Skip to main content

🤖 Artificial Intelligence in Yume-chan

Yume-chan is the Discord bot with the most AI features in 2026 — 15 AI functions powered by 5+ AI models including GPT-4o-mini, Gemini 2.0 Flash, Llama 3.3 70B, Runware (421+ image models), and Pollinations. From AI image generation and AI-powered moderation to natural language commands and an AI DJ, Yume-chan integrates more AI capabilities than any other Discord bot. 4 features are completely free.

🎯 Feature Summary​

🆓 FREE (4 features)​

FeatureDescriptionAI Model
Basic Image Generation (/image)Image generation with daily limitRunware / Pollinations
Puppy AI Chat (/puppy)Virtual assistant with personalityGemini 2.0 + Llama 3.3
AI DJ Tracking (/music dj)Automatic music trackingGemini 2.0 Flash
NLP Commands (@Yume-chan)Natural language commands by mentioning the botGemini 2.0 Flash + Groq

🌟 STARTER (+1 feature)​

FeatureDescriptionAI Model
Welcome AI BackgroundsAI-generated backgrounds for welcome messagesPollinations / Runware

Total available: 5 features

💎 PRO (+2 features)​

FeatureDescriptionAI Model
Full AI DJ (/music dj)5 smart DJ modes + recommendationsGemini 2.0 Flash
Tournaments AnalyticsPredictive tournament analysisStatistics (AI planned)

Total available: 7 features

👑 ULTRA (+8 features)​

FeatureDescriptionAI Model
AI Auto-ResponsesAutomatic answers with generative AIGPT-4o-mini
AI Context Word FilterContext analysis before penalizing wordsGPT-4o-mini
AI Duplicate DetectionParaphrase-based spam detectionGPT-4o-mini
AI Toxicity DetectionAdvanced toxicity detectionGPT-4o-mini
AI Scam DetectionScam and phishing detectionGPT-4o-mini
AI Raid DetectionRaid pattern analysisGPT-4o-mini
AI Video Generation (/ai video)5s videos with 13 models (Kling, Veo 3, MiniMax, etc.)Runware
AI Welcome VideoAI welcome video with ffmpeg overlay + Components V2Kling AI

Total available: 15 features

🧠 AI Models Used​

OpenAI GPT-4o-mini​

  • Used in: 6 ULTRA features
  • Advantages: High precision, multilingual, advanced reasoning
  • Cost: ~$0.0001-0.0003 per analysis
  • Best practices: Chain of Thought, JSON Schema, Temperature 0

Google Gemini 2.0 Flash​

  • Used in: AI DJ, Puppy AI, NLP Commands
  • Advantages: Free, fast, experimental
  • Cost: $0 (free)
  • Features: Thinking mode, multimodal (future)

Groq Llama 3.3 70B​

  • Used in: Puppy AI (fallback), NLP Commands (fallback)
  • Advantages: Ultra fast (< 1s), high availability
  • Cost: $0 (free)
  • Use: Fallback when Gemini fails

Runware​

  • Used in: Image Generation
  • Advantages: 421+ models, FLUX, Stable Diffusion
  • Cost: $0.0006-0.20 per image
  • Features: Text-to-Image, Image-to-Image, Upscaling

Pollinations.ai​

  • Used in: Image Generation (fallback), Welcome backgrounds
  • Advantages: Free, no limits
  • Cost: $0 (free)
  • Use: Fallback when Runware fails

📊 Distribution by Tier​

FREE ─────────────────────────── 4 features
  │
  ├─ STARTER ──────────────────── +1 (Total: 5)
  │
  ├─ PRO ──────────────────────── +2 (Total: 7)
  │
  └─ ULTRA ────────────────────── +8 (Total: 15)

⚡ Highlighted Features​

🏆 Top 5 Most Complete​

  1. AutoMod AI (ULTRA)

    • 6 AI modules
    • GPT-4o-mini with Chain of Thought
    • Reduces false positives 70-80%

  2. Auto-Responses (ULTRA)

    • Configurable knowledge base
    • Real-time generative responses
    • Complete analytics

  3. AI DJ (PRO)

    • 5 smart modes
    • 15+ detected genres
    • Musical profiles

  4. Image Generator (/image)

    • 421+ available models
    • Multiple styles
    • 4x-8x upscaling

  5. NLP Commands (FREE) (@Yume-chan)

    • Natural language commands
    • Gemini 2.0 Flash + Groq + Regex fallback
    • Supports music, images, economy, pets, games

