Inspiration

I'm not just a student. I'm also a builder, a creator, someone who makes YouTube videos, experiments with side hustles, dives deep into AI, and is constantly exploring anything interesting. The challenge I kept running into wasn't motivation. I had plenty of that. What I lacked was consistency and a clear system to manage everything happening at once.

I'd start strong on a goal, lose track after a few days, and end up stuck in the cycle of planning and restarting. On top of that, I wanted to document my journey publicly but always struggled to find the time or words to do it consistently.

So I decided to build something that solves exactly that problem, for me and for anyone who feels the same way.

What it does

Momentum AI is a goal execution companion that helps you stay consistent, not just motivated. You start by entering your goal, deadline, daily time availability, and experience level. The AI generates a full personalized roadmap with milestones, weekly phases, and first week priorities tailored specifically to your inputs.

Every day you do a quick emotional check-in, selecting from states like Energized, Focused, Calm, Tired, or Overwhelmed. The app tracks your emotional state and uses it to personalize your Sprint Mode experience, including your motivational messages and session framing. Your mood is also factored into the Build-in-Public content generator so your posts reflect how you're genuinely feeling, not just what you accomplished.

Sprint Mode lets you launch a focused pomodoro-style work session with an immersive countdown timer, your top tasks for that session, and an AI generated motivational message tied to your actual goal. When the sprint ends, your Momentum Score increases and a Narrative Memory message appears reflecting on your emotional journey, for example: "Yesterday you were feeling overwhelmed. Today you completed a full sprint. You're rebuilding."

The Build-in-Public generator lets you turn your daily progress into ready-to-post content. Choose LinkedIn, X, or Dev Log format, describe what you accomplished, and the AI produces a properly formatted, natural-sounding post that incorporates your emotional state and progress context. No more staring at a blank page. The dashboard ties everything together with your roadmap, Momentum Score, streak tracking, Today's AI Insight, and Quick Stats all in one place.

How I built it

I used Claude to refine and structure my prompts before bringing them into MeDo, and used ChatGPT for a second opinion on the overall idea and flow. The actual building happened entirely through MeDo using a phased, iterative approach rather than one large prompt.

Phase 1 established the full frontend skeleton, all pages, navigation, and visual design. Phase 2 added AI powered goal decomposition, the roadmap timeline, weekly phases, first week priorities, and the Today's AI Insight card, all powered by a Supabase Edge Function with Gemini 2.5 Flash. Phase 3 added Sprint Mode with the Narrative Memory post-sprint screen, emotional check-in logic and the Build-in-Public content generator.

Every subsequent prompt refined specific components without regenerating the whole app. This kept the architecture clean, the UI consistent, and the credit usage efficient. The multi-turn conversational development made the entire process feel intuitive and surprisingly fast.

Challenges I ran into

Nothing that blocked progress seriously, which itself was impressive. The one thing worth mentioning: I prompted the agent to fix a loading overlay animation that appeared in two different parts of the app. Since both originally shared the same animation logic, I expected the fix to apply to both at once. It only updated one. A small follow-up prompt fixed the second one. Not a blocker at all, just an interesting reminder that being specific in prompts matters even for small fixes.

Everything else went smoother than expected. The multi-page app generation with proper mechanics, the elegant UI, the debugging agent, and the final testing features all worked really well. The multi-agent setup MeDo uses behind the scenes made development feel easy and intuitive in a way I genuinely did not expect going in.

Accomplishments that I'm proud of

The biggest one is that this solves a real problem I actually face. The goal planner, emotional adaptation system, Sprint Mode, and Build-in-Public generator are all features I will personally use after this hackathon.

The second is that I built something this polished in a day or two without formal web development experience. A few years ago this would have felt impossible. The fact that it works, looks good, and does something meaningful is genuinely exciting to me.

What I learned

Prompting is a skill. The quality of what MeDo generates is directly tied to how clearly and specifically you describe what you want. Breaking the build into phases rather than asking for everything at once made a massive difference in output quality and consistency.

I also learned that building for yourself first is the right instinct. When you are solving your own problem, every feature decision feels obvious and the product ends up more focused.

What's next for MomentumAI

Right now I kept authentication out intentionally. I wanted zero friction for new users to try it, and storing data locally means people can share their goals freely without worrying about privacy. If the project gets good traction from the community and judges, I would love to take it further.

The next features I have in mind are a community layer where people working toward similar goals can find each other, share progress, and build momentum together, and a consistency leaderboard that adds a healthy competitive element to staying on track. Auth would come alongside the community feature naturally once there is a reason for users to have persistent identities across devices.

Built With

Share this project:

Updates