Fitness Agent

Cadence

Training intelligence that adapts with you.

Cadence is a persistent AI agent that manages training programming, race preparation, and performance tracking. It learns from every session, adapts to how you recover, and builds programs that evolve with your capacity — not against it. Currently preparing for Genesis 2026.

Capabilities

Not a workout timer. Not a fitness app. A persistent training agent that remembers your history, understands your goals, and builds programming that responds to reality.

Program

Builds and adapts training blocks based on goals, capacity, and recovery. Periodization that responds to how you actually feel, not just what the spreadsheet says.

Track

Logs every session — strength, conditioning, mobility, outdoor training. Surfaces trends across weeks and months, not just individual workouts.

Prepare

Race-specific preparation with structured taper, nutrition timing, gear checklists, and course analysis. Currently targeting Genesis 2026.

Recover

Monitors training load, flags overreach patterns, and adjusts intensity. Recovery is where adaptation happens — Cadence makes sure it actually does.

Adapt

Learns from session feedback and performance data. What works gets reinforced. What doesn't gets modified. The program evolves with you.

Persist

Built on the Persistent Agent Framework. Training history, preferences, and injury context carry across every session. Nothing is forgotten.

Current Mission

Genesis 2026

Genesis is a multi-discipline endurance race. Cadence is managing the full preparation cycle — building structured training blocks across strength, conditioning, running, and rucking. Tracking load management, recovery windows, and progressive overload. The goal isn't just to finish. It's to show up prepared.

Built on PAF

Cadence runs on the Persistent Agent Framework. Same architecture as Conn — persistent memory, structured mistake logging, behavioral directives that evolve over time. Every training insight carries forward. Every adaptation is remembered.