  6. Puppy AI (FREE) (/puppy)

    • Unique RPG personality
    • 25-message memory
    • Multi-provider with fallback

💰 Estimated Costs​

FREE Tier​

  • AI Cost: $0/month (only free models)
  • Limits: Basic on image generation

STARTER Tier​

  • AI Cost: ~$0.50-2/month (welcome backgrounds)
  • Limits: Increased on image generation

PRO Tier​

  • AI Cost: ~$0.50-2/month (Gemini is free)
  • Limits: Full AI DJ without restrictions

ULTRA Tier​

  • AI Cost: ~$3-10/month (depending on GPT-4o-mini usage)
  • Limits: No limits on AI features
  • Estimate: With 10,000 messages/day = ~$3-10/month
Cost Optimization

All AI systems are optimized to minimize costs:

  • ✅ Fallbacks to free models when possible
  • ✅ Frequent response caching
  • ✅ Configurable thresholds to reduce unnecessary processing
  • ✅ Smart rate limiting

🛠️ AI Architecture​

Backend​

bot/features/
├── auto-responses.js      # AI Auto-Responses (GPT-4o-mini)
├── automod.js            # 6 AI AutoMod modules (GPT-4o-mini)
├── firulais-ai.js        # Puppy AI (Gemini + Llama)
├── ai-dj.js              # AI DJ (Gemini 2.0 Flash)
├── image-generator.js    # Image Gen (Runware)
├── welcome-goodbye.js    # Welcome AI (Pollinations)
├── nlp-commands.js       # NLP Commands (Gemini + Groq)
└── tournaments.js        # Analytics (statistics)

Frontend​

server-panel/src/pages/
├── AutoResponses.jsx     # Auto-Responses Config
├── AutoMod.jsx          # AI AutoMod Config
├── Music.jsx            # AI DJ Config
├── Welcome.jsx          # Welcome AI Config
└── Tournaments.jsx      # Tournaments Config

Implemented Best Practices​

  • ✅ Chain of Thought (CoT): Step-by-step reasoning
  • ✅ JSON Schema Validation: Structured responses
  • ✅ Temperature 0: Maximum consistency
  • ✅ Few-shot Learning: Examples for better accuracy
  • ✅ Multi-provider Fallbacks: High availability
  • ✅ Error Handling: Robust error management
  • ✅ Rate Limiting: Anti-spam protection

🔮 Upcoming AI Features​

In Development​

  • AI Vision Analytics - Real image analysis in tournaments (GPT-4 Vision)
  • Voice Transcription - Voice transcription (Whisper)

Planned​

  • AI Sentiment Analysis - Chat sentiment analysis
  • Smart Translation - Contextual translation
  • Content Summarization - Conversation summaries
  • Image Captioning - Image descriptions
  • AI Moderation Rules Generator - Auto-generate rules

🎓 Feature Guides​

Getting Started​

  1. NLP Commands (@Yume-chan) - Talk to the bot in natural language (FREE)
  2. Puppy AI (/puppy) - Start here (FREE, easy to use)
  3. Image Generation (/image) - Create images with AI
  4. AI DJ (/music dj) - Smart music (requires PRO)

Advanced Moderation​

  1. AI AutoMod - Full configuration (requires ULTRA)
  2. Auto-Responses - Automatic support (requires ULTRA)

Administration​

  • Server Panel: panel.yumechanbot.com
  • Full documentation: This wiki
  • Support: Official Yume-chan server

📈 Project Statistics​

  • Total AI Features: 15
  • Models Used: 5 different
  • AI code lines: ~11,000+
  • Backend files: 8
  • Frontend files: 5
  • Best Practices: 100% implemented

🆘 Support​

Common Issues​

"AI doesn't respond"

  • Check your premium tier
  • Make sure the feature is enabled in the panel
  • Check logs in the server panel

"Incorrect responses"

  • Adjust thresholds in settings
  • Update knowledge base (Auto-Responses)
  • Report the case in the support server

"Feature locked"

  • Verify you have the required tier
  • Go to /premium to upgrade your tier

Contact​

📚 Additional Resources​

Command Consolidation v2

AI commands have been reorganized:

  • /imagen has been consolidated into /image (with subcommands: create, enhance, transform, upscale, random, model, styles, limit, gallery + custom-styles group)
  • Puppy AI is now used with /puppy (subcommands: chat, model)
  • AI DJ is accessed via /music dj (group within /music)
Continuous Updates

This documentation is regularly updated with new AI features. Last updated: 2026-02-25