SummitMind: Finishing a Local-First Memory App for the GitHub Finish-Up-A-Thon

#githubchallenge #devchallenge #githubcopilot #webdev

GitHub “Finish-Up-A-Thon” Challenge Submission

This is my submission for the GitHub Finish-Up-A-Thon Challenge.

What I Built

SummitMind is a local-first memory app for understanding thoughts, decisions, projects, and recurring patterns.

Most note apps help you store information. SummitMind focuses on the next step: helping you notice what keeps coming back.

The app runs in the browser with no account, no backend, and no tracking. Memories stay in localStorage unless the user exports them.

Live demo:
https://p-r-e-m-i-u-m.github.io/SummitMind/?demo=1&report=1

Repository:
https://github.com/P-r-e-m-i-u-m/SummitMind

Before

The old version had a good idea, but it still felt unfinished.

It could:

save memories
search the archive
group memories by month
show simple pattern cards
import and export JSON

But the product loop was incomplete. You could capture a memory, but you could not comfortably revise it, mark it as important, filter by theme, or turn the archive into a review.

After

The revived version now supports a complete reflection flow:

Capture a memory.
Edit it when the context changes.
Pin memories that still matter.
Filter the archive by generated tags.
Generate a private monthly reflection report.
Copy the report as Markdown or print the archive.

I also added:

demo data for quick evaluation
a GitHub Pages live demo
a demo screenshot and judge guide
keyboard shortcuts for search and save
accessibility polish with a skip link and live status updates
print styles
CI checks for formatting and JavaScript syntax

How GitHub Copilot Helped

GitHub Copilot helped me turn a vague unfinished roadmap into concrete implementation steps.

The most useful parts were:

breaking the revival into small product improvements
refactoring the browser-only JavaScript without adding a framework
adding edit, pin, tag filtering, and report behavior while preserving old saved data
improving UI copy so the product felt calm and human
shaping the before/after story for the challenge submission

I still reviewed the code manually, tested the app in Chrome, generated screenshots, and kept the scope intentionally small.