• 

  Avatar

• 

  Application screenshot

• 

  Video1

• 

  Video2

• 

  Video3

• 

  Waifu

🚺 Waifu Maker – AI Voice Generation & Chat

Waifu Maker is a modern web application that brings anime-style virtual companions to life. It combines ElevenLabs voice synthesis with Mistral AI for intelligent, emotionally expressive conversations—featuring a 2D animated avatar that talks and reacts in real time.

💡 Inspiration

We wanted to create a virtual companion that feels emotionally present—someone who can talk to you, remember you, and feel alive. Inspired by anime waifus and voice-based AI, we fused the best of both worlds: cutting-edge voice tech and interactive avatars.

🛠 What It Does

• 🔊 Real-time voice generation with ElevenLabs
  
• 🧠 Natural AI conversations via Mistral
  
• 🖼️ Lip-synced 2D animated avatars that "talk" as they speak
  
• 🧩 Custom personalities with memory and emotional tone
  
• 📜 Full conversation history with voice playback
  

🧱 How We Built It

• Frontend: React with modern UI (glassmorphism, dark mode, responsive design)
  
• Backend: Flask API handling voice generation and AI chat
  
• Voice: ElevenLabs API for high-quality TTS
  
• AI Chat: Mistral AI for natural language dialogue
  
• Avatar: GIF-based avatar system with dynamic lip-sync logic
  
• Sync: Smart GIF selection based on message length (short/long speech)
  

🧗 Challenges We Ran Into

• Syncing avatar lip movements with real-time speech
  
• Managing timing between TTS output and GIF animation
  
• Handling API latency while keeping the app responsive
  
• Balancing emotional tone in conversation and voice
  
• Designing a UI that feels sleek yet cozy and anime-inspired

🏆 Accomplishments We’re Proud Of

• Achieved smooth, believable avatar lip-sync using only GIFs
  
• Created a modular personality system for AI customization
  
• Seamless integration of voice, chat, and visual feedback
  
• Built a compelling user experience in a short development cycle
  

📚 What We Learned

• The power (and quirks) of ElevenLabs' real-time TTS
  
• How to create engaging, expressive UIs with minimal animation libraries
  
• Nuanced design challenges of AI-driven emotional interaction
  
• Ways to map voice, text, and visual feedback into a unified user experience
  

🔮 What’s Next for Waifu Maker

• 🎭 Custom avatar uploads and styles
  
• 🧠 Long-term memory and evolving personality
  
• 📱 Mobile app version
  
• 🗣️ Voice input for full back-and-forth conversations
  
• 🧑‍🤝‍🧑 Multi-character support (harem mode, obviously)
  
• 💕 Emotional state tracking and adaptive responses
  

Built With

• elevenlabs
• fal.ai
• flask
• mistral.ai
• react.js