approve_pending_client
ChatGPTLet a client who is waiting (pending_coach) in. While coaching is free nobody new waits, since there are no client limits; this is for an older request. If a plan limit is ever back on and the plan is full, the coach frees a seat by removing a client (remove_client). Only an owner or admin can approve an organization client.
cancel_client_invite
ChatGPTCancel an open Simple Macros coaching invite identified by its email address or invite_id from get_coach_invite_link. Invalidates its link in Simple Macros without sending an email. Only invites sent by this coach can be canceled. Confirm with the coach before canceling.
create_favorite
ChatGPTSave a reusable favorite. Give EITHER from_meal_id (copy a logged meal's items; that meal then shows as this favorite in the app) OR items[] (same shapes as log_meal). servings = how many servings the whole thing makes (default 1; a batch recipe might make 6). On a name clash you get error_code 'name_taken'; ask the user, then retry with replace_existing: true or a new name. Each item is one of: { kind: 'food', food_item_id, quantity, unit? } from our catalog; { kind: 'favorite', saved_meal_id, portions? } for a saved favorite; { kind: 'menu_item', restaurant, menu_item_id or name, options?, remove_options?, quantity? } for a restaurant item (get_restaurant_menu); { kind: 'combo', combo_id } for a whole idea from get_restaurant_meal_ideas; or { kind: 'custom', name, amount, unit, calories, protein_g, carbs_g?, fat_g?, ... } with macros for the WHOLE amount, only for recipes and foods search_foods doesn't have. A custom item that matches a catalog food is rejected with error_code 'resolvable' and candidates: resend with the food_item_id, or with allow_custom: true if it's genuinely different. Restaurant food can't be a custom item: use menu_item or combo.
delete_favorite
ChatGPTDelete a favorite. Meals already logged with it keep their food (they hold their own copy). Ask the user first, then call with confirm: true.
delete_meal
ChatGPTDelete a logged meal and all its items. Ask the user first, then call with confirm: true (without it nothing is deleted). Saved favorites are not affected.
delete_weight
ChatGPTDelete the weigh-in for a date. Ask the user first, then call with confirm: true (without it nothing is deleted).
duplicate_favorite
ChatGPTCopy a favorite into a new, independent one (named '<name> (copy)' unless you give a name). Useful for 'the same, but with one change'.
fetch
ChatGPTFetch one result from search by its id (meal:<id>, favorite:<id>, restaurant:<slug>).
find_nearby_restaurants
ChatGPTStores near a place for the chains Simple Macros has menus for in the user's country, nearest first. Pass near (a city, ZIP or postal code, or address the user gave); without it, the client's approximate location is used when it shares one. Optionally restaurant to find one chain. Distances are in the user's unit (distance_unit: miles in the US and UK, km elsewhere). Then use get_restaurant_meal_ideas for what to order.
get_client_days
ChatGPTRead one active client's records for a date range: logged meals with every food and its recorded nutrients, stored steps, weigh-ins and daily goals. Resolve client only among the coach's authorized relationships. Dates use the client's timezone; default today. Select only the needed categories with include. Returns dated records, not averages, trends, adherence scores or comparisons. Calculate those from the returned records, using the correct denominator. Never treat null steps as zero or an empty meal log as zero food intake. Today is marked partial. Workouts are device-only and unavailable here. Follow next_start until has_more is false before describing the whole range. Client sharing consent includes connected AI assistants. Tracked fitness data is returned only for active relationships that authorize the signed-in coach or organization admin. Clients can revoke sharing in Settings, and permissions are checked on every request. Results are tracking summaries, not medical diagnoses or treatment recommendations.
get_coach_invite_link
ChatGPTThe coach's code and permanent link to send a client (the client opens it, installs Simple Macros and approves sharing), plus email invites nobody has accepted yet. For each organization they belong to, also their org code and link: a client who uses it joins the org and is assigned to this coach. Use for 'what's my coach code?', 'send me my link' and 'did Tammy accept my invite?'. The QR code is on simplemacros.com/coach/invite. To change the personal code use set_coach_code.
