Skip to content

[Feature]: AI meal logging from narrated meal descriptions #1703

Description

@sacgsxr

What problem does this solve?

Meal logging is still a little cumbersome when the meal is easy to describe but inconvenient to photograph or enter item-by-item. For example, I may know exactly what I ate and want to quickly narrate it:

1/2 grilled chicken breast, a handful of mixed greens, a slice of cheddar cheese, 1 small Roma tomato, one jalapeno...

Today the AI photo flow can estimate ingredients from an image, but there does not appear to be an equivalent text/voice-first path where I can describe a meal naturally and have SparkyFitness turn that into editable logged ingredients.

Describe the feature you'd like

Add an AI-assisted meal logging option that accepts a narrated or typed meal description and converts it into structured food/ingredient entries, similar to how AI food photo estimation identifies ingredients from a photograph.

Desired workflow:

  1. User taps a "describe meal" / voice input option from meal logging.
  2. User speaks or types a natural-language meal description.
  3. SparkyAI parses the description into ingredients, quantities, and likely serving sizes.
  4. The app shows an editable review screen before logging.
  5. User can adjust ingredients, quantities, serving units, and nutrition matches before saving.

Example input:

1/2 grilled chicken breast, handful of mixed greens, slice of cheddar cheese, 1 small Roma tomato, one jalapeno

Example parsed result:

  • Grilled chicken breast, 0.5 breast or estimated grams
  • Mixed greens, 1 handful / estimated cups or grams
  • Cheddar cheese, 1 slice
  • Roma tomato, 1 small
  • Jalapeno, 1 pepper

This could reuse much of the existing AI food-photo estimation/review behavior, but with natural-language text or speech as the input instead of an image.

Which area does this relate to?

  • Mobile app (iOS)
  • Mobile app (Android)
  • Server / API
  • AI / SparkyAI
  • Nutrition tracking

How important is this to you?

Would significantly improve my workflow

Alternatives you've considered

Current workarounds are either logging each item manually or taking a photo even when the meal is easier to describe than photograph. Voice dictation into a generic text field may capture the words, but it still does not parse the meal into structured, editable ingredients.

Additional context

Related ideas/features:

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    Status
    Backlog

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions