Creates account
Chooses privacy defaults and confirms age eligibility. Continue with Apple authenticates only.
The flows make the strategy operational: memory first, exact inputs remain available, uncertainty is reviewed, and no weekly recommendation bypasses data validation.
Solid cards are actions. Dashed cards are decisions. The coloured top rule marks a trust-sensitive checkpoint.
Set a credible starting point, then hand off to the real home screen.
Chooses privacy defaults and confirms age eligibility. Continue with Apple authenticates only.
Enter editable fields manually or choose the optional Apple Health assist.
The requested body and activity categories are explained before the separate system permission.
Imported height and weight stay editable. Missing fields and private equation sex are entered manually.
Lose, maintain, gain, or recomposition. Unsafe rates are unavailable.
Chooses a landing weight for Lose, Maintain, or Gain; Recomposition uses an optional body-fat percentage with a comparable measurement source.
Available recent steps and workouts suggest a category that the person confirms or edits; otherwise use the manual questions.
Sees the basis and can choose a manual target.
Eligible and manual-only paths preserve their source context. Each shows exactly one CTA: Go to Today.
Food logging is available as a primary home-screen action, outside onboarding.
The primary memory-first path for recurring food.
Based on time, day, and the person’s own history—not a public guess.
One tap repeats. Edit opens portion and ingredient controls.
Change amount, remove an item, or save this as a new variation.
The new Food Log Events point back to the Meal Template version used.
Today refreshes calories, protein, and next action; Food preserves the itemized daily record.
When exact data exists, prediction should get out of the way.
Both inputs remain prominent on the add-food surface.
Personal history first, then verified regional data, then public entries.
Gram, serving, barcode, and provenance are visible.
A local correction can be saved without corrupting the public source.
Weight and common portions remain interchangeable.
The item becomes a Food Log Event. The person can later group events or save a reusable Meal Template.
Speed without pretending that spoken portions are exact.
“Chicken wrap with cheese and a small latte.”
Matches items from personal memory and credible food sources.
Low-confidence food or portion estimates are visually distinct.
If the estimate can materially change calories or protein, confirmation is required.
Uses natural units, grams, or a remembered portion.
Only confirmed values enter weekly analysis.
Create once, then let each person log the amount on their own plate.
Adds ingredients and records final prepared weight.
Sends a private link to a partner.
Sees ingredients and nutrition source, not the cook’s targets or diary.
350 g and 500 g can come from the same prepared batch.
Each log belongs to its person; the shared object remains intact.
Update the new batch or repeat the previous version.
The weekly value loop and the product’s most important integrity safeguard.
Separates supported goal outcomes from controllable behaviours such as validated logging, protein coverage, steps against personal intent, and training consistency.
Missing dinner, unusually low intake, or unconfirmed drafts become review items.
Confirms intake completeness separately from any corrections or review status.
Shows exactly which days and measurements remain in the analysis.
The calendar week remains context. Eligible calibration uses the approved rolling evidence window and separates observation, exclusion, interpretation, and confidence.
Types or speaks optional constraints, then confirms the editable transcript and every structured label before use.
Keep, adjust, collect more evidence, flexible week, or manual-only review is explained with exact values and dates.
No target changes before acceptance. The final decision remains inspectable and undoable.
Connect weekly progress, measures, check-ins, and plan evidence without overstating causality.
Reviews averages, totals, coverage, and the accepted plan decision.
Completed weeks enter validation before any recommendation is generated.
Weight and compatible body measurements remain available; private progress photos are marked Premium.
Inspects baseline inputs, sources, target origin, formula version, and decision history.
Keep the plan, archive it and confirm a new goal, destination, and baseline, or reset recommendation calibration through a separate warning flow.
New-plan and reset actions retain food, measures, photos, check-ins, prior plans, and manual-only safety eligibility.
The neutral wireframes sit with the flows they implement. They keep product structure distinct from the Organic site language while making the confirmed architecture reviewable.
Apple Health consent, editable body data, manual-only safety, four goals and destinations, target review, setup acknowledgement, and the handoff to Today.
Daily consumption, weekly context, direct logging, unified food entry, reconciled diaries, weekly progress, plan management, and Premium photo moments.
Empty states explain what to log and why. They never manufacture an insight.
Drafts, suspicious days, and weak sources remain visibly unresolved.
“Keep the plan” is a valid recommendation when confidence is low.