NutraPlanner

NutraPlanner nutrition calculator API

Free, read-only nutrition calculators over HTTP: the same formulas as our web calculators, returned as JSON. No sign-up and no API key.

The NutraPlanner calculator API is free and read-only: no API key and no sign-up. It runs 9 calculators (BMI, BMR, TDEE, calorie target, macros, protein, body fat, water and ideal weight). Send GET /api/public/v1/calculators/{name} and get JSON back, up to 120 requests per minute per IP address. The OpenAPI 3.1 description is at /openapi.json.

Overview

The public API runs the formulas behind the free calculators (BMI, BMR, TDEE, calorie and macro targets, protein, body fat, water and ideal body weight) and returns each result as JSON.

  • No account, no API key, no sign-up.
  • Stateless: each request is computed and answered. The application does not store request inputs.
  • Inputs are metric: kilograms and centimetres.
  • Results are general estimates for healthy adults. They are not medical advice.
  • There is no API for practitioner or client account data. Client records, notes, meal plans and appointments are only available in the signed-in app.

Quick start

Send a GET request with query parameters. For example, total daily energy expenditure:

Request
curl -G 'https://nutraplanner.com/api/public/v1/calculators/tdee' \
  -d 'sex=female' \
  -d 'age=35' \
  -d 'weight_kg=68' \
  -d 'height_cm=165' \
  -d 'activity=moderate'

Response
{
  "bmr_kcal": 1375,
  "tdee_kcal": 2132,
  "activity_multiplier": 1.55,
  "formula": "mifflin_st_jeor"
}

Authentication

None. Every endpoint is anonymous and CORS is open, so you can call it from a server, a script or a browser. Send no credentials.

Rate limits

120 requests per minute per client IP address, shared across all endpoints. Over the limit, the API answers 429 with a Retry-After header giving the seconds to wait.

Results are deterministic and sent with a Cache-Control header: cache them rather than repeating the same call.

Discovery

Machine-readable descriptions for agents and code generators:

Endpoint reference

Every endpoint is a GET under https://nutraplanner.com/api/public/v1. This reference is generated from the same definition the API runs on.

Endpoint, parameter and field descriptions come straight from the API definition, in English.

Body Mass Index

GET /api/public/v1/calculators/bmi

operationId
calculateBmi
Web calculator
BMI calculator

Body Mass Index (kg/m²) with the WHO adult category.

Query parameters: Body Mass Index
ParameterTypeDescription
weight_kgrequirednumber, ≥ 20 and ≤ 500Body weight in kilograms.
height_cmrequirednumber, ≥ 100 and ≤ 250Height in centimetres.
Response fields: Body Mass Index
FieldTypeDescription
bminumberBody Mass Index, kg/m², two decimals. The category is taken from this rounded value, so the two never disagree at a cut-off.
categorystring: underweight, normal, overweight or obeseWHO adult category.
Example request: Body Mass Index
curl -G 'https://nutraplanner.com/api/public/v1/calculators/bmi' \
  -d 'weight_kg=68' \
  -d 'height_cm=165'

Example response: Body Mass Index
{
  "bmi": 24.98,
  "category": "normal"
}

Basal metabolic rate

GET /api/public/v1/calculators/bmr

operationId
calculateBmr
Web calculator
TDEE calculator
Method
How Mifflin-St Jeor, Harris-Benedict and Katch-McArdle compare

Basal metabolic rate (kcal/day) by Mifflin-St Jeor (default), revised Harris-Benedict, or Katch-McArdle.

