← Back to projects
Berm

Capabilities

  • Interactive maps
  • GPX and KML import
  • Group rides and events
  • JWT authentication

Frontend

Angular
Ionic
TypeScript
Leaflet

Backend

NestJS
Node.js
PostgreSQL
Neon
Docker
Vercel

Berm

In Progress

Berm started because my MTB routes were scattered across Wikiloc, my watch and folders of GPX files. It imports GPX and KML routes (Wikiloc too, waypoints included), draws them following bike-friendly trails, shows them in 2D and in a 3D viewer with terrain, and stores workouts from FIT and TCX files with their stats and the bike they were ridden on. It is a monorepo with a NestJS API on PostgreSQL with PostGIS, an Angular and Ionic PWA and an Astro landing page, built working with AI agents.

The challenge

A rider wants to plan a route, see it in 3D, know how much each section climbs and share it with their group, and today that takes three or four different apps. The challenge was to bring it together in one that works on a phone, without a shared route giving away where you live.

Screenshots

Implementation

  • A route editor that follows trails

    Free drawing or snapping to trails with BRouter's MTB profile, with pencil, eraser, undo, reverse and close loop; the elevation profile updates as you draw. The toolbar folds and can be dragged around to draw with a finger on an iPad.

  • 3D viewer with a linked elevation profile

    MapLibre with terrain, satellite or topographic layers and an animated flyover. The elevation profile is linked to the 2D and 3D maps: move along the chart and see where you are on the map.

  • Privacy zones

    Draw the areas you don't want to show, like home or work, and public routes are trimmed automatically where they enter or leave them.

  • Explore by layers

    Public routes with search and filters, spots, events and group rides with their meeting point, plus your own markers, all on the same map.

  • Connected to AI agents

    A custom MCP server lets an agent list, create, import, rename and edit the track of routes: a ride can be planned by talking to the AI.

  • Bikes and mileage

    Each bike adds up the kilometres of the workouts and routes ridden on it, and you pick it in the detail of each ride.

Frontend