IT-Diet – Next Feature: AI-Based Calorie Estimation
I’m extending my IT-Diet system with automatic calorie estimation.
The idea: The meal is photographed, the AI recognizes the components, and the nutritional values are estimated.
Since I have a ChatGPT Pro subscription but no API access, my workflow looks like this:
Drag the photo from a folder into the chat
Paste the prompt from the clipboard
Download the result as JSON
The JSON is then automatically imported into the system
The only manual step is uploading the photo into the chat window – everything else runs automatically.
Example
Here is the result that the AI automatically determined from the photos. It’s incredible.
Perfect – let’s break down the dish into its individual components and estimate the calories:
**Plate contents**
4 small boiled potatoes
One small potato (\~80 g) has about 70 kcal.
4 pieces = approx. 280 kcal
Herring in tomato sauce (about 3 pieces, \~100 g)
Herring (in sauce) has about 200 kcal per 100 g.
Estimated amount here: \~100 g → 200 kcal
Blue cheese (\~30 g piece)
Blue cheese has about 350 kcal per 100 g.
30 g = 105 kcal
Butter (\~10 g piece)
Butter has about 720 kcal per 100 g.
10 g = 72 kcal
**Drink**
Red wine (\~150 ml glass)
Red wine has about 85 kcal per 100 ml.
150 ml = 128 kcal
**Total**
Potatoes: 280 kcal
Herring in tomato sauce: 200 kcal
Cheese: 105 kcal
Butter: 72 kcal
Red wine: 128 kcal
Total: approx. 785 kcal