Query parameters: Basal metabolic rate
ParameterTypeDescription
sexrequiredone of female or maleSex used by the equation.
agerequirednumber, ≥ 18 and ≤ 120Age in years (adults).
weight_kgrequirednumber, ≥ 20 and ≤ 500Body weight in kilograms.
height_cmrequirednumber, ≥ 100 and ≤ 250Height in centimetres.
formulaoptionalone of mifflin_st_jeor, harris_benedict or katch_mcardle (default: mifflin_st_jeor)BMR equation. katch_mcardle requires body_fat_pct.
body_fat_pctoptionalnumber, ≥ 2 and ≤ 70Body-fat percentage. Required only for formula=katch_mcardle.Estimate it with the body-fat endpoint
Response fields: Basal metabolic rate
FieldTypeDescription
bmr_kcalintegerBasal metabolic rate, kcal/day.
formulastring: mifflin_st_jeor, harris_benedict or katch_mcardleEquation used.
Example request: Basal metabolic rate
curl -G 'https://nutraplanner.com/api/public/v1/calculators/bmr' \
  -d 'sex=female' \
  -d 'age=35' \
  -d 'weight_kg=68' \
  -d 'height_cm=165'

Example response: Basal metabolic rate
{
  "bmr_kcal": 1375,
  "formula": "mifflin_st_jeor"
}

Total daily energy expenditure

GET /api/public/v1/calculators/tdee

operationId
calculateTdee
Web calculator
TDEE calculator
Method
How Mifflin-St Jeor, Harris-Benedict and Katch-McArdle compare

Total daily energy expenditure (kcal/day): BMR multiplied by an activity factor.

Query parameters: Total daily energy expenditure
ParameterTypeDescription
sexrequiredone of female or maleSex used by the equation.
agerequirednumber, ≥ 18 and ≤ 120Age in years (adults).
weight_kgrequirednumber, ≥ 20 and ≤ 500Body weight in kilograms.
height_cmrequirednumber, ≥ 100 and ≤ 250Height in centimetres.
activityrequiredone of sedentary, light, moderate, active or very_activeActivity level; multiplier applied to BMR (sedentary ×1.2, light ×1.375, moderate ×1.55, active ×1.725, very_active ×1.9).
formulaoptionalone of mifflin_st_jeor, harris_benedict or katch_mcardle (default: mifflin_st_jeor)BMR equation. katch_mcardle requires body_fat_pct.
body_fat_pctoptionalnumber, ≥ 2 and ≤ 70Body-fat percentage. Required only for formula=katch_mcardle.Estimate it with the body-fat endpoint
Response fields: Total daily energy expenditure
FieldTypeDescription
bmr_kcalintegerBasal metabolic rate, kcal/day.
tdee_kcalintegerTotal daily energy expenditure, kcal/day.
activity_multipliernumberMultiplier applied to BMR.
formulastring: mifflin_st_jeor, harris_benedict or katch_mcardleBMR equation used.
Example request: Total daily energy expenditure
curl -G 'https://nutraplanner.com/api/public/v1/calculators/tdee' \
  -d 'sex=female' \
  -d 'age=35' \
  -d 'weight_kg=68' \
  -d 'height_cm=165' \
  -d 'activity=moderate'

Example response: Total daily energy expenditure
{
  "bmr_kcal": 1375,
  "tdee_kcal": 2132,
  "activity_multiplier": 1.55,
  "formula": "mifflin_st_jeor"
}

Daily calorie target

GET /api/public/v1/calculators/calorie-target

operationId
calculateCalorieTarget
Web calculator
Calorie calculator

Daily calorie target (kcal/day): TDEE adjusted for a weight goal (about ±250 or ±500 kcal/day).

