned with images, provides detailed analysis
- Can reference previously shared images in conversation
- Combines image context with conversation flow
- A brief analysis is provided for all images, and a more detailed analysis is only provided when the image is directly mentioned to the bot
### Image Processing System
#### Two-Tier Analysis:
- Brief Analysis: Quick 1-2 sentence descriptions for all images
- Detailed Analysis: Comprehensive analysis when directly mentioned
- Supports multiple image formats
- Maintains image context within conversations
### Conversation History Management
- Maintains a rolling message history (up to 20 messages per channel) to provide context for responses
- Tracks message references to understand conversation flow
- Processes both direct mentions (@SmolBot) and reply-based interactions for flexible engagement
- Efficiently manages separate conversation contexts across multiple channels
- Implements Discord typing indicators during response generation to show activity
- Uses message history to inform AI responses with relevant context
- Cleans and formats message history before sending to AI model
### Emoji System
- Strict emoji validation and formatting
- Supports both static and animated Discord emojis
- Case-insensitive emoji matching
- Automatic emoji formatting and cleanup
### AI Integration
- Uses Groq API for text and vision processing
- Multiple model fallback system:
- Primary: llama-3.1-70b-versatile
- Fallback: llama-3.2-3b-preview
- Emergency Fallback: llama-3.1-8b-instant
- Intelligent retry mechanism with exponential backoff
- Timeout handling for API calls
### Technical Features
- Robust error handling and logging
- Message cache system
- Automatic channel monitoring
- Event-based architecture
- Configurable system messages
- Typescript/Node.js implementation
### Performance Features
- Timeout handling for API calls
- Exponential backoff for retries
- Memory-efficient caching
- Optimized image processing
- Concurrent request handling
## Technical Requirements
- Discord.js for bot interactions
- Groq API for AI processing
- Node.js runtime
- TypeScript support
- File system access for configuration
- Environment variable configuration
- Logging system (Pino)
## Integration Points
- Discord API
- Groq API
- File System
- Environment Configuration
- Channel Cache System
- Emoji Management System
This bot is particularly notable for its sophisticated handling of context, images, and emojis, while maintaining a consistent personality. The multi-tier fallback system and robust error handling make it particularly resilient to API issues or rate limiting.typescript
