account.get_context
ChatGPTUse this when the user asks who they are in Eixo Run, whether they are a runner or coach, which teams or organizations they can access, or before running coach-only or athlete-specific Eixo Run queries. This reads account context only and never changes training data.
coach.get_athlete_context_brief
ChatGPTUse this when a coach asks for context on one accessible athlete before writing feedback or making a training decision: recent workouts, plan status, planned-workout matches, feedback threads, care inbox signals, and best efforts. This is read-only.
coach.get_attention_brief
ChatGPTUse this when a coach asks who needs attention, who disappeared, who is below plan, who reported pain, onboarding gaps, or recent personal-best moments.
coach.get_roster_activity_brief
ChatGPTUse this when a coach asks what their athletes have done recently, wants a roster-level view of recent workouts, asks which athletes received feedback, or wants to review feedback they gave across athletes. This reads only athletes accessible through the selected organization and only returns feedback authored by the authenticated coach.
coach.list_athletes
ChatGPTUse this when a coach needs to find athletes they can access before asking about a specific runner, roster segment, or org athlete.
eixo.open_app
ChatGPTOpen Eixo Run from the sidebar or beside this conversation. Shows signed-in runner training analysis with Pacer, or the coach's accessible organizations and athlete tools. Accepts empty arguments. Reads data without changing training, feedback or account settings.
media.prepare_training_visual_brief
ChatGPTUse this when the user wants an image, poster, share card, race-build graphic, or photo-based visual generated from Eixo Run training data.
runner.analyze_training
ChatGPTUse when a runner asks Pacer for an on-demand training evaluation, daily or frequent feedback, volume/consistency/pace/feeling charts, or a comparison between recent and previous periods. Also reads one coach-authorized athlete when athleteUserId is provided. Returns deterministic, chart-ready evidence from stored workouts, equally sized rolling periods, metric coverage, provenance and limitations. For prescribed workout context or saved HR/pace zones, combine the existing plan or zone-calibration tools. Never infers fitness improvement from aggregate pace and never changes plans or writes feedback.
runner.get_coach_feedback_threads
ChatGPTUse this when the signed-in runner asks what their coach said, wants feedback history, asks about feedback on specific workouts or planned workouts, or wants open threads to review. This is read-only and does not reply, dismiss, or change feedback.
runner.get_feedback_brief
ChatGPTUse this first when the signed-in runner asks about their own recent post-workout feedback, feedback patterns, feeling-score distribution, comments, tiredness, pain, heat, strong efforts, or workouts worth discussing with a coach. This is the dedicated feedback analysis tool and reads only the signed-in runner's own workouts.
runner.get_home_brief
ChatGPTUse this when the signed-in runner asks for a complete Eixo Run overview of their current running context: recent workouts, active plans, upcoming sessions, next race, coach feedback, personal bests, and shoes. This reads only the signed-in runner's own account.
runner.get_personal_bests
ChatGPTUse this when the signed-in runner asks for PRs, personal bests, best 5K/10K/21K/42K, or record history. This uses Eixo Run's materialized personal-best summary and does not recompute or mutate records.
runner.get_plan_brief
ChatGPTUse this when the signed-in runner asks about their current training plan, upcoming planned workouts, cancelled or incomplete sessions, AI/self/coach plan source, or how executed workouts matched the plan. This is read-only.
runner.get_race_brief
ChatGPTUse this when the signed-in runner asks about upcoming races, registered/going/completed races, race recommendations, race-course context, or existing race strategy summaries. This is read-only and does not mark interest or registration.
runner.get_shoe_brief
ChatGPTUse this when the signed-in runner asks about their running shoes, shoe mileage, active shoes, workouts without a shoe assigned, or best workouts by shoe. This reads only the runner's stored shoe and workout data.
runner.get_zone_calibration_brief
ChatGPTUse this when the signed-in runner asks about Z2, easy runs that felt or looked too hard by heart rate, pace-vs-heart-rate mismatch, threshold, lactate-threshold heart rate, or whether recent runs support their current Eixo Run zones. This reads the runner's saved threshold profile plus summarized recent run streams; it does not calculate, edit, or save new zones.
training.get_snapshot
ChatGPTUse this when the user asks for a recent training summary, consistency check, current plan context, personal-best overview, next race context, or a coach asks for one athlete's snapshot. Do not use this for post-workout feedback, feeling-score, comment, tiredness, or pain pattern analysis; use runner.get_feedback_brief for those requests.
training.get_workout
ChatGPTRead one specific completed workout by its canonical Eixo Run id discovered in training.search_workouts or runner.analyze_training. Returns recorded facts, runner reflection and the existing canonical planned-versus-executed comparison when available. For a coach athlete, also supply orgId, source and externalId from the authorized search result; the backend enforces access and the returned workout id must match. Use this for explain-this-workout follow-ups; never search by a timestamp to guess identity. Does not open a panel, change a plan, generate a persisted reading or send feedback.
training.search_workouts
ChatGPTUse this when the user asks for specific workout records or ranked/filterable workout lists, including questions like best 10K above 25 Celsius, runs in a city, hot-weather efforts, or shoe-specific sessions. If the request is about feedback patterns, feeling scores, tiredness, pain, comments, or workouts to discuss with a coach, use runner.get_feedback_brief instead.