Momentum

Inspiration

Through some informal sampling, and through looking at my own experience, I noticed that most people struggle to organise and optimise their days. That's what pushed me to build a website that answers this problem.

What it does

Momentum collects daily data about its user — through short forms, the tasks they plan, and how far they get through their goals. From that data, it learns how that specific person actually works: which hours they get things done, which day of the week is their strongest, and which habits change the rest of their day. For example, it might raise the importance of reading, because statistically its presence improves the user's results and their end-of-day score.

It then acts on what it has learned. It suggests lightening an overloaded day, adding something to an empty one, and it turns long-term goals — running a half-marathon, reading five books — into regular sessions spread across the week.

One rule runs through the whole product: Momentum never changes anything on its own. It proposes, the user decides, and every decision can be undone. And when it doesn't have enough data to know something, it says so instead of inventing an answer.

How we built it

Momentum is built with React 19, TanStack Start, Tailwind and shadcn/ui, and runs entirely in the browser. Everything Momentum measures goes through a single layer that attaches a confidence level to each signal — below a certain threshold, nothing is displayed and nothing is decided.

I used Visual Studio Code with Claude Code and Claude to guide the structure and the prompting of the application, Lovable to get a visual rendering as I made changes, and ChatGPT for the logo.

Challenges we ran into

While developing Momentum, I found it difficult to prioritise the tasks to do, and in which order, so that the result would be as good as possible without wasting time. I also ran into a lot of bugs that cost me a lot of time to spot, diagnose and fix.

Accomplishments that we're proud of

This first hackathon experience made me realise that I am able to develop a working website on my own within a limited amount of time. I'm also happy with the final visual result of the site.

What we learned

Through this adventure I learned how to use the AI agents that dominate the world today, and that you need to master them in order not to be overwhelmed. It gives me the opportunity to give free rein to my imagination and to build many innovative projects in the future.

What's next for Momentum

After this hackathon, I'd like to finish the website properly, make it fully operational, and potentially start sharing it with people around me and expanding it from there.

Built With

• chatgpt
• claude
• claude-code
• css
• git
• github
• i18n
• localstorage
• lovable
• radix-ui
• react
• shadcn-ui
• svg
• tailwindcss
• tanstack-start
• typescript
• visual-studio-code
• vite
• vitest