This is a submission for the "Hacktoberfest Open-Source AI Challenge Week 4: Ghost in the Machine" (https://dev.to/challenges/hacktoberfest-week4-2026-10-26)
👻 The Last Login — Your Memories. Your Wishes. Your Control.
What I Built
Imagine having thousands of photographs, personal documents, voice recordings, letters, and important memories stored across your devices.
But what happens to all these digital belongings when you are no longer able to manage them?
Who will know what to preserve, what to share, and what should remain private?
That question inspired me to build The Last Login, a privacy-first, local AI-powered digital estate planner.
The Last Login helps people organize their digital belongings, preserve the stories behind them, document personal wishes, and prepare an encrypted digital legacy archive for the future.
I interpreted the Ghost in the Machine theme as the digital traces we leave behind — the memories, stories, and personal information that continue to exist in our devices.
✨ Key Features
- Life Inventory: Organize and manage important digital belongings.
- Living SĂ©ance: Use local AI to help document the stories and memories behind personal items.
- Keep, Release & Exorcise: Decide what to preserve, prepare for sharing, or exclude.
- Legacy Letters: Create personal letters, stories, and instructions with AI assistance.
- Sealed Ofrenda: Create an encrypted archive of selected legacy content and restore it later.
- Memorial Preview: Preview only the content explicitly approved by the user.
Our philosophy is simple: AI assists, but humans decide.
We don't recreate or impersonate deceased people. We preserve real memories and personal wishes.
Demo
🚀 Live Application: [https://the-last-login-green.vercel.app/]
🎥 Demo Video: [PASTE YOUR DEMO VIDEO URL, IF AVAILABLE]
Code
đź’» GitHub Repository: [https://github.com/rudrarajsinh920/The-Last-Login]
The project is built with a modular architecture using React, TypeScript, Node.js, Express, and SQLite.
How I Built It
I built The Last Login using modern open-source technologies, with a strong focus on local AI processing, privacy, and user control.
AI — Gemma through Ollama
I chose Gemma because I wanted to explore running AI locally instead of depending entirely on external cloud AI APIs.
Using Ollama, the application can run a compatible Gemma model on the user's machine.
I use local AI to assist with documenting memories and generating personal content.
When the local model is unavailable, supported operations can use an offline fallback.
Technology Stack
- Frontend: React, TypeScript, Vite, Tailwind CSS
- Backend: Node.js, Express.js, TypeScript
- Database: SQLite with FTS5
- AI: Gemma through Ollama
- Encryption: AES-256-GCM and PBKDF2-SHA256
- Validation: Zod
Security and Privacy
For the encrypted archive, I implemented AES-256-GCM encryption and password-based key derivation using PBKDF2-SHA256.
Users select what they want to include in their archive and remain responsible for protecting their passwords and recovery information.
The application is designed around local-first processing and explicit user decisions.
Why Does Open Innovation Matter?
Open innovation made it possible for me to experiment with local AI rather than building the entire experience around a closed cloud API.
Using Gemma and Ollama helped me explore how AI-powered applications can work closer to the user's own device.
It also gives developers more flexibility to experiment with models, customize AI workflows, and build applications without making every AI interaction dependent on an external paid API.
For a project dealing with personal memories and sensitive digital belongings, that flexibility matters.
I believe open innovation makes it possible to build technology that gives users more control over their data and how AI interacts with it.
My Agent Session
[OPTIONAL: ADD YOUR DEVRELAY AGENT SESSION LINK OR EMBED YOUR SESSION HERE]
The Spirit of It
The most meaningful part of building The Last Login was thinking about the stories behind ordinary digital files.
A photograph might look like just another image in a folder, but for someone, it could represent an important family memory.
A voice recording might contain a story that deserves to be preserved.
A personal letter might carry wishes that someone wants their loved ones to understand.
I wanted to build something that respects those memories instead of turning them into another dataset for AI.
The Last Login is not about bringing people back through AI. It is about helping people preserve what matters while they are here to make those choices themselves.
Your memories. Your wishes. Your control.
Prize Categories
[ADD THE ELIGIBLE PARTNER PRIZE CATEGORIES YOU ARE ACTUALLY ENTERING, BASED ON THE OFFICIAL CHALLENGE RULES.]
Built with curiosity, open-source AI, and the belief that our digital lives deserve thoughtful preservation.
Top comments (1)
The “humans decide” rule is especially important here: the archive should preserve an owner’s choices, not infer a legacy policy from their files. Since recovery is the hard part of encrypted personal archives, it would be valuable to make the sealed archive format and restore test explicit—versioned metadata, an integrity check, and a guided export that the owner can verify before they need it. That turns encryption into a durable promise rather than only an implementation detail.