cancel_action
ChatGPTCancels an unconfirmed pending mutation. Nothing is written.
Healthcare
Overview
Connect an existing Zolt account to review nutrition, adaptive TDEE, weight, hydration, workouts, and consent-gated wearable summaries; get food suggestions based on remaining targets; and prepare account changes that require a separate explicit confirmation before execution. Zolt handles sensitive health and wellness information and is intended for users age 13 and older. It is a consumer wellness service, not medical advice, diagnosis, treatment, or emergency support.
Tools
confirm_action
ChatGPTExecutes one prepared mutation only after the user explicitly approved its complete preview in a later turn. Tokens are user-bound, single-use, and expire after ten minutes.
draft_food_log
ChatGPTUse this whenever the user says they ate, had, drank, consumed, or wants to add or log any food or beverage. Trigger even if they do not say 'log', and use it for both one food and multiple foods. Include every named food as a separate query. Resolves the complete intake statement into an editable, uncommitted Zolt food-log draft. The tool does not write data. Use its verified matches and serving nutrition to build an optimistic itemized confirmation draft immediately. Choose a reasonable default for unresolved variants or units, label each assumption beside its item, and invite the user to refine it. Ask a blocking question only when no defensible draft can be formed. Never answer with general-knowledge calorie estimates instead. Use 'cal' or 'calories' for dietary energy in user-facing responses to match Zolt; do not rewrite it as 'kcal'.
get_adherence
ChatGPTUse this when the user asks how often they met calorie or protein targets across up to 366 days. Uses explicit targets when supplied, otherwise the current goal's latest confirmed targets.
get_bodyweight
ChatGPTGet the user's bodyweight log history. Returns entries sorted by date, with source information (mcp, healthkit, etc.).
get_checkins
ChatGPTGet the user's weight goal check-in history. Returns TDEE at check-in, calorie targets, macro adjustments, trend weight, and scale weight over time. Useful for understanding progress on a cut or bulk.
get_custom_foods
ChatGPTList the user's custom foods with nutrition, serving data, and commit_food_log-compatible identity fields.
get_daily_nutrition_status
ChatGPTUse this when the user asks what they have consumed or have left today. Returns calories/macros consumed, current goal targets, remaining amounts, and percentages for one date.
get_daily_summary
ChatGPTReturns one combined daily view of nutrition, targets/TDEE, hydration, weight progress, synced health aggregates, wearable workouts, and Zolt workouts. In user-facing responses, label dietary energy as 'cal' or 'calories' to match Zolt; do not rewrite it as 'kcal'.
get_exercise_history
ChatGPTReturns all-time paginated sessions and progression metrics for one performed exercise.
get_food_frequency
ChatGPTUse this when the user asks what foods they eat most, how often they ate something, or when they last logged it across up to 366 days.
get_food_history
ChatGPTUse this when the user says 'my usual', 'same as last time', or asks when/how often they logged a food. Returns real source IDs and typical servings for reuse in an editable food draft.
get_food_logs
ChatGPTUse this whenever the user asks what they ate or logged, wants to review or correct an existing food log, or wants to copy, move, replace, or delete logged food. Gets food log entries for one date or an inclusive date range of at most 14 days. Returns source food IDs/types, serving selections, meal sections, and macros so an assistant can review, repeat, or delete entries accurately.
get_food_recommendations
ChatGPTUse this when the user asks what to eat or what fits their remaining macros. Ranks foods from their own recent history, custom foods, and recipes; results are suggestions only and are never logged automatically.
get_health_summary
ChatGPTRead consent-gated daily Apple Health and Google Health aggregates for sleep, recovery vitals, steps, and wearable expenditure. No raw samples are returned and the values are not medical advice.
get_health_workouts
ChatGPTGet aggregate workout summaries synced from Apple Health or Health Connect. Returns no GPS routes, raw heart-rate samples, notes, events, or device identifiers.
get_hydration
ChatGPTReturns hydration entries and daily totals for a date range. Entries imported from health platforms are read-only and must be corrected at their source.
get_nutrition_stats
ChatGPTUse this for totals, averages, trends, macro split, micronutrients, coverage, or high/low days across up to 366 days. Missing days are reported as missing, never treated as zero-consumption days.
get_nutrition_summary
ChatGPTGet aggregated calorie and macro totals for a date range (max 14 days). Returns per-day breakdowns and overall averages. Useful for assessing adherence to a diet plan.
get_recent_meals
ChatGPTUse this when the user asks what they ate on a prior day, wants the same meal again, or asks what they usually eat for a meal section. Returns complete grouped meals with source and serving fields.
get_recipes
ChatGPTList the user's saved recipes with ingredients, total nutrition, serving data, and commit_food_log-compatible identity fields.
get_tdee
ChatGPTReturns the user's current adaptive TDEE, calorie target, macro targets, active goal, and source check-in date. In user-facing responses, label dietary energy as 'cal' or 'calories' to match Zolt; do not rewrite it as 'kcal'.
get_training_summary
ChatGPTReturns deterministic workout frequency, duration, working-set, volume, distance, and period-comparison metrics.