Query parameters: Daily calorie target
ParameterTypeDescription
sexrequiredone of female or maleSex used by the equation.
agerequirednumber, ≥ 18 and ≤ 120Age in years (adults).
weight_kgrequirednumber, ≥ 20 and ≤ 500Body weight in kilograms.
height_cmrequirednumber, ≥ 100 and ≤ 250Height in centimetres.
activityrequiredone of sedentary, light, moderate, active or very_activeActivity level; multiplier applied to BMR (sedentary ×1.2, light ×1.375, moderate ×1.55, active ×1.725, very_active ×1.9).
goalrequiredone of lose, mild_lose, maintain, mild_gain or gainWeight goal (lose -500, mild_lose -250, maintain +0, mild_gain +250, gain +500 kcal/day).
formulaoptionalone of mifflin_st_jeor or harris_benedict (default: mifflin_st_jeor)BMR equation.
Response fields: Daily calorie target
FieldTypeDescription
tdee_kcalintegerTotal daily energy expenditure, kcal/day.
adjustment_kcalintegerGoal adjustment applied to TDEE, kcal/day.
target_kcalintegerDaily calorie target, kcal/day.
Example request: Daily calorie target
curl -G 'https://nutraplanner.com/api/public/v1/calculators/calorie-target' \
  -d 'sex=female' \
  -d 'age=35' \
  -d 'weight_kg=68' \
  -d 'height_cm=165' \
  -d 'activity=moderate' \
  -d 'goal=mild_lose'

Example response: Daily calorie target
{
  "tdee_kcal": 2132,
  "adjustment_kcal": -250,
  "target_kcal": 1882
}

Macronutrient grams from a calorie target

GET /api/public/v1/calculators/macros

operationId
calculateMacros
Web calculator
Macro calculator

Grams of protein, carbohydrate and fat for a calorie target and a split preset (4/4/9 kcal per gram). Every preset keeps protein within the 10–35% AMDR.

Query parameters: Macronutrient grams from a calorie target
ParameterTypeDescription
caloriesrequirednumber, ≥ 500 and ≤ 10000Daily calorie target, kcal.
splitoptionalone of balanced, low_carb, high_protein or high_carb (default: balanced)Split preset as protein/carbs/fat shares (balanced 25/45/30, low_carb 25/35/40, high_protein 30/45/25, high_carb 20/60/20).
Response fields: Macronutrient grams from a calorie target
FieldTypeDescription
protein_gintegerProtein, grams/day.
carbs_gintegerCarbohydrate, grams/day.
fat_gintegerFat, grams/day.
splitstring: balanced, low_carb, high_protein or high_carbSplit preset used.
Example request: Macronutrient grams from a calorie target
curl -G 'https://nutraplanner.com/api/public/v1/calculators/macros' \
  -d 'calories=2000'

Example response: Macronutrient grams from a calorie target
{
  "protein_g": 125,
  "carbs_g": 225,
  "fat_g": 67,
  "split": "balanced"
}

Daily protein target

GET /api/public/v1/calculators/protein

operationId
calculateProteinTarget
Web calculator
Protein calculator
Method
The evidence behind the g/kg goal bands

Daily protein target (grams) from body weight and an activity goal, in g/kg.

Query parameters: Daily protein target
ParameterTypeDescription
weight_kgrequirednumber, ≥ 20 and ≤ 500Body weight in kilograms.
goalrequiredone of sedentary, active, build_muscle or athleteGoal (sedentary 1 g/kg, active 1.2 g/kg, build_muscle 1.6 g/kg, athlete 2 g/kg).
Response fields: Daily protein target
FieldTypeDescription
protein_gintegerProtein, grams/day.
g_per_kgnumberGrams of protein per kilogram of body weight.
Example request: Daily protein target
curl -G 'https://nutraplanner.com/api/public/v1/calculators/protein' \
  -d 'weight_kg=68' \
  -d 'goal=active'

Example response: Daily protein target
{
  "protein_g": 82,
  "g_per_kg": 1.2
}

Body-fat percentage (U.S. Navy method)

GET /api/public/v1/calculators/body-fat

operationId
calculateBodyFatNavy
Web calculator
Body-fat calculator

Body-fat percentage from circumference measurements by the U.S. Navy method. hip_cm is required for sex=female.

