Running Score

Running Score Dashboard
Analyze Activity
Have scores about your last run
Get AI comments and suggestions
And of course profile & theme support

Inspiration

I (the author of the application) is often told I have several postural and stride issues during runs, however applications often are just collecting tons of statistics and don't have inputs nor analytics for real-world run sequences pictures.

What it does

Running Score enables to upload photos of a run session and screenshots / exports of activities from usual applications to have an analysis and improvement suggestions from an AI.

How I built it

Bolt.new primarily generated code over recurring prompt iterations, fine tuning & debug done manually using usual IDE and usage of Supabase AI assistant + Microsoft Copilot for Supabase related operations and generation of some texts and images for the application.

Challenges I ran into

To be honest, besides the large number of tokens used even for simple prompts as soon as the project grows a bit, there were some small pebbles and dumb mistakes to fix in AI generated content but throughout the challenge but it was mostly very smooth.

AI was used to generate all structural, skeletons, initial contents for each component / asset and manual edits were done only for localized modifications (wasting tokens in prompts to fine tune a CSS property or portions of a few lines of stupid and bugged code is not worth it).

Accomplishments that we're proud of

As I started the application for my own need only in the second half of the challenge and worked casually on it in just a few sessions throughout the two remaining weeks, I didn't think at all I would build that much of an extendable platform around my initial idea and cover it with proper documentation and tests.

Prior to use AI tools it would have taken me several weeks to do what I did in just a few days and it would have been of lesser quality because for pet/side projects many aspects I was able to cover on Running Score would have been discarded and ignored as using too much time.

What I learned

Primarily a refreshing experience, a Stark workflow where iterating on apps by vocally specifying prompt commands and have a visual feedbacks while we're away of the keyboard might not be so far :)

What's next for Running Score

The version for the hackathon will be frozen as Running Score while a continuation of the development will be published on a more adequate infrastructure with more features to implement and more open testing.