SmartFIT Intelligence
Describe a ride or run in plain language, review the parsed route and performance plan, compare it with the active Athlete Twin, and generate with explicit targets rather than hidden assumptions. Explore SmartFIT Intelligence.
SMARTFIT 1.8 · ADAPTIVE ACTIVITY STUDIO
Build activities that behave like an athlete, not a spreadsheet.
Create cycling and running activities from mapped roads, trails or exact GPX tracks. SmartFIT can learn from each athlete's latest 5–25 genuine activities to adapt behavior, recognize repeated route corridors, build a private personal heatmap, and keep route-specific cycling or running patterns separate.
Training Session Builder
Build route-fitted cycling and running recovery, endurance, tempo, threshold, interval and progressive sessions with simulated pace, HR, cadence and power behavior. Explore training sessions.
Adaptive Activity Library
Import genuine FIT/GPX/TCX activities into a private per-athlete library. Choose a latest 5, 10, 15, 20 or 25 activity learning window for the Athlete Twin, repeated Route Twins and Personal Heatmap 2.0.
Per-athlete Activity Twin
Learn cycling and running behavior separately, with recency weighting for HR recovery, cadence, coasting, power or pace, trajectory movement and recording style.
Per-athlete connected accounts
Strava remains a same-user display-only connection and is excluded from SmartFIT learning.
Passkey-first accounts
Sign up and sign in without a site password. Device user verification protects access while athlete profiles and saved projects remain separated by account.
FIT & GPX studio
Edit, repair and convert FIT data; preserve sensor-only records during GPS loss; import exact GPX tracks; generate route-aware cycling and running activities.
FIT Tools Suite
Use focused utilities for time adjustment, combining split recordings, field merging/stripping, corrupt-time repair, break and peak cleanup, elevation changes, sport/device changes and conversion. Explore FIT tools.
Adaptive cycling and running activity generation with route-aware human behavior
SmartFIT models a legal route, a physical rider or runner path and the final GNSS measurement as separate layers. This lets repeated loops vary naturally without allowing GPS noise to redefine the activity distance. Cycling logic supports coasting and power behavior; running logic uses pace, step cadence, gradient and runner trajectory rules.
When a route corridor is recognized, SmartFIT can use its Route Twin before the generic Athlete Twin. The private heatmap is rebuilt from the included learning window and stays separated by account and athlete. Existing activity tools remain part of the same project: FIT inspection and repair, timestamp editing, supported sensor-stream changes, GPX/TCX conversion, route building, device presets, lap generation, personal trajectory learning and activity export.