Query parameters: Body-fat percentage (U.S. Navy method)
ParameterTypeDescription
sexrequiredone of female or maleSex used by the equation.
height_cmrequirednumber, ≥ 100 and ≤ 250Height in centimetres.
neck_cmrequirednumber, ≥ 20 and ≤ 100Neck circumference, cm.
waist_cmrequirednumber, ≥ 40 and ≤ 250Waist circumference, cm.
hip_cmoptionalnumber, ≥ 40 and ≤ 250Hip circumference, cm. Required for sex=female.
Response fields: Body-fat percentage (U.S. Navy method)
FieldTypeDescription
body_fat_pctnumberEstimated body fat, percent, one decimal (always 2–70).
Example request: Body-fat percentage (U.S. Navy method)
curl -G 'https://nutraplanner.com/api/public/v1/calculators/body-fat' \
  -d 'sex=female' \
  -d 'height_cm=165' \
  -d 'neck_cm=33' \
  -d 'waist_cm=76' \
  -d 'hip_cm=98'

Example response: Body-fat percentage (U.S. Navy method)
{
  "body_fat_pct": 28.9
}

Daily water intake

GET /api/public/v1/calculators/water

operationId
calculateWaterIntake
Web calculator
Water intake calculator

Estimated daily fluid intake (mL) from body weight and activity, in mL per kg.

Query parameters: Daily water intake
ParameterTypeDescription
weight_kgrequirednumber, ≥ 20 and ≤ 500Body weight in kilograms.
activityrequiredone of sedentary, active or athleteActivity (sedentary 30 mL/kg, active 35 mL/kg, athlete 40 mL/kg).
Response fields: Daily water intake
FieldTypeDescription
water_mlintegerDaily fluid, millilitres.
ml_per_kgnumberMillilitres per kilogram of body weight.
Example request: Daily water intake
curl -G 'https://nutraplanner.com/api/public/v1/calculators/water' \
  -d 'weight_kg=68' \
  -d 'activity=active'

Example response: Daily water intake
{
  "water_ml": 2380,
  "ml_per_kg": 35
}

Ideal and adjusted body weight (Devine)

GET /api/public/v1/calculators/ideal-weight

operationId
calculateIdealBodyWeight
Web calculator
Ideal weight calculator

Ideal body weight (kg) by the Devine formula. With weight_kg, also the adjusted body weight used clinically when actual weight exceeds ideal.

Query parameters: Ideal and adjusted body weight (Devine)
ParameterTypeDescription
sexrequiredone of female or maleSex used by the equation.
height_cmrequirednumber, ≥ 100 and ≤ 250Height in centimetres.
weight_kgoptionalnumber, ≥ 20 and ≤ 500Actual body weight in kilograms, for adjusted body weight.
Response fields: Ideal and adjusted body weight (Devine)
FieldTypeDescription
ideal_weight_kgnumberIdeal body weight, kg, one decimal.
adjusted_weight_kgnumber | nullAdjusted body weight, kg; null when weight_kg is not given.
Example request: Ideal and adjusted body weight (Devine)
curl -G 'https://nutraplanner.com/api/public/v1/calculators/ideal-weight' \
  -d 'sex=female' \
  -d 'height_cm=165' \
  -d 'weight_kg=68'

Example response: Ideal and adjusted body weight (Devine)
{
  "ideal_weight_kg": 56.9,
  "adjusted_weight_kg": 59.7
}

Errors

Errors use RFC 9457 problem details (Content-Type: application/problem+json) with type, title, status and, usually, detail. The type URI points to the matching entry below.

Problem titles are shown exactly as the API returns them, in English.

Invalid query parameters

HTTP 400 · invalid-parameters

A parameter is missing, unknown, repeated or out of range. The errors array names each parameter and what is wrong with it.

Inputs cannot be computed

HTTP 422 · not-computable

Each parameter is valid on its own, but together they cannot be computed, for example formula=katch_mcardle without body_fat_pct.

