Skip to main content

Nutrition API · private preview

Voice food logging API for health apps.

Let users say what they ate. Your app uploads the voice note, and BiteRight returns the foods, an estimated portion, calories and nutrients as JSON, in English or Arabic.

Private preview. The team confirms fit, sandbox access and setup; an enquiry does not issue keys. Production access, final specifications and pricing are not available yet.

From a voice note to a food log.

  1. Record

    Your app records a short voice note, for example “two eggs, wholemeal toast and a flat white”.

  2. Send

    Upload the note with the meal date to the voice food-log endpoint over HTTPS.

  3. Interpret

    BiteRight turns the spoken description into foods, likely ingredients and portion estimates.

  4. Store

    Your app receives the food log as JSON, with calories and nutrients, and displays it in its own way.

Nutrient values come from curated food-composition records, not from a generated answer. See the food-data sources and which nutrients come back.

Why add voice meal logging.

No database search

Users describe the meal instead of scrolling through near-identical database entries.

Works without a photo

A meal that is already eaten, a dark restaurant or a packed lunch can still be logged by describing it.

Captures what a photo hides

Users can mention the oil, butter, sugar or sauce that a camera cannot see.

Easier for some users

Speaking can suit people who find typing on a small screen hard. Test this with your own users before relying on it.

A voice request.

Send the meal date and one or more voice notes as an authenticated multipart upload. Each team uses its assigned base URL and access token.

date
Meal date, for example 2026-11-02
voices[]
Voice note file(s), for example .m4a

Before the first voice request, set up a dietary profile and daily intake (RDI). The response has the same shape as photo and text logs: the recognised dish, total weight, ingredients and nutrient objects of the form { name, amount, unit }.

Illustrative example from the private-preview documentation, not a final specification.

curl --request POST \
  --url "$BASE_URL/api/b2b/v1/food-log-voice" \
  --header "Authorization: Bearer $ACCESS_TOKEN" \
  --form 'date=2026-11-02' \
  --form 'voices[]=@breakfast.m4a'

English and Arabic voice input.

Users can describe meals in English or Arabic. The food data includes a Bahraini database and localised Middle Eastern dishes, and mixed dishes are broken into likely ingredients. Test with meals typical for your users during evaluation.

Arabic and Middle Eastern food →

Photo and text too.

Voice sits alongside photo and text logging, so users can pick the input that fits the moment. All three return food and nutrient estimates for your app to display and store.

Food recognition API →

Voice food logging API questions.

Add voice logging to your app.

Share your product, your users' languages and cuisines, and expected volume. The team reviews the fit and confirms sandbox access and setup.

Please leave out patient or personal health information.