get_coaching_plan
ChatGPTRead the signed-in coach's current client capacity: active and pending client counts, configured limits, remaining capacity, and whether client detail access is restricted by capacity. Also returns their coaching organizations and membership roles; organization capacity is available only to owners and admins. Null limits mean no configured limit. Does not change account access or client assignments.
get_days
ChatGPTRead the signed-in user's own records for a date range: logged meals with every food and its recorded nutrients, stored steps, weigh-ins and daily goals. Use days or start/end in their timezone; default today. Select only the needed categories with include. Returns dated records, not averages, trends, adherence scores or comparisons. Calculate those from the returned records, using the correct denominator. Never treat null steps as zero or an empty meal log as zero food intake. Today is marked partial. Workouts are device-only and unavailable here. Follow next_start until has_more is false before describing the whole range.
get_favorite
ChatGPTOne favorite with every item (item_id, name, amount, calories, protein, carbs, fat) for the WHOLE batch; divide by servings for one serving.
get_food
ChatGPTFull details of one catalog food by food_item_id: per-100g macros (calories, protein, carbs, fat, fiber, sugars, sodium, cholesterol), the default serving, and the 'as prepared' state when the label has one (e.g. dry pasta vs cooked). Needed to log by weight ('150 g') or as prepared.
get_goals
ChatGPTRead fixed daily goals, exact bounds, scheduled changes, revision, and pending coach proposals. Dates are inclusive local dates. Steps always use a minimum. Net carbs means total carbs minus fiber. Empty or incomplete tracking cannot establish that a maximum was met.
get_profile
ChatGPTThe user's profile and today's goals: calories, protein, steps, goal direction and pace, activity level, current/target weight, TDEE, timezone, and this connection's access level. setup_complete is false until the account finishes its short setup; then setup_url is a link for the user to finish it on the website.
get_restaurant_meal_ideas
ChatGPTSimple Macros' Meal Generator: real orders at a chain that hit a calorie target with at least a protein floor, built from the official menu. With no calories, the target is what the user has LEFT today (calories and protein), so 'what can I eat at X to fill out my day' needs only the restaurant. With calories and no protein_g, the protein floor is that many calories' worth of their daily goals (the app's ratio), so '750 calories that hits my macros' needs only calories. Pass protein_g ONLY when the user is explicit (a number, or 'my remaining protein'); never fill it in yourself. must_include / exclude to steer it. Each idea has a combo_id: log it with log_meal items [{ kind: 'combo', combo_id }]. The numbers are the chain's official published nutrition, and every item carries breakdown_per_item (base plus each option). The breakdown shows how a total is built from the chain's own portion sizes (a side cup is a different portion from an entrée portion).
get_streaks_and_badges
ChatGPTCurrent and longest streak for weigh-ins, calorie goal (within ±500), protein goal and steps; every badge earned with its date; and the next badge in each category with how far away it is when we can tell. Badges are awarded by the app, so a badge earned from here shows up after the app next opens.
invite_client_by_email
ChatGPTEmail someone a one-time invite to connect with the coach in Simple Macros (the same invite the portal sends; it expires in 14 days). Whoever accepts it and approves sharing becomes the client. With organization, the client joins that organization instead: a plain org coach's invite assigns the client to themselves; an owner or admin may pick a coach with assign_to (or leave the client unassigned). Confirm the address with the coach first.
list_favorites
ChatGPTThe user's saved favorites (reusable meals and recipes), most-used first, with servings, per-serving calories/protein, tags, restaurant and item names. query matches names, items and restaurant fuzzily; tag filters (breakfast, lunch, dinner, snack, sides, drinks, restaurant). Check here first when the user says 'my usual' or names a meal.
