A Private Feedback Platform for a Coach — and His Players

Over the last week I built this for my friend Kha and his volleyball coaching business, Volleyball University. It started with an AI tool that watches his game film and files it for his editors automatically — that solved his content problem. But it surfaced a bigger one: Kha had years of coaching feedback locked inside individual videos, with no way to share it — not with the coaches he's training, and not with the players it was meant for.

So we built the other half of the system: a full web app that turns his feedback into two things at once — a training library for future coaches, and private feedback for individual players. Same clips, two audiences, one product.

The upload, done in three clicks

The heart of it is how fast he can turn a coaching moment into a saved, shareable breakdown. He uploads two short clips — his spoken feedback, and a demo of the correct technique — and the app does the tedious part.

The AI — the same volleyball-trained engine from the video pipeline — listens to his feedback and drafts the whole breakdown: a title, the skill, what the player did wrong, what right looks like, and a drill to fix it. He reads it, tweaks the wording, and saves. What used to live only in his head now becomes a structured, searchable entry in seconds.

The hard part wasn't the AI. It was privacy.

Here's the tension that made this a real engineering problem: every piece of player feedback needs to do two contradictory things at once. A player must see only their own feedback. But that same feedback should also feed the coaches' library as a teaching example — without ever revealing which player it belongs to.

Get that wrong and you've either broken a kid's privacy or made the library useless. So the privacy model isn't a feature bolted on top — it's the foundation:

• Players never log in. Each gets a unique, unguessable private link. They tap it and land straight on their own feed — the lowest-friction thing possible for a player or parent.
• Coaches never see who. The library is served through a separate, anonymized layer that structurally cannot expose a player's identity — the name simply isn't in the data a coach can reach.
• The coach controls what's shared. One toggle keeps a sensitive breakdown player-only, off the library entirely.
• Links can be revoked. If a link is ever forwarded or a player leaves, one click issues a new link and instantly kills the old one.

That last 20% — proving no coach and no anonymous visitor could reach a player's identity through any path — is the part that took the real work. It's enforced at the database level, not just hidden in the interface, and it's covered by a suite of automated tests that try to break it on every change.

Built to be his, and to last

A few principles ran through the whole build — the same ones that make automation trustworthy instead of impressive:

• He owns his data. Videos live in his own Google Drive; the app just links to them. No surprise storage bills, no lock-in.
• It's genuinely his brand. A custom "film room" look — his logo, his colors — so it feels like his platform, not a template.
• It ships and holds up. Deployed on a live URL, backed by real accounts and security rules, and hardened by an independent code review before it went out.
• It grows with him. Coach invites, player rosters, a custom domain — all straightforward next steps on the same foundation.

The result

One coach now has a system that spans his entire workflow: film gets understood and filed on its own, feedback becomes a searchable library for the coaches he's building up, and each player gets a private, personal stream of exactly what to work on. Two projects, one connected platform — and a small coaching business that now runs like a much bigger one.

That's the pattern I keep coming back to: the best software doesn't just automate a task. It quietly makes a small business punch above its weight — and then gets out of the way.