A health tracking app for a busy loved one
This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend My girlfriend is great: she's a UI/UX designer who's great at her job, a passionate artist who inspires me, and just an all around fun person who I love ardently. She also doesn't take care of herself — too busy for tha

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend My girlfriend is great: she's a UI/UX designer who's great at her job, a passionate artist who inspires me, and just an all around fun person who I love ardently. She also doesn't take care of herself — too busy for that. I think we all have been here at some point (some point being the entirety of our college life + professional career). Lumen's goal is simple: help my girlfriend feel better by doing small things to boost her health, mood and make her day better by the cumulative action of healthy habits. Lumen is an app built on Typescript — React Native with Expo, and Node.js. The goal was simple, build an app that's so easy to use there would be no excuses for not tracking. A bit more about Lumen: It is a tracker. Not a diagnosis solution, or a replacement for medical advice. It is an app where one can track their health, moods, exercise activity, and keep a record of it. It, quite importantly, also has an AI tool that gives you suggestions to help you recover from a low spell, or keep yourself steady if you're on the recovery path. I wanted my girlfriend to spend a minute or two on the app every day, and whenever she wants, it provides her with real, achievable tasks to boost her energy and keep her going. The app is also offline first, helping her use it whenever she has time and syncing it when the server is up (more about this later) Here's a video and some screenshots from the app. https://github.com/anuraggopalakrishna/lumen How I Built It I used Opencode + Deepseek v4.1 to plan the tasks out, and used Deepseek v4.1 Flash and Muse Spark 1.3 to build the app. The core of the app is the suggestions module. I used Qwen3 4b for this, with local inference. This is sensitive data, data that I do not want handed over to any corporation or anywhere outside trusted sources (i.e my machine, where it is safe and can be wiped on demand). With this, and completely local hosting, the data is closed, while the tech is open. The app is completely hosted on my machine, so one text message and all the features are up :P Data security. It's the heart of this project and without open innovation we just cannot have it. There is no true self-built self-purposed software without Open source, and even in the age of AI it rings true. Will update with feedback soon :)
Key Takeaways
- •This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend My girlfriend is great: she's a UI/UX designer who's great at her job, a passionate artist who inspires me, and just an all around fun person who I love ardently
- •This story was reported by Dev.to, covering developments in the dev space.
- •AI advancements continue to reshape industries — read the full article on Dev.to for complete coverage.
📖 Continue reading the full article:
Read Full Article on Dev.to →

