Food & nutrition

Create foods and recipes

Build reusable foods and recipes when the catalog is not the right match.

All devices

Your User Library is for foods and recipes you choose to keep. Review names, portions, ingredients, and nutrition before saving so the result matches what you intend to log.

Choose the right starting point

• Custom food: enter a food and its available nutrition directly.
• Recipe: assemble ingredients or review an imported recipe before saving it.
• Open Food Facts: search the external catalog and import a reviewed result into your User Library.

Choose the path that matches your source. A label, recipe page, dictated description, photo, and provider search each require a different review.

Describe a dish by voice

Use a picture on an eligible iPhone

These paths help discover recipe candidates. Neither path treats a spoken description or photograph as authoritative nutrition.

Mac presents manual, recipe, provider-assisted, and Open Food Facts starting points together.

URL import is one of the Mac entry points when a public recipe already exists.

Dictation can land in a reviewed food after you confirm the match.

Review before saving

Confirm the title, source, serving, ingredients, and nutrition that are available. Provider-assisted results are suggestions; a photograph or dictated description does not create authoritative nutrition automatically.

Open the candidate recipe

Accept only after reviewing it

Check the recipe source, ingredients, yield, and calculated nutrition before making it part of your library.

Keep the reviewed food for reuse

Reopen its nutrition at any time

Saving creates a reusable personal food. Review it again whenever the source recipe or your serving assumptions change.

iPhone and iPad

On an eligible iPhone, Take a Picture uses Apple Vision classification inside an Apple Intelligence workflow. The option is hidden when Apple Intelligence is unavailable. BeFit Tracker uses the analysis to search public recipe websites for possible matches. You choose and review the recipe before importing it. Mobile food entry can also use dictation or a recipe URL.

The same candidate-and-review structure is visible in the dictated workflow:

1. Review the description

2. Choose a trusted match

Correct the description first, then choose among recipe candidates rather than accepting the first result automatically.

3. Inspect ingredients and nutrition

4. Keep the reviewed food

The accepted recipe becomes a reusable food in your User Library.

Paste a recipe URL, review the retrieved recipe, and accept it only when the details look right.

Check the transcript before searching

Compare the recipe with the dish

Correct recognition errors before searching, then reject candidates whose ingredients or preparation do not match.

After accepting the reviewed recipe, choose the meal, date, and serving amount for this particular log entry.

On iPad, paste the public recipe URL and continue to the same source and nutrition review before accepting it.

A reviewed recipe remains inside a focused sheet on iPad before it is accepted into the User Library.

Mac

Mac can use a configured OpenAI, Claude, or local Ollama provider for supported assisted workflows. Provider availability, network access, and the model you configure affect the result.

Review the dictated query before sending it to the configured AI integration.

Compare candidate sources, ingredients, and serving assumptions before choosing a recipe.

A photo helps identify a likely dish; it does not measure ingredients, portion size, or nutrition from pixels.

Only the food you explicitly accept becomes a reusable User Library item.

Mac presents manual, recipe, provider-assisted, and Open Food Facts starting points together.

Dictation helps narrow the candidate recipe before you decide what to save.