Not found

HTTP 404 · not-found

Nothing exists at that path: no calculator by that name, or no Canadian Nutrient File food with that code (the CNF search finds a food's code). The endpoint reference lists every calculator.

Method not allowed

HTTP 405 · method-not-allowed

The API is read-only. Use GET (HEAD and OPTIONS are also answered).

Too many requests

HTTP 429 · rate-limited

The per-IP rate limit was exceeded. Wait the number of seconds in the Retry-After header, then try again.

Internal error

HTTP 500 · internal-error

The server failed while computing the result. The request itself may be fine: try again, and if it keeps failing, contact us.

Canadian Nutrient File API

Health Canada's Canadian Nutrient File 2026, plus the 2015 foods it withdrew, as three read-only endpoints. Same rules as the calculators: no key, GET only, the same per-client rate limit, problem+json errors. Values are per 100 g of edible portion, exactly as Health Canada published them.

  • GET /api/public/v1/cnf/foods?q=…&group=…&limit=… Search by English or French name or Health Canada's alternate names (q), or list a CNF food group (group); limit 1 to 50.
  • GET /api/public/v1/cnf/foods/{key} One food by its CNF food code: every nutrient with its value, the edition it comes from, its basis and the same food's 2015 value, plus Health Canada's household measures. A code that CNF 2026 gave to a different food has its 2015 food at {code}-2015.
  • GET /api/public/v1/cnf/nutrients Every nutrient of both editions: code, symbol, English and French name, unit, published decimals, and whether 2026 added or dropped it.
Search for oat beverages
curl 'https://nutraplanner.com/api/public/v1/cnf/foods?q=oat%20beverage&limit=5'

Which value, from which edition

Each nutrient says where its value comes from: cnf2026 (CNF 2026's own), carried_forward_from_cnf2015 (a nutrient CNF 2026 dropped, such as starch or oxalic acid, with its 2015 value), not_in_cnf2026 (CNF 2026 still publishes the nutrient but not for this food; value is null), or historical_cnf2015 (a food CNF 2026 withdrew). value_2015 is the same food's 2015 value, for comparison.

Licence and attribution

The data is Health Canada's, under the Open Government Licence – Canada. Every response carries the attribution in source, and a Link header names the licence. Keep the attribution when you show the data, and never suggest that Health Canada endorses your use. NutraPlanner's API is unofficial and not endorsed by Health Canada.

The same data as pages: the CNF 2026 Explorer.

The same data is a public MCP connector with no sign-in, for Claude, ChatGPT or any MCP client: how to add the NutraPlanner Recipe Analyzer connector. Its tools search foods, get a food, compare editions and total a recipe from CNF codes and amounts; nothing else of NutraPlanner is reachable through it.

NutraPlanner Recipe Analyzer connector (MCP) URL
https://nutraplanner.com/api/public/v1/cnf/mcp

AI assistants (MCP)

NutraPlanner runs a Model Context Protocol (MCP) server, so an AI assistant that supports connectors can search foods and recipes, draft recipes, build meal-plan drafts and draft services for your practice. This is separate from the calculator API above, and it needs a NutraPlanner account.

What it does for a practice, and which plans include it: AI assistants in NutraPlanner.

Server address

Use this address when an app asks for the server URL. Always use nutraplanner.ca without "www": a connection made through www is not accepted.

NutraPlanner MCP server address
https://nutraplanner.com/api/mcp

How to connect

  • Connector with sign-in (recommended). In Claude on the web, Claude Desktop or ChatGPT, add a custom connector, paste the server address and connect it. You are sent to NutraPlanner to sign in and approve. The steps and menu names differ by app and plan, so follow the app's own instructions for custom connectors.
  • Claude Code. Run this once in a terminal, then run /mcp inside Claude Code and choose NutraPlanner to sign in.
Claude Code command
claude mcp add --transport http nutraplanner https://nutraplanner.com/api/mcp

  • Personal access token. For tools that take a configuration file or an HTTP header instead of a connector, create a token under Connected apps in your NutraPlanner settings and send it as a bearer token in the Authorization header.

