DEV Community

juan barrera
juan barrera

Posted on

Kayrol — Day 0: Building AI highlight reels for athletes (in public)

#buildinpublic #saas #startup #ai

I started building Kayrol this weekend. Here's what Day 0 looks like:
the problem, the stack, and the plan for the next 4 weeks.

I'm Juan, a CS student + mid-level dev who is currently an student-athlete. Last year I was exactly where my target customers
are: grinding scholarships, building a portfolio, hoping someone
would notice me.

The Problem

Athletes have HOURS of game footage. Scouts watch 30 seconds. We're
losing talent to bad marketing of ourselves.

That's the gap I'm filling.

What I'm Building

Kayrol — a SaaS that auto-generates highlight reels from raw game
footage using AI + gives coaches/scouts performance analytics per match.
Bilingual ES/EN from day one.

Why it matters:

  • Scholarship recruiting agencies in LATAM handle 50-60 athletes/year, charge $1,500-3,000 per athlete
  • They manually edit highlights → I automate it
  • Athletes pay $25/game or $59/month (vs $400+ Hudl/Veo)

The competition:

  • Athlete AI (USA, multi-sport mediocre, English-only, $5-20/month)
  • Trace ($180-300/year + mandatory hardware)
  • Veo Camera ($2,000+ hardware)
  • Hudl (expensive, built for teams not individuals)
  • Traditional agencies (NCSA, IFX, New Vision: $1,000-3,000 per athlete)

My unfair advantage:

  • Software-only, no hardware required
  • Bilingual from day one
  • Built for soccer specifically
  • I am the customer (Student athlete)
  • 80% cheaper than incumbents

What I Built (Day 0)

Saturday, May 23, 2026 — project started:

  • Next.js 16 + TypeScript (full i18n setup for ES/EN)
  • Supabase for auth + database
  • Cloudflare R2 for video storage
  • Dev environment ready, repo structured for scale

Tech decisions made me ask: Does this stack actually make sense for
video processing at scale? (Feedback welcome on this)

Validation So Far

  • 15 athletes interviewed — pain point confirmed
  • 1 recruiting agency (close friend, 50-60 athletes/year) committed to paying $199 USD for a 30-day pilot with 10 athletes
  • Same agency said they'd pay $500-1,000+/month for full version

The 4-Week Plan

Week 1-2: Auth flow + video upload infrastructure

Week 3: Video processing + basic analytics

Week 4: Landing page + pilot launch with first agency

I'm dedicating 30-40h/week to this. Solo build, no co-founder yet.

What I'm Looking For

  1. Tech feedback: Does Next.js + Supabase + R2 make sense for
    video-heavy SaaS at scale? Better alternatives?

  2. Recruiting agencies interested: 30-day pilot, 10 athletes, $199.
    (Already have 1 committed, looking for 2-3 more)

  3. Athletes/coaches: What would actually make you use something
    like this? What's missing?

The Journey

I'm building this in public. Not because it's trendy, but because:

  • I need accountability
  • Feedback loops matter
  • Founders building something similar can learn from my mistakes

Next post: Auth flow + first video ingest working.

Repo: github.com/jdexbarr/Kayrol (will make public after MVP)

Website: kayrol.com (landing page coming soon)

Let's go 🚀

Top comments (0)