Logging a Meal Without a Photo: How Voice and Text Input Work

CaloTec-AI does not require a photo of your plate. You can type a meal description or dictate it, review the text, and submit it for analysis. Here is the actual flow, including the points that require your review.

Text Input

Enter what you ate in the meal field. A useful description includes the dish, any known amount, and extras: "two slices of bread with one egg and a teaspoon of oil" provides more information than "sandwich."

After submission, an AI service processes the description and returns estimated food items and nutrition values. A new analysis requires an internet connection.

Voice Input

Tapping the microphone starts Android's speech-recognition service. Depending on the device and available service, audio may be processed on-device or online. CaloTec-AI receives the transcript, places it in the text field, and does not submit it for analysis on the user's behalf.

Pause and read the field before submission. Food names, numbers, and units can be transcribed differently from what you said. You can correct the text or add a missing detail.

What the AI Service Receives

When you request an analysis, the food description is sent to Gemini. If Gemini is unavailable, the app may contact DeepSeek as a backup. The service returns an estimate, not a photograph or a direct measurement of the meal.

How to Review the Result

Check the food items, amounts, and values before saving. With a home recipe or restaurant meal, the system may assume a different serving size or recipe from the one you had.

You can edit the items. If the result is far from the meal, submitting a clearer ingredient list can be more useful than correcting one number.

No Photo Does Not Mean No Data Transfer

Avoiding a photo reduces the type of information you provide, but the text still leaves the device for AI analysis. Read the current privacy policy and do not include personal information that is not needed to analyze the food.