Approval and second factor

Connecting with sign-in shows an approval screen that lists what the app asked for. You approve with a code from your second factor (an authenticator app or an emailed code), so a hijacked session alone cannot connect an app. (NutraPlanner's own demo accounts for app-directory reviewers approve without a code; they can never be given client access.) A personal access token needs the same code when you create it. You can revoke any connection or token at any time under Connected apps.

What an assistant can access

Access is limited to what you choose when you approve or create the connection, and never exceeds what your own account can do.

  • Read your recipes.
  • Save recipe drafts for you to review.
  • Search the food database.
  • Save your own foods and food lists as drafts for you to review: no diet or allergy check can use a drafted list, and nothing a client receives shows a drafted food until you review it.
  • Read the diets and allergies a recipe or plan can be checked against.
  • Read your meal-plan templates.
  • Create and edit meal-plan templates.
  • Read your services, prices, hours, programs, forms and practice settings, never a client's record or your identity details.
  • Draft services for you to review and activate. Off unless you turn it on.
  • Draft intake forms for you to review: a drafted form is never sent until you review it. Off unless you turn it on.
  • Draft programs for you to review: no session can open until you review each module. Off unless you turn it on.
  • Propose changes to your settings for you to apply: nothing changes until you apply them in NutraPlanner. It needs the access to read your practice setup, which is ticked with it. Off unless you turn it on.
  • Save message snippets for you: a snippet is never sent on its own, and the assistant never reads yours. Off unless you turn it on.

Tools

These are the tools an assistant can call. It sees only the ones your connection's access and your plan allow, titled in your account's language. Everything it writes is a draft for you to review in NutraPlanner.

