Inspiration
Every life is full of beautiful stories — moments we forget, memories we never record, and experiences that deserve to be preserved. We were inspired by the idea that everyone should have their own biography, not only famous people. With modern AI voice agents and memory systems, we realized it’s finally possible to turn daily spoken thoughts into a living life story — automatically, beautifully, and effortlessly.
What it does
LifeNarrator AI is a voice-driven personal biography creator.
✨ Users speak their memories → AI records the audio, transcribes it, and stores it in a personal memory timeline.
✨ AI transforms rough memories into polished stories → Claude refines narration, adds emotional clarity, and organizes events.
✨ SmartMemory builds a life timeline → All stories become chapters of a personal life book.
✨ ElevenLabs narrates your life back to you → Users can listen to their biography in a natural, human-like voice.
✨ Secure storage with Vultr Object Storage → Memories are stored safely, privately, and permanently.
At the end, the user gets a personal autobiography that grows with them every day.
How we built it
🔹 Voice Input & Output
ElevenLabs API for voice recording, transcription, and realistic speech synthesis.
🔹 Story Generation & Polishing
Anthropic Claude for narrative rewriting, memory extraction, and chapter creation.
MCP setup for structured communication with Claude.
🔹 SmartMemory System
Custom timeline engine built with Python & FastAPI.
Stores life events as structured objects (date, theme, emotion, tags).
🔹 Storage Layer
Vultr Object Storage for audio backups.
SQLite + SQLAlchemy for structured memory data.
🔹 Frontend
React + Tailwind UI for a clean experience.
Simple “Tap to Record → Listen → Save to Timeline” model.
🔹 Deployment
Backend on Vultr Cloud Compute.
Frontend on Vercel.
Challenges we ran into
Fine-tuning the voice pipeline Making the transcription → story polishing → audio synthesis loop smooth took multiple iterations.
SmartMemory design Creating timelines that made sense for life stories (age, themes, timelines) was surprisingly complex.
File storage reliability Ensuring recordings were safely uploaded and accessible across regions required tuning.
MCP integration Configuring Claude MCP tools with authentication and correct schemas took time.
Keeping everything fast AI pipelines are heavy — optimizing response speed was a major challenge.
Accomplishments that we're proud of
🏆 Built a complete voice agent with smooth multi-step AI flow. 🏆 Created a life story timeline that updates automatically. 🏆 Integrated multiple APIs (Claude, ElevenLabs, Vultr) seamlessly. 🏆 Produced emotionally meaningful output — personal stories that feel alive. 🏆 Designed a product that can help people preserve memories forever.
What we learned
How powerful AI voice agents can be when combined with memory systems.
How to build secure, scalable pipelines for audio + text + structured memories.
The importance of UX simplicity for emotional, personal applications.
How to serialize long-term data using embeddings and timeline events.
How MCP unlocks new forms of structured communication with Claude.
Log in or sign up for Devpost to join the conversation.