# Product ## Platform ios The customer product is a native iOS app. Web surfaces are companion landing pages, marketing content, product education, and research material rather than a web version of the tracker. ## Users The beachhead audience is English-speaking serious repeat nutrition trackers who: - track calories and protein most days for weight loss, weight maintenance, weight gain, body recomposition, or performance; - cook recurring meals and value exact portions, saved routines, and a compact daily record; - often share prepared meals with a partner while keeping goals, portions, diaries, and recommendations private; - have experience with products such as MyFitnessPal, Lose It!, Lifesum, or MacroFactor and are frustrated by repetitive logging, uncertain food data, or advice based on incomplete logs; - will correct questionable data when correction is fast, inspectable, and reusable; - want progress explained without food moralizing, shame, or compulsory coaching. Clinically supervised users and coach-led workflows are not the initial audience. Automated nutrition recommendations are intended only for eligible adults within the product's documented safety boundary. ## Product Purpose Build the most frictionless calorie tracker for serious nutrition users while protecting the integrity of the data used for guidance. The product helps a person answer three questions: 1. **Today:** What should I do next? 2. **Food:** What did I record on this day? 3. **Journal:** How is my plan progressing, and what should I review next? Success means familiar food takes seconds to log, the entire day remains easy to read, shared meals do not require duplicate recipe entry, and weekly guidance is offered only when the available food and body data can support it. ## Positioning This is a memory-first nutrition decision system, not another generic AI calorie tracker. Its differentiated mechanism connects immutable Food Log Events, optional reusable Meal Templates, private per-person portions, correction history and provenance, explicit day-validity states, explainable weekly decisions, and longitudinal Journal evidence. Exact inputs take priority when exact data exists; assistance is used when uncertainty already exists and remains editable. The product promise to protect is: > Remember what works. Show what changed. Recommend only what the data can support. The initial wedge is a tracker that remembers recurring meals and lets partners share one prepared meal while logging different private portions. The deeper product value is trust: recommendations never silently use incomplete days or present estimates as facts. ## Operating Context - The primary product is a native iOS app with optional Apple Health input for activity classification and body data. - Daily use happens around meals, shopping, cooking, meal preparation, and retrospective diary review. Common tools include a food scale, barcode scanner, search, voice input, saved meals, recipes, and prepared batch weights. - A Food Log Event preserves what was recorded at a point in time. A Meal Group can organize selected events without replacing them. A reusable Meal Template is created only when the person chooses to save a group or recipe for future use. - Shared prepared meals use one recipe or Meal Template as the source while each person records an independent portion. Sharing must not reveal either person's weight, goals, other meals, diary, or recommendation history. - Weekly check-in requires the person to review suspicious days, confirm intake completeness, and preserve correction history before quantitative coaching. - Journal brings weekly progress, check-ins, measurements, progress photos, current-plan evidence, and previous decisions into one private record while distinguishing observation from possible explanation. - Web landing pages and marketing content explain the product, support acquisition, and communicate migration, privacy, trust, and product value. They do not substitute for the native tracker. ## Capabilities and Constraints ### Current product direction - The home screen has two layers. Today is primary: it shows the current diary state, calories and protein, every logged item, per-meal totals, and an explicit start/continue tracking action. A compact "This week so far" layer is secondary: it shows clearly labelled weekly averages or accumulated totals, data coverage, and progress toward the person's weekly intent without judging an individual day. - Food is the second top-level tab. It owns the detailed diary for today and past dates, item selection, Meal Group creation, selected-item nutrition and sharing, daily steps/weight/body-fat context, and reversible day reconciliation. Review status remains separate from intake completeness. The current-plan start-versus-now summary belongs in Journal. - Journal is the third top-level tab. It owns week-by-week progress, the Weekly check-in entry point, measurements and premium progress photos, current-plan review, new-plan creation, and recommendation-algorithm reset. - Every goal asks for a landing destination during onboarding and new-plan setup: weight for Lose, Maintain, and Gain; body-fat percentage for Recomposition. When Recomposition lacks a credible sourced baseline, the person can explicitly defer the destination. The destination is user intent, not a guaranteed outcome or an instruction to exceed safety and pace limits. - Starting a new plan archives the current