Skip to content
heyrory

Angling guide and community appOwn product · live

Fish Patagonia

A bilingual guide and community for fly anglers heading to Argentine and Chilean Patagonia, and for the guides, lodges and shops they book. I designed and built it end to end: first in 2024, then again from scratch in 2025.

Visit the project (opens in a new tab)
Role
Own product: architecture and full stack
Year
2024
Status
Live
Stack
  • Full stack
  • AI
  • Next.js
  • PostGIS
  • Offline PWA
[PROJECT SCREENSHOT - 2400×1260]

The brief

Share the fishing without burning the spots

Planning a fishing trip to Patagonia means piecing together scattered information: which permit covers which water, when each season opens, who guides where. Fish Patagonia puts it in one place, with a map of rivers, lakes, species, guides, lodges and shops, river pages grouped by region with permit and season details, weather and conditions, courses, events and a forum. Guides, lodges, shops and transport operators each sign up under a role of their own.

One design problem shaped the community features. Anglers want to share their catches, but a photo can give away a hidden spot, which anglers call spot burning. Uploads are converted on the server, which strips their location data; the original coordinates are stored encrypted and never shown, and the public catch map counts reports per river instead of plotting points. Catch reports can also be logged with no signal and sync once the phone reconnects, without creating duplicates.

Permit rules and season dates differ by region and change over time, so a weekly job reads the official regulation pages and uses Gemini to pull out the rules and season dates, running as a dry run by default and within a daily AI budget. Results are checked against a schema and held for an admin to approve, and the river page carries an AI curated notice until that happens. Underneath is Next.js on Postgres, with PostGIS for map queries, permissions for six roles declared in the schema with ZenStack, and a service worker for offline use.

Outcome

  • 8Regulatory regions the river pages are organised by, across Argentina and Chile
  • 6User roles, including angler, guide, lodge and shop, each with its own permissions
  • 90+Access rules declared in the schema across 31 data models