list_my_clients
ChatGPTList only the signed-in coach's connected clients: names, connection status, organization name when applicable and the relationship reference used to select a client. Includes active clients and pending requests. No fitness records or check-in assessments are returned. Use get_client_days for an active client's records. Access is checked on every request.
list_my_reviews
ChatGPTReviews clients left for the coach, newest first, with the coach's reply if any, plus the average rating.
log_meal
ChatGPTTrack one meal in Simple Macros; it shows on the Track tab like a meal logged in the app. Each item is one of: { kind: 'food', food_item_id, quantity, unit? } from our catalog; { kind: 'favorite', saved_meal_id, portions? } for a saved favorite; { kind: 'menu_item', restaurant, menu_item_id or name, options?, remove_options?, quantity? } for a restaurant item (get_restaurant_menu); { kind: 'combo', combo_id } for a whole idea from get_restaurant_meal_ideas; or { kind: 'custom', name, amount, unit, calories, protein_g, carbs_g?, fat_g?, ... } with macros for the WHOLE amount, only for recipes and foods search_foods doesn't have. A custom item that matches a catalog food is rejected with error_code 'resolvable' and candidates: resend with the food_item_id, or with allow_custom: true if it's genuinely different. Restaurant food can't be a custom item: use menu_item or combo. Optional date (YYYY-MM-DD, default today), time (HH:mm), and label. Returns the meal and what's left for the day.
log_weight
ChatGPTLog OR correct the day's weigh-in, in the chosen unit (default: the user's preference). A day has exactly one weigh-in: if one already exists for that date it is updated (the result says updated: true and gives previous_weight_lb), never duplicated. Defaults to today and now.
lookup_barcode
ChatGPTFind a packaged product in our catalog by its UPC/EAN barcode. Several candidates can share a barcode; they come ranked with verified ones first. No match means we don't have it yet (adding a product needs the app's label-photo flow), so tell the user and use a custom item if they want it logged. Log it with log_meal as { kind: 'food', food_item_id, quantity }. Check list_favorites and get_frequent_foods first for the user's own versions.
propose_client_goals
ChatGPTPropose fixed daily goals to an active client by relationship_id. Does not change goals until the client explicitly accepts. Optional start_date and end_date are inclusive in the client timezone. Uses the same minimum, maximum, target and range rules as personal goals.
remove_client
ChatGPTEnd coaching with a client, or decline one who is waiting for a seat. The coach stops seeing their data right away and the seat frees up. An active client gets a notice. This is the only way to free a seat; to come back, the client reconnects with the coach's code and approves sharing again. An organization client leaves the whole organization, and only an owner or admin can remove them (to move them to another coach, use assign_client). Confirm with the coach first.
report_issue
ChatGPTSend the Simple Macros team a problem report when something went wrong here: wrong or surprising numbers, a meal or weigh-in that didn't log right, a food or menu item we're missing, or an error message. Offer it when the user says something is broken or wrong, and ask before sending. details is what the team uses to find and fix the problem, so make it specific: what the user asked for, which tools you called with the inputs you sent and what came back (the numbers, names, dates and any error text), and what was wrong or what the user expected instead. We add the account, the assistant and our recent tool-call log ourselves. The team replies to the user by email.
resend_client_invite
ChatGPTResend an open Simple Macros coaching invite identified by its email address or invite_id from get_coach_invite_link. Emails the recipient a fresh link valid for 14 more days, invalidates the old link, and revives an expired invite. Only invites sent by this coach can be resent. Confirm with the coach before sending.
respond_goal_proposal
ChatGPTAccept or decline one pending coach proposal only after the user explicitly chooses. Read get_goals first and show the changes and dates. Acceptance updates only proposed metrics and requires the current revision.
respond_to_review
ChatGPTPost or replace the coach's one public reply to a client review stored in Simple Macros. Only updates reviews of this signed-in coach within Simple Macros; it does not contact external review platforms. The reply is displayed on the coach's Simple Macros profile. An empty response removes the reply. It replaces any earlier reply, so confirm the wording with the coach first.
