Echo

Create lifelike AI avatars of loved ones to preserve their voice, personality, and memories forever. Connect through video, chat, and voice interactions.

Comment

🔊 Echo – Preserve Memories with AI

Reconnecting hearts through digital remembrance

Echo is a human-centered AI platform that lets you recreate and interact with digital personas of loved ones through voice, chat, and lifelike video. Built entirely with Bolt.new, Echo combines realistic avatars (Tavus), emotional voice cloning (ElevenLabs), and memory-based AI conversations to create an experience that feels real, respectful, and deeply personal.

🐙 Why Echo for Octopus Hackathon?

Echo reflects the octopus spirit of intelligent multitasking:

  • Multiple “tentacles” of tech: AI, avatars, voice, video, memories, ethics
  • Built for real-world emotional impact (grief, remembrance, closure)
  • Accessible UX for non-technical users, not just developers

It aligns strongly with the judging criteria:

  • Creativity & Uniqueness – Digital remembrance + AI persona of loved ones
  • Functionality – Working prototype with chat, avatar, and voice
  • Responsiveness – Real-time conversations and video sessions
  • Innovation – Ethical memory modeling + consent-based architecture

✨ Features

Capability Description
🧬 Upload Memories Add photos, audio clips, letters, diary entries, and stories
🧠 AI Personality Engine Learns tone, style, quirks, and emotional context
🎭 Lifelike Avatar Realistic talking avatar generated via Tavus
🔊 Voice Cloning Voice reproduction powered by ElevenLabs
💬 Conversational Chat Talk to your Echo in natural language
📞 Video Call Mode Simulated video calls with real-time emotional responses
⏳ Time Capsule Messages Schedule future messages (birthdays, anniversaries, etc.)

💡 Inspiration

“My grandmother passed away two years ago. I kept wondering—what if I could just call her again? That idea sparked Echo.
It’s not just about remembering someone. It’s about talking to them—again.”

Grief is universal. Echo doesn’t replace people; it preserves their essence through memories and stories.

🛠 Tech Stack

Area Tool / API
Frontend Bolt.new (React)
Avatar Video Tavus API
Voice Clone ElevenLabs
AI Chat OpenAI / Claude
Realtime Video LiveKit
Backend Node.js, Firebase
DB & Storage Supabase, MongoDB

🧪 How It Works

  1. Upload Memories
    Users upload photos, audio clips, letters, chats, and written memories.

  2. Build the Persona

    • Tavus generates a realistic talking avatar.
    • ElevenLabs clones the voice from audio samples.
    • The LLM creates a memory model from the uploaded content (stories, traits, tone).

  3. Interact with Echo

    • Chat using text or voice.
    • Initiate a simulated video call where the avatar talks back.
    • Send or receive time capsule messages on special dates.
Flow:
User → Uploads memories → Persona builder → Avatar + Voice + Memory model → Chat & video interactions

Built With

Share this project:

Updates