VLinks

• 

  Splash Screen

• 

  Home page

• 

  Home page

• 

  Choose a "key keeper"

• 

  Create digital vault

• 

  Dashboard

• 

  Create digital vault

• 

  Create profile

• 

  AI Call with deceased's avatar

• 

  Messages sealed in time for specific loved ones

• 

  To keep the essence

• 

  Speak out the unspoken

Inspiration

Why is the funeral industry still in the Stone Age? We live in an era where AI drives our cars, diagnoses diseases, and writes our code. Yet, when we lose someone, technology abandons us. We are left with nothing but a cold headstone and static JPEGs. The inspiration for VLinks came from a simple, painful realization: Death is the only aspect of human life untouched by intelligence.

We wanted to change the narrative of loss. We didn't want to build another photo album app; we wanted to bridge the gap between the living and the departed. We asked ourselves: "In the age of LLMs and Spatial Computing, why can't we say goodbye on our own terms? Why can't a father give a speech at his daughter's wedding, even if he’s no longer physically there?"

What it does

VLinks is the world's first AI-Native Memorial Ecosystem. It transforms static memories into a dynamic Digital Consciousness.

The platform features two core pillars:

Echo AI (Interactive Presence): Users upload voice samples and personality traits of a loved one (or themselves). We process this data to create an interactive AI avatar. This allows grieving families to "speak" to their loved ones again via text or voice, hearing their specific tone, idioms, and laughter.

Time Vault (Future Presence): This allows users to record messages or generate AI interactions for specific future milestones (e.g., a grandchild’s graduation in 2040). These messages are encrypted and time-locked, ensuring the departed can still offer blessings at life’s most important moments.

How we built it

We built VLinks by integrating state-of-the-art Generative AI models with a secure, user-centric frontend.

Frontend: We used Next.js and React to create a serene, comforting, and responsive UI.

The "Brain" (LLM): We leveraged OpenAI's GPT-4o (or replace with your specific model, e.g., Claude/Llama) for the conversational engine. We utilized advanced Prompt Engineering and RAG (Retrieval-Augmented Generation) to inject specific memories and personality traits, ensuring the AI behaves like the specific individual, not a generic bot.

The "Voice" (Audio Synthesis): We integrated ElevenLabs API for high-fidelity voice cloning to capture the emotional nuance of the human voice.

Backend & Security: We used Python/FastAPI for the backend logic and implemented strong encryption for the Time Vault to ensure data privacy and integrity over long periods.

Challenges we ran into

The "Uncanny Valley": Early versions of the voice models sounded too robotic. Fine-tuning the latency and emotional inflection to make it feel "warm" rather than "creepy" was a major hurdle.

Hallucinations vs. Memory: Preventing the AI from inventing false memories while keeping the conversation flowing was difficult. We had to implement strict guardrails to ensure the AI respected the factual history of the person.

Emotional Design: designing a UI for grief is incredibly delicate. We had to iterate many times to ensure the user experience felt supportive and dignified, not transactional or "techy."

Accomplishments that we're proud of

Real Emotional Connection: During our testing, we successfully simulated a conversation with a team member's late relative (using old audio). The moment the AI responded with their specific catchphrase and tone, it was genuinely moving. It proved our concept works.

Seamless Integration: We successfully combined voice cloning and LLMs with low enough latency to allow for a near-real-time voice conversation.

Privacy First: Building a robust architecture for the Time Vault that respects the sanctity of posthumous data.

What we learned

Prompt Engineering is Psychology: We learned that capturing a person isn't just about data; it's about capturing their linguistic quirks and worldview.

Grief needs "Agency": People don't want to replace the dead; they want a way to process the loss. Giving them control over the "Digital Consciousness" aids in that healing process.

AI for Good: We learned that AI's most powerful application isn't just productivity—it's preserving humanity.

What's next for VLinks

Visual Avatars: We plan to integrate Avatar generation (using tools like HeyGen or SadTalker) to add a visual layer to the Echo AI.

Spatial Computing Support: Bringing VLinks to Apple Vision Pro, allowing users to sit in a virtual living room and have a "face-to-face" conversation with the digital presence.

Blockchain for Time Vault: Exploring decentralized storage to guarantee that Time Vault messages remain accessible for decades, independent of any single server failure.

Built With

• netlify
• next.js
• react
• supabase
• tailwindcss
• typescript