search
ChatGPTSearch the user's logged meals (last 92 days), saved favorites, and the restaurants Simple Macros has menus for.
search_foods
ChatGPTSearch the Simple Macros food catalog (USDA whole foods, packaged/branded products, and community-scanned items) for 1 to 10 foods at once. Returns candidates per query with food_item_id, name, brand, default serving and per-serving calories/protein/carbs/fat. An empty list means we don't have it: then use a custom item. Log it with log_meal as { kind: 'food', food_item_id, quantity }. Check list_favorites and get_frequent_foods first for the user's own versions.
search_restaurants
ChatGPTThe restaurant chains Simple Macros has full menus for (official nutrition, every option), limited to chains with locations in the user's country. Search by name, or omit query to list them all. Use the slug with get_restaurant_menu, get_restaurant_meal_ideas and log_meal. For stores near a place, use find_nearby_restaurants.
set_coach_code
ChatGPTChange the coach's code (and so their link). Either a code they choose, 3 to 12 letters and numbers such as their name (SARAHLEE): spaces and dashes are dropped and it's saved in capitals, and the old code keeps working so links and QR codes already handed out still reach this coach. Or random: true for a new random code, for when a code was shared somewhere it shouldn't be: then the current code AND every old one stop working, so confirm with the coach first. Up to 5 code changes a day.
update_coach_profile
ChatGPTRead or change the signed-in coach's profile stored in Simple Macros: name, business, headline (80 characters max), specialties (up to 5 of: Weight loss, Muscle gain, Strength, Endurance, Sports performance, General health, Plant-based, Women's health, Men's health, Over 50, Bodybuilding, Habit coaching, Postpartum), how they coach (online, in person or both), city/region/country, languages, a price note, website, Instagram and TikTok usernames, and whether they're taking new clients. Only updates this coach's Simple Macros profile. Website and social usernames are stored as profile text; no websites or social accounts are contacted or edited. Only the fields passed change; call with no fields to read the profile. Show the coach the new text before saving. The photo is changed on simplemacros.com/coach/settings.
update_favorite
ChatGPTEdit a favorite's template: rename, change servings or tags, add_items, remove_item_ids (item_id from get_favorite), or set_items to replace all. Future logs use the new version; meals already logged keep their copy.
update_goals
ChatGPTChange the user's daily goals from today on: calorie_diff (calories per day vs maintenance/TDEE, negative = deficit, e.g. -500), daily_protein_goal (grams), step_goal, activity_level. Returns the new goals and the TDEE math behind them. Past days keep their old goals.
update_meal
ChatGPTEdit a logged meal by meal_id (from get_day / get_meals): add_items, remove_food_ids (the item food_id values from get_day), set_quantity to change one item's amount, set_items to replace everything, or move it with date / time / label. Items use the same shapes as log_meal. Returns the updated meal and what's left for the day.
update_profile
ChatGPTChange target weight, weekly pace, goal direction (lose | maintain | gain), or timezone (IANA name like America/Chicago). These are profile fields only: they do NOT change the daily calorie goal. To change calories, use update_goals (e.g. calorie_diff -500 for about 1 lb/week).
create_favorite
Claudedelete_favorite
Claudedelete_meal
Claudedelete_weight
Claudeduplicate_favorite
Claudefetch
Claudefind_nearby_restaurants
Claudeget_day
Claudeget_favorite
Claudeget_food
Claudeget_frequent_foods
Claudeget_history
Claudeget_meals
Claudeget_nutrient_report
Claudeget_profile
Claudeget_restaurant_meal_ideas
Claudeget_streaks_and_badges
Claudeget_weight_history
Claudelist_favorites
Claudelog_meal
Claudelog_weight
Claudelookup_barcode
Claudereport_issue
Claudesearch
Claudesearch_foods
Claudesearch_restaurants
Claudeupdate_favorite
Claudeupdate_goals
Claudeupdate_meal
Claudeupdate_profile
Claude
