Vector Tiles (MVT) with pg_tileserv, PostGIS, Leaflet and Python Flask

Tired of slow maps? Say goodbye to heavy WMS and hello to blazing-fast vector tiles! In this mini tutorial, we’ll build a lightning-fast vector tile server using PostGIS + pg_tileserv + Leaflet — no GeoServer needed. Fully dynamic, lightweight, and perfect for mobile-friendly GIS apps.

🧠 What is pg_tileserv?

pg_tileserv is a super-lightweight vector tile server developed by CrunchyData. It serves raw .pbf tiles directly from your PostGIS database, no caching or tiling middleware required.

It’s built using Golang, and it outputs tiles in Mapbox Vector Tile (MVT) format, meaning you can render beautiful, dynamic maps in Leaflet or MapLibre with total frontend control.

⚙️ Step-by-Step: Serve Vector Tiles in Minutes

1. Download & Install

   • 🔗 pg_tileserv GitHub

   • Install Go: https://go.dev/dl

   • Check: go version

2. Build & Run

cd D:\pg_tileserv
go build -o pg_tileserv.exe
set DATABASE_URL=postgresql://postgres:admin123@localhost:5432/your_db
pg_tileserv.exe

1. Visit the Interface

   • 👉 http://localhost:7800/

   • See available layers (e.g. public.grid_lines)

2. Render in Leaflet
   Use Leaflet.VectorGrid to render .pbf tiles with full control — color, filter, interactivity — all from the frontend!

   Feature	Raster Tiles (WMS)	Vector Tiles (pg_tileserv)
   Styling	Static (SLD)	Dynamic (JS, CSS)
   File Size	Heavy	Compact .pbf
   Rendering	Server-side	Client-side
   Speed	Slower	⚡ Ultra-fast
   Interactivity	Limited	Popups, filters, hover
   Zoom Levels	Fixed	Fully flexible

   💡 Bonus Tips

   • Add SQL filtering to control tiles (like CQL for WMS) dynamically.

   • Use bounding box parameters for real-time updates.

   • Combine with Flask API for full-stack GIS control.

   ---

   📦 Stack Used

   • PostgreSQL + PostGIS

   • pg_tileserv

   • Python + Flask

   • Leaflet + VectorGrid