  • search_foods Searches the food database, or lists your practice's own foods with whether each is still an AI draft. It can also list food lists (your own and NutraPlanner's, by name) and show what one of your own lists holds and whether a diet or allergy check uses it. It never returns nutrient values, nor what a NutraPlanner list holds.
  • save_food Saves a food of your own as an AI draft for your review, from a Nutrition Facts label you share, a web page you name (NutraPlanner never opens it) or a copy of a NutraPlanner food, never from the assistant's memory. Energy, protein, fat, carbohydrate and sodium are required; a value the label doesn't give stays unknown, never zero. Nothing a client receives shows it until you review it. It changes only an unedited AI draft that you created through an assistant, whichever of your connected apps or tokens made it.
  • save_food_list Creates a food list of your own as an AI draft, or changes one that no diet or allergy check uses: it adds or removes foods and nests your own lists or NutraPlanner's, up to a hundred changes at a time. A changed list is an AI draft again, and no diet or allergy check can use it until you review it. The assistant sees only the name of a NutraPlanner list, never what it holds, and a list already in use changes only in NutraPlanner.
  • search_recipes Searches your recipes and the NutraPlanner library, by diet, meal, time and nutrients, in the order you choose.
  • get_recipe Reads one recipe with its ingredients and nutrition. It can also re-size the recipe to a number of servings and give its amounts in metric or US measures, worded as NutraPlanner prints them; the nutrition per serving stays the same.
  • analyze_recipe Works out the nutrition of a recipe without saving it.
  • save_recipe Saves a recipe as a draft for your review. An adapted recipe keeps a note of what changed from the original.
  • import_recipes Brings your own recipes over as private drafts, a few at a time: the assistant reads each one, keeping every ingredient line as written, and NutraPlanner matches each line to a food and works out the nutrition. A line it cannot match can be saved by name for you to finish. The recipes are checked as usual, but the assistant is not told how they fare against diets or allergies until you review them, and NutraPlanner never opens a web address you give it.
  • tag_recipes Adds or removes meal types and occasions (such as Christmas) on your own recipes, and reports the tags each recipe had so the change can be undone. It never changes a recipe's ingredients, nutrition or diet checks, and never sets a diet, allergy or religious tag.
  • list_profiles Lists the diets and allergies recipes and plans can be checked against, and which of them NutraPlanner enforces, removing any dish that fails one.
  • create_meal_plan Creates a meal-plan template and fills it from the recipe library. It can also make a variant of one of your templates, or of a library template, for an allergy or a diet pattern NutraPlanner enforces by removing any dish that fails it (e.g. gluten-free, vegan), and only from a template whose own diets and allergies NutraPlanner enforces too: a new draft without the dishes that do not fit, refilled where a fitting dish exists, leaving the original as it was.
  • get_meal_plan Reads a meal plan with its daily totals, and gives you a link to print a template from NutraPlanner. The assistant never receives a printout. For a template it can also give the grocery list, week by week and aisle by aisle in your own units, and each day's nutrition against the template's own target, with healthy-adult reference ranges beside it: a range to read, not a goal, and not for children or clinical diets.
  • search_meal_plans Searches meal-plan templates — the published library, your practice’s shared ones and your own — by name and diet, in the order you choose.
  • suggest_dishes Offers ranked recipes for one meal slot of a plan.
  • set_dish Places or removes one dish in a plan.
  • regenerate_meal_plan Fills empty slots, re-rolls days or rebalances portions in a plan. On a template an assistant made, it can also remove the dishes that no longer pass its enforced allergies and diet patterns, and any imported recipe you haven't reviewed yet. It refills a slot when it can, and reports any it leaves empty or lighter. It never redoes a template you have changed in NutraPlanner since server version 2.5.0.
  • get_client_context Reads what an assistant may know about one client you selected in the practice panel, by a random code: dietary restrictions, rounded targets and plans — never a name, contact details, dates or notes. No connection is given client information, so no connection lists this tool. Not available: no connection is given client information.
  • propose_appointment Proposes a visit for one client you selected in the practice panel. Nothing is booked: you confirm or decline it in NutraPlanner, and the assistant never gets the time back. Not available: no connection is given client information.
  • propose_task Proposes a task for you about one client you selected in the practice panel. It waits on your NutraPlanner dashboard, labelled as drafted by an AI assistant, and is never sent to the client. Not available: no connection is given client information.
  • get_practice_setup Reads your practice setup: your services (length, price, who may book them, whether each is live or an AI draft), their weekly hours and closed or special dates (never the reason you gave), the questions you added to your intake forms, your programs (their modules and structure, never who is enrolled), your reminder timings, branding colours and invoice terms. It never reads a client's record or your identity details, and never returns a booking link. In text people in your practice wrote, the first and last name of a client you can open or of a team member is replaced, and email addresses and phone numbers are removed, before it is sent; what the assistant drafted itself comes back as written.
  • draft_service Drafts a service with its weekly hours as an AI draft that nobody can book until you review and activate it in NutraPlanner, and can attach intake forms you already have if your plan and role allow it. It changes only AI drafts you haven't edited (from any assistant you've connected), and only what it sends. It never sets the currency, the name printed on invoices or the booking link.
  • draft_form Drafts an intake form as an AI draft that is never sent until you review it in NutraPlanner: the sections every drafted form keeps (the client's details, health history, medications and supplements, allergies and intolerances, goals and the consent wording for your province), any other standard section you ask for, and questions of its own. It never writes the consent or leaves out those sections, and it redoes only AI drafts you haven't edited (from any assistant you've connected), as a new version. It points out its own questions about allergies, disordered eating, weight or identification numbers, and any text written in one language only, and the review screen marks them. A form for clients in Québec can't be marked reviewed until every question has its French. It asks for lab values through the Recent lab results section, and refuses a question that writes a reference range or says whether a value is normal.
  • draft_program Drafts a program as an AI draft that no session can open until you review it in NutraPlanner, module by module: its name, length and delivery, modules by day with a lesson each, and lesson notes, meal plan templates or your forms as items. It never writes the public description, a price, the visibility or a session, and it changes only drafts of its own you haven't edited or reviewed. It flags wording for you to edit or keep (promised results, diet or restriction words, weight-stigmatizing words, calorie or per-kg numbers, named priority allergens); in a program about a medication it refuses medication advice and adds two safety lines you can edit in NutraPlanner.
  • propose_practice_change Proposes changes to your live settings in a change set you apply in NutraPlanner: a service's price, booking and cancellation terms (loosening only), weekly hours and closed or special dates; your branding colours; your payment terms, payment days and invoice prefix; your reminder times. Nothing changes until you apply the set, and you see which changes concern existing bookings or email your clients before you do. It never learns whether a booking exists, and it never writes contact details into your payment terms.
  • save_snippet Saves a message snippet you can insert into a message or a note, or changes one it saved that you haven't edited. A snippet is never sent on its own, and the assistant never reads your snippets. Snippets it saves show "From your AI assistant" until you edit them; in a practice that runs a medication program it refuses medication advice.
  • open_practice_panel Shows the practice panel inside the assistant, where you choose clients yourself. The assistant receives only a random code for the client you choose. Not available: no connection is given client information.
  • create_panel_ticket Used only by the practice panel to sign itself in to NutraPlanner. It returns no client information. Not available: no connection is given client information.

