add_workout_plan_exercise
ChatGPTInsert one exercise with its planned sets at a zero-based index. Use an active exercise definition ID from list_exercises. Existing exercises and sets are retained. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.
add_workout_plan_set
ChatGPTInsert one planned set at a zero-based index within one existing planned exercise. Existing sets are retained. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.
create_custom_exercise
ChatGPTCreate one account-owned exercise definition for reuse in workout plans. This does not add the exercise to a plan. Use a new requestId for each intended creation.
create_saved_meal
ChatGPTSave one reusable meal preset without logging food or changing older records. Provide complete nutrients for one serving. Retry with the same requestId and arguments only.
create_workout_plan
ChatGPTCreate one workout plan with its complete ordered exercise and set graph in one write. Copy active exerciseDefinitionId values from list_exercises. This does not schedule or complete a workout. Retry with the same requestId and arguments only.
delete_custom_exercise
ChatGPTArchive an account-owned custom definition after reading get_exercise for its current revision. Library definitions cannot be archived; existing workout snapshots and audit history remain.
delete_workout_plan
ChatGPTImmediately archive a workout plan while retaining its sync and audit history.
delete_workout_plan_exercise
ChatGPTRemove one exercise from the current workout plan.
delete_workout_plan_set
ChatGPTRemove one planned set from an existing exercise. The final set cannot be removed; use delete_workout_plan_exercise to remove that exercise instead. Completed workout history stays unchanged. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.
get_account_access
ChatGPTRead current Helios account access. This does not start a trial or purchase.
get_change_set
ChatGPTRead one durable change-set receipt created by this authenticated MCP client.
get_exercise
ChatGPTRead one library or account-owned exercise definition.
get_nutrition_targets
ChatGPTRead daily targets for today in the user's time zone. Targets may be empty until the user sets them. Keep localDate and revision for set_nutrition_targets. Targets remain active until the next change.
get_saved_meal
ChatGPTRead one active meal preset.
get_saved_meals
ChatGPTRead up to twenty-five known saved-meal presets in one call. Results keep the requested order; a missing, archived, or inaccessible preset has meal=null and does not block other reads. Use list_saved_meals to discover IDs. Previous meals are separate food-record views.
get_training_note
ChatGPTRead one active linked training note by its stable identifier.
get_training_summary
ChatGPTReturn a compact account-scoped summary using only the training read scopes granted to this connection.
get_workout_history
ChatGPTRead one completed or voided session with immutable performed data and corrections.
get_workout_plan
ChatGPTRead one workout plan in its current state.
get_workout_plans
ChatGPTRead up to ten known plan IDs in one call. Results keep the requested order; a missing or inaccessible plan has plan=null and does not block other reads. Use list_workout_plans to discover IDs. Request fewer IDs if the full graphs exceed the result-size limit.
list_exercises
ChatGPTList active library and account-owned custom exercise definitions. Use nextCursor with the same source and query to continue.
list_previous_meals
ChatGPTList distinct meals from active food records, newest first. Use nextCursor to continue; offset remains supported for older clients.
list_recent_change_sets
ChatGPTList recent durable change receipts created by this authenticated MCP connection. Use nextCursor to continue with the same connection.
list_saved_meals
ChatGPTList active saved meal presets for this account. Use nextCursor to fetch the next page; offset remains supported for older clients. Previous food records are separate.
list_training_notes
ChatGPTList active linked training notes for the authenticated account, newest first. Use nextCursor from the prior page to continue.
list_workout_history
ChatGPTList completed or explicitly voided workout sessions, newest first. Use nextCursor from the prior page to continue.
list_workout_plans
ChatGPTList workout plans, newest first. Use nextCursor to read the next page with the same includeArchived filter. Stop when nextCursor is null.
move_workout_plan_exercise
ChatGPTMove one existing planned exercise to a zero-based index in the same plan. No exercises or sets are added or removed. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.
move_workout_plan_set
ChatGPTMove one existing planned set to a zero-based index within its current exercise. No sets are added, removed, or moved to another exercise. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.
remove_saved_meal
ChatGPTArchive a preset and keep earlier food records.
save_previous_meal
ChatGPTCopy a food record into a reusable meal preset.
set_nutrition_targets
ChatGPTSet calories, protein, carbs and fat from today until the next change. Use integer milli-kcal and milligrams, with explicit zero allowed for macros. Read get_nutrition_targets first. expectedLocalDate is a stale-read check, not a start-date choice. No end date or mobile approval is needed.
set_workout_plan_details
ChatGPTChange one plan's name or notes. Omitted fields stay unchanged; null clears notes. Exercises and sets stay unchanged. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.
undo_change_set
ChatGPTImmediately append a compensating change when the target reversible change is still current.
update_custom_exercise
ChatGPTReplace an account-owned custom exercise definition. Read get_exercise for exerciseId and expectedRevision first. Library definitions cannot be changed; existing workout snapshots keep their recorded names.
update_saved_meal
ChatGPTReplace one preset's complete definition. Read get_saved_meal for mealId and expectedRevision first. Existing food records keep their snapshots. Require calories, protein, carbs, and fat for one standard serving. Zero must be measured, never a missing value. Select sources automatically: available package data for the exact product, its matching manufacturer page, then AFCD or USDA FoodData Central for matching generic food. Use Open Food Facts as a lead; compare with the label when available. If incomplete, check another eligible source without asking the user to validate a provider. Keep the same product, serving, units and preparation. Store the source name and URL or package reference in sourceReference. Convert kcal and grams to integer milli-kcal and milligrams once. Keep incomplete data as a draft; never invent values.
update_workout_plan
ChatGPTReplace one plan's full name, notes, exercises, and sets. Read get_workout_plan first and pass its current expectedRevision. Existing exercises or sets not in the new graph are removed; use the separate plan detail, exercise, or set tools for a small change.
update_workout_plan_exercise
ChatGPTChange one planned exercise's definition, default rest, or notes. Omitted fields and its planned sets stay unchanged. Use its nested planExerciseId, not the library definition ID. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.
update_workout_plan_set
ChatGPTChange one planned set's kind, target reps, load in grams, RPE, or rest. Omitted fields stay unchanged; null clears a nullable target. Completed workout history stays unchanged. Read get_workout_plan first for its current revision and nested exercise or set IDs. This change returns a receipt that can be undone while the plan remains current.