Walker

Real-time treadmill tracking. Connect a Bluetooth walking machine, track your walks honestly, and compete on a live leaderboard.

Production: https://walker.akerud.se

What It Does

Walker connects to your Bluetooth treadmill, detects when you're actually walking (not just standing on the belt), and computes honest calories using peer-reviewed biomechanics models. Data streams to a server that powers a live dashboard with leaderboards, profiles, heatmaps, and history.

Steps detect, speed measures, server computes.

Quick Start

walker login                 # authenticate (opens browser)
walker walk                  # connect to treadmill and start tracking

Open the dashboard at https://walker.akerud.se to see your stats, compete on the leaderboard, and explore your walking history.

Supported Devices

Device	Protocol	Status
UREVO SpaceWalk E1L (URTM041)	Proprietary BLE (0xFFF0)	Supported
FTMS treadmills	Bluetooth FTMS	Planned

CLI Commands

walker login [--dev]       # authenticate (opens browser)
walker logout [--dev]      # remove credentials
walker walk [--dev]        # connect to treadmill and report
walker simulate [--dev]    # simulate walking without hardware
walker enumerate           # scan for BLE treadmills
walker set-weight 78       # set weight in kg
walker listen [--dev]      # run the server

How Calories Work

Walker uses the Ludlow–Weyand (2017) minimum mechanics model for walking and the Minetti (2002) polynomial for running, dispatched at the 7 km/h gait transition. Calories are computed from speed, incline, weight, and duration — never stored, always calculated fresh. This means formula improvements apply retroactively to all historical data.

Only active kcal is shown — the exercise-only number above resting metabolic rate. Resting calories are deliberately excluded: they inflate the number without reflecting effort, which turns leaderboards into a "who sat on the belt longest" contest.

Architecture

Rust server + vanilla JavaScript dashboard. No frameworks, no build step.

• Client sends speed + walking/idle state to the server via HTTP
• Server manages segments (continuous periods of walking or idle) in PostgreSQL
• Dashboard is a single-page app served as static files, themed with CSS variables
• WebSocket pushes live updates to connected viewers

See CLAUDE.md for the full spec — it's the source of truth for how everything works.

Environment Variables

Variable	Description
DATABASE_URL	PostgreSQL connection string (defaults to local Postgres in --dev)
WALKER_BASE_URL	Base URL for OAuth callbacks
WALKER_GITHUB_CLIENT_ID	GitHub OAuth App client ID
WALKER_GITHUB_CLIENT_SECRET	GitHub OAuth App client secret
WALKER_GOOGLE_CLIENT_ID	Google OAuth client ID
WALKER_GOOGLE_CLIENT_SECRET	Google OAuth client secret

Local Development

./reset_db.sh                        # fresh Postgres in Docker
cargo run -- listen --dev             # server with seed data at localhost:3000
cargo run -- login --dev              # click "Dev Login" in browser
cargo run -- simulate --dev --count 5 # simulate walkers

Deployment

Production runs on Render.com. The Dockerfile builds a server-only binary (no BLE dependencies) with dependency caching.

License

MIT