Prompts

Starting points the assistant can offer you, in your account's language. Claude shows them; ChatGPT does not show MCP prompts, so there you ask in your own words. You see only the ones your connection's access and your plan allow.

  • get-started Reads what this connection can do and offers you a few first tasks.
  • plan-a-week Builds a meal-plan template from a goal and fills it from the recipe library.
  • plan-a-series Builds a series of weekly meal-plan templates that differ in their dishes, all on the same daily target unless you set another for a week, after asking you what should change from one week to the next.
  • write-a-recipe Writes a recipe and saves it as a draft, with its nutrition and diet checks.
  • adapt-a-recipe Adapts an existing recipe and saves the result as a new draft, leaving the original as it was.
  • import-recipes Brings your own recipes over as private drafts, a few at a time, keeping each ingredient line as written. A line NutraPlanner cannot match is matched with you, or saved by name for you to finish.
  • set-up-a-service Drafts the services you describe, with their hours, for you to review and activate.
  • draft-an-intake-form Drafts an intake form after asking which languages your clients answer in and which standard sections you want. The form is never sent until you review it.
  • build-a-program Builds a program after asking its purpose, its length, cohort or self-paced, and what each module hands over. No session can open until you review each module.

Resources

Documents an assistant can read as context.

  • nutraplanner://connection What this connection can do: its access, the tools and prompts it sees, the plan features it lacks, whether client work is possible here, and what is left of today's limits. Reading it needs no particular access.
  • nutraplanner://profiles The diets and allergies recipes and plans can be checked against, and which of them NutraPlanner enforces, removing any dish that fails one.

Limits

Each tool has a per-minute limit, and some have a daily limit per person and per practice. When a call is refused for going over a limit, the error says when the assistant can try again, and a well-behaved assistant waits and retries.

Versioning

The server is version 3.1.0. Within a major version, changes are additive: new tools and new optional fields, with existing ones left as they are. ChatGPT keeps a copy of a connector's tool list, so after an update, open the connector in ChatGPT and press Refresh. A new kind of access is not added to an existing connection: reconnect the app, or create a new token, and tick it.

Privacy

Calls are limited to your own account and workspace. We record each call so we can investigate misuse: when, which tool, with which credential and how it went, kept 12 months. The request and our response themselves are kept 90 days, and only when the assistant saves or analyses a recipe. No connection is given client information. See the privacy policy.

Support

Questions about connecting, or a problem with a tool? Write to hello@nutraplanner.com.