get_trends
ChatGPTCombines nutrition, weight, activity, sleep, recovery, hydration, and workout trends over one to four weeks.
get_user_profile
ChatGPTGet the user's profile including name, gender, height, weight unit, activity level, and starting bodyweight. Useful for nutrition coaching context.
get_weight_goal
ChatGPTGet the user's weight goals. Returns goal type (lose/gain/maintain), target weight, starting weight, macro split percentages, and timeline.
get_weight_progress
ChatGPTUse this when the user asks about trend weight, pace, distance from their goal, or whether their cut/bulk is on track. Uses a canonical seven-calendar-day average and the active goal.
get_workout_detail
ChatGPTReturns one completed Zolt Train workout with every exercise and set.
get_workout_history
ChatGPTReturns completed Zolt Train strength/cardio workouts and sets. Supports date, exercise, type, and cursor filters. Separate from wearable workouts.
get_workout_prs
ChatGPTCalculates rep-count personal records and estimated one-rep maxes directly from canonical completed workout sets.
get_workout_templates
ChatGPTRead saved Zolt Train workout templates, including complete exercises, sets, reps, rest, weights, supersets, progression, folder and sync timestamp. Use template_id for a single template; lists paginate by offset. Deleted templates are excluded. Weight values are pounds.
prepare_bodyweight
ChatGPTStages one bodyweight value for review. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_custom_food
ChatGPTStages a custom-food create, replacement, or deletion. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_food_log
ChatGPTValidates and resolves the final itemized food-log draft. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_food_log_changes
ChatGPTStages edits, moves, copies, replacements, or deletions of existing food logs. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_recipe
ChatGPTStages a recipe create, replacement, or deletion. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_water
ChatGPTStages one hydration amount for review. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_water_change
ChatGPTStages a hydration correction or deletion for review. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_weekly_checkin_commit
ChatGPTStages the final weekly target selected from preview_weekly_checkin. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_weight_goal
ChatGPTStages a weight-goal create, update, completion, or deletion. It stores an expiring server-side draft but does not execute the requested health action; it returns a confirmation token.
prepare_workout
ChatGPTValidates a completed Zolt Train workout and stores an expiring server-side confirmation preview. It does not log the workout.
prepare_workout_changes
ChatGPTStores an expiring server-side draft for a full replacement or deletion of an existing completed Zolt Train workout. It does not change the workout and returns a confirmation token.
prepare_workout_template
ChatGPTStage creation of a reusable Zolt Train template, not a completed workout. Extract screenshot/text workouts with your own vision, resolve canonical exercises using search_exercises, and submit the complete draft. Returns a preview for explicit user confirmation; no template is saved yet. New templates are unfiled; Train assigns their order on sync.
prepare_workout_template_update
ChatGPTStage a full content replacement of an existing Train template. Read it first and preserve all unchanged exercise details in draft. Requires its updated_at timestamp, rejects stale previews, and preserves folder, order, usage and creation metadata. confirm_action saves only after explicit confirmation.
preview_weekly_checkin
ChatGPTUse this when the user asks to check in on a weekly-cadence weight goal. Returns recent weight pace and a server-calculated calorie/macro proposal for the client to display and edit; it does not write anything.
resolve_food_log_draft
ChatGPTResolve up to five distinct foods for a concrete food-log draft, including your existing custom foods and recipes. Returns one selected match and at most two alternatives per item, with serving data for prepare_food_log. No food is logged until the prepared draft is explicitly confirmed. Use this to refine food choices as well as create a new draft. Free accounts receive 30 distinct food resolutions per UTC day; repeating the same query and brand that day does not charge again.
search_exercises
ChatGPTSearches exercises the user has performed or created. Use before preparing a workout to preserve stable exercise identity.
search_foods
ChatGPTUse this to resolve or refine one food inside an active food-log draft, or when the user explicitly wants to search Zolt's food database. For a new statement about food the user ate or drank, use draft_food_log instead, whether it contains one food or many. Searches Zolt's verified food database plus the user's custom foods and saved recipes. Do not substitute general-knowledge estimates or broad calorie ranges. When a product variant, size, sushi type, quantity, or unit is ambiguous, search the known terms first, select the best defensible match and standard serving, and show an optimistic pending draft with the assumption attached to that item. Use this while building or editing the client's food-log draft, including after visually identifying food in an attached photo. Every result includes food_id, food_type, and serving options for the draft and eventual commit_food_log call. Use the selected serving to show the user each pending item's quantity, explicit unit, calories, protein, carbs, and fat before confirmation. Prefer an editable assumption over a blocking question; ask first only when no reasonable match or serving can be selected. Use 'cal' or 'calories' for dietary energy in user-facing responses to match Zolt; do not rewrite it as 'kcal'. Use short food/brand queries; never invent IDs or nutrition values.
Capabilities
InteractiveWrites
Example Prompts
Click any prompt to copy it.
Links
App Stats
48
Tools
3
Prompts
ChatGPT
Platforms
Category
Nutrition PlanningWorks with
ChatGPT
Data refreshed daily