Track nutrition and fitness via webhook with OpenAI, Google Sheets and Slack alerts

Webhook receives user input → AI analyzes food/activity → Google Sheets updates → Slack alerts on limits/risks → daily scheduled reset

Who

Fitness app developers, health platform operators, personal trainers, nutritionists, automation engineers who need automated diet and activity tracking

You get

After a user sends a food or activity description, the system automatically estimates calories and protein, updates the daily summary in Google Sheets, sends Slack alerts when thresholds are exceeded or health risks are detected, and resets statistics daily via schedule trigger

You need

n8n · google-sheets · slack

Steps

1. Configure Webhook to receive user input

   Add a Webhook node (Receive User Input) in n8n, set path to e.g. /diet-input, and set responseMode to responseNode. This endpoint will receive POST requests containing phone and message.

2. Connect Google Sheets and map user data

   Add a Google Sheets node (Fetch User Data), select Search operation, match the Phone column against incoming body.phone. Enter your document ID and sheet name, ensuring column names match the template.

3. Validate user exists

   Add an If node (Validate User Exists), check whether the previous output object exists ($json exists). Only continue if the user is found in the sheet, avoiding unnecessary AI processing for unregistered users.

4. Configure AI Nutrition Analyzer

   Add an AI Agent node (AI Nutrition Analyzer), set promptType to define, and paste the following prompt into the Text field. This node must be connected to an OpenAI Chat Model sub-node.

   Execution prompt · Source text。AI Nutrition Analyzer node Text field (promptType=define)

   Copy prompt

   =You are an advanced nutrition and fitness AI.
   
   User Goal: {{ $json.Goal }}
   Daily Calorie Target: {{ $json.Daily_Calorie_Target }}
   
   User Input:
   "{{ $('Receive User Input').item.json.body.message }}"
   
   Tasks:
   1. Detect type:
      - FOOD
      - ACTIVITY
      - UNKNOWN
   
   2. If FOOD:
      - Estimate calories
      - Estimate protein
      - Detect meal type (breakfast/lunch/dinner/snack)
   
   3. If ACTIVITY:
      - Estimate calories burned
   
   4. Detect risk:
      - OVER_EATING
      - LOW_PROTEIN
      - UNHEALTHY
      - NONE
   
   Return STRICT JSON:
   
   {
     "type": "FOOD | ACTIVITY | UNKNOWN",
     "calories": number,
     "protein": number,
     "calories_burned": number,
     "meal_type": "breakfast | lunch | dinner | snack",
     "risk": "NONE | OVER_EATING | LOW_PROTEIN | UNHEALTHY",
     "suggestion": "short advice",
     "reply": "friendly message"
   }

5. Configure AI model and credentials

   Add an OpenAI Chat Model sub-node (AI Model (GPT)), select gpt-4.1-mini (or gpt-4.1 / gpt-4), and connect OpenAI API credentials. Ensure the credential is active and has available quota.

6. Process nutrition data

   Add a Code node (Process Nutrition Data), and run the following JavaScript to merge AI-returned JSON with existing user data, calculating cumulative calories, protein, and risk count.

   Execution prompt · Source text。Process Nutrition Data node JS Code field

   Copy prompt

   const raw = $input.first().json.output;
   
   let parsed;
   
   try {
     parsed = JSON.parse(raw);
   } catch (e) {
     parsed = {
       type: "UNKNOWN",
       calories: 0,
       protein: 0,
       calories_burned: 0,
       risk: "NONE",
       suggestion: "AI parsing failed",
       reply: "Sorry, I couldn't understand. Try again."
     };
   }
   
   // Get user data
   let user = $('Fetch User Data').first().json;
   
   let totalCalories = parseInt(user.Consumed_Calories || 0);
   let protein = parseInt(user.Protein || 0);
   let riskCount = parseInt(user.Risk_Count || 0);
   
   // FOOD
   if (parsed.type === "FOOD") {
     totalCalories += parsed.calories;
     protein += parsed.protein;
   }
   
   // ACTIVITY
   if (parsed.type === "ACTIVITY") {
     totalCalories -= parsed.calories_burned;
   }
   
   // Risk tracking
   if (parsed.risk !== "NONE") {
     riskCount += 1;
   }
   
   return [{
     json: {
       ...user,
       ...parsed,
       Consumed_Calories: totalCalories,
       Protein: protein,
       Risk_Count: riskCount,
       message: $('Receive User Input').first().json.body.message
     }
   }];

7. Check calorie threshold and alert

   Add an If node (Check Calorie Threshold), checking if Consumed_Calories > Daily_Calorie_Target. If exceeded, trigger a Slack node (Alert: Calorie Limit Exceeded) to send an alert.

   Execution prompt · Source text。Alert: Calorie Limit Exceeded node Text field

   Copy prompt

   =You have exceeded your daily calorie limit.  Consumed: {{ $('Process Nutrition Data').item.json.Consumed_Calories }} Target: {{ $('Process Nutrition Data').item.json.Daily_Calorie_Target }}
   {{ $('Process Nutrition Data').item.json.suggestion }}
   
   
   
   Calorie Limit Exceeded
   
   User: {{$json.Name}}
   Consumed: {{$json.Consumed_Calories}}
   Target: {{$json.Daily_Calorie_Target}}
   
   {{$json.suggestion}}

8. Detect health risks and escalate alerts

   Add an If node (Check Health Risk) checking if risk != 'NONE', triggering a Slack node (Alert: Health Risk). Then add an If node (Check Critical Risk) checking if Risk_Count >= 3, triggering a Slack node (Alert: Critical Health Issue) for critical escalation.

   Execution prompt · Source text。Alert: Health Risk node Text field

   Copy prompt

   =Health Risk Detected  User: {{ $('Process Nutrition Data').item.json.Name }} Risk: {{ $('Process Nutrition Data').item.json.risk }}  
   {{ $('Process Nutrition Data').item.json.suggestion }}

   Execution prompt · Source text。Alert: Critical Health Issue node Text field

   Copy prompt

   =CRITICAL ALERT  User: {{ $('Process Nutrition Data').item.json.Name }} Multiple unhealthy patterns detected.  Immediate attention required.

9. Return API response and update user data

   Add a Respond to Webhook node (Send API Response), returning JSON with the AI reply, this meal's calories, and today's total. Then add a Google Sheets node (Update User Data), select Update operation, match by Phone column, and write back Consumed_Calories, Protein, Risk_Count, Status, and Last_Meal.

10. Configure daily scheduled reset

    Add a Schedule Trigger node (Daily Reset Trigger), set the daily execution time. Connect to a Google Sheets node (Fetch All Users) to read all users, then to a Google Sheets node (Reset Daily Stats) with Update operation, matching by Phone, resetting Consumed_Calories, Protein, Risk_Count to 0 and Status to NEW_DAY.

You are setting up an n8n AI nutrition tracking workflow. First confirm you have: 1) an n8n instance; 2) a Google Sheets with correct column names; 3) OpenAI API credentials; 4) Slack API credentials. Then import the workflow JSON or create nodes manually, replacing all document IDs, Sheet IDs, channel IDs, and credential references.