KinesteX

Application Lifecycle

Events for startup, loading and leaving KinesteX.


EventData FieldsDescription
kinestex_loadeddate: string (Date.toString(), e.g. "Thu Oct 29 2026 10:00:00 GMT+0000")The page has loaded and is listening. Send (or re-send) your launch payload now. It does not mean verification or the camera is ready. Once per page load
kinestex_launcheddata: string ("DD MM YYYY HH:MM:SS", device-local)Your launch payload was verified; KinesteX is starting up. Sent after every successful verification
exit_kinestexdate: string (ISO 8601, UTC), time_spent: string ("HH:MM:SS")The user closed KinesteX from a top-level screen: back or close on the workout or plan page you launched, a home page, the leaderboard with an empty backUrl, plan onboarding, the AI Trainer chat, a challenge or assessment intro, an error page. Close your KinesteX view. With exitUrl in your launch payload, KinesteX also redirects there
exit_kinestex (results fallback)data: { workout: string, date: string }Rare: the user left a results screen and KinesteX could not tell which integration launched it. Treat it like any exit_kinestex
main_page_openeddate: string (ISO 8601)The Home page opened. Sent just before home_page_opened
home_page_openeddate: string (ISO 8601)Home Page integration screen opened. Fires exactly once per mount (the legacy KinesteX home: once it finishes loading)
experience_openedexerciseId: string, exerciseTitle: string, isGame: booleanUser tapped a challenge or game card on the Home Page. Followed by challenge_opened or game_opened
challenge_openedexerciseId: stringThe tapped Home Page card is a challenge
game_openedgameId: stringThe tapped Home Page card is a game
personalized_plan_openedreferral: "Home"User opened the personalized plan from the Home Page

exit_kinestex can also come from the SDK itself, without date or time_spent (see Receiving Data), so read time_spent as optional.


Loading and readiness


EventData FieldsDescription
models_loadedmessage: string, modelIds?: string[]Motion models are ready. With modelIds (message "Successfully loaded N model(s)"): a model download finished with at least one model loaded, for the Camera component's launch exercises or your load_models command. modelIds lists the ids that loaded, as you sent them. Without modelIds (message "All models loaded"): Camera component, every model in exercises is loaded. A Camera component launch can get both
model_warmedupmessage: stringCamera component only: the pose tracking engine is ready. Sent each time the camera page opens, so again after the user returns from the camera selector
speech_fetch_completesuccessCount: number, failureCount: number, modelIds?: string[]Spoken audio finished downloading. Without modelIds: the rest-speech phrases (Camera component restSpeeches, workout preload, Custom Workout and AI Trainer workouts); not sent when every phrase was already cached. With modelIds (and failureCount: 0): the mistake-feedback audio of the models just loaded, downloaded in the background, so it can arrive after models_loaded
warningdata: string, or message: stringA notice that blocks nothing. data: "No model IDs provided": the Camera component was launched without exercises (or with an empty list). message: Camera component bookkeeping edge cases (a rep or mistake arrived with no current exercise). Not the same event as error_occurred with severity: "warning"

Custom Workout has its own loading events (custom_workout_ready, commit_screen_shown, all_resources_loaded, workout_committed): see Workout Events.