plan, preserves its diary and Journal history, and establishes a new confirmed baseline. Resetting the recommendation algorithm preserves the current plan and all user records but irreversibly clears its learned calibration before rebuilding from confirmed inputs. - Daily weight and body-composition signals, calories, protein, alcohol, steps, training type and minutes, and qualitative notes can roll up into the weekly layer. Missing or intentionally untracked days are never treated as zero, and the product does not create a moralized daily score. - Repeat logging, search, barcode scanning, and confirmed voice drafts create Food Log Events. Repeating a Meal Template creates new events and never rewrites prior history. - Familiar meals should be repeatable in no more than two taps and typically under ten seconds. - Recipes and prepared meals are gram-native: ingredient weights, final cooked weight, arbitrary portions, and arbitrary grams consumed remain supported. - Food records expose source and serving basis. Local corrections can be saved without corrupting a public source. - Users can browse, export, correct, and delete their complete history. Retention must come from value rather than data captivity. - Recommendations state their evidence, excluded data, uncertainty, confidence, and proposed change. A person can accept, defer, reject, edit, undo, or choose a manual target. - Suggested actions, commentary, patterns, and coaching remain configurable without removing the core tracking tools. - Tracking language is adherence-neutral and does not label foods or imperfect days as good or bad. ### Safety and data integrity - Missing intake is never treated as zero. - Drafts and uncertain values are not facts until reviewed. - Incomplete or intentionally untracked days do not silently drive quantitative target changes. - Progress photos are private by default. - Resetting the recommendation algorithm never bypasses automation eligibility. A manual-only user remains manual-only and receives no automated maintenance estimate or target. - Formula versions, recommendation inputs, evidence windows, and edits remain attributable and auditable. - Automated energy recommendations are not for people under 19, pregnancy or breastfeeding, a current or target BMI below 18.5, disclosed eating disorders or active treatment, or medical circumstances that materially alter nutrition needs. In these cases, every goal—including Maintain—uses a manual-target-only path: neutral tracking and manual targets may remain available with appropriate clinical direction, but the product must not show an automated neutral maintenance estimate. - The current recommendation algorithm is a researched v0.1 proposal and requires dietitian, medical-safety, implementation, and product validation before launch. ### Commercial model and open decisions - The trustworthy daily tracker—including core logging, barcode input, reusable meals, complete history, measurement tracking, and portable export—stays free. - Individual Premium is **$4.99 per month** or **$29.99 per year**. Its confirmed features are editable AI photo-recognition drafts and the private Photo Journal. - Duo is **$49 per year for two people**. It grants Premium to two linked accounts while each person keeps private goals, portions, diaries, measurements, and recommendations. Core meal sharing remains available without Duo. - The future pricing strategy may include an optional “Buy me a coffee” contribution for free users. It is not a current interface element and would not unlock product capability, change service, or create a subscription. - Trial design, regional storefront pricing, billing implementation, launch markets, and future premium boundaries remain open. - Coach collaboration, MCP integrations, broad AI coaching, social features, challenges, and clinician workflows are future possibilities outside the initial critical path. - The product has no confirmed name yet. “OpenCal” is not the product name. ## Brand Commitments - No product name is confirmed. - Voice is calm, direct, precise, non-moralizing, and transparent about uncertainty. - Assistance must preserve control: sources stay visible, probabilistic drafts stay editable, recommendations require approval, and undo remains available. - Privacy, data ownership, explainability, and the separation of observed facts from interpretation are product commitments rather than compliance footnotes. - Do not lead positioning with “AI calorie counting.” Lead with memory, exact repeat workflows, shared portions, and trustworthy decisions. ## Evidence on Hand - [`index.html`](index.html) — strategy index and current decision summary, including positioning, monetization, and beachhead audience. - [`COMPETITOR_RESEARCH.md`](research/COMPETITOR_RESEARCH.md) and [`competitor-research.html`](research/competitor-research.html) — evidence from MacroFactor, Lifesum, and MyFitnessPal users; current user needs, incumbent strengths, switching triggers, and product recommendations. - [`COMMERCIAL_ANALYSIS.md`](research/COMMERCIAL_ANALYSIS.md) and [`commercial-analysis.html`](research/commercial-analysis.html) — market model, beachhead audience, confirmed price ladder, go-to-market loops, commercial risks, roadmap, and decision gates. - [`calorie-app-prd-v0.2.md`](product/calorie-app-prd-v0.2.md) and [`product-vision.html`](product/product-vision.html) — living product vision, core jobs, Today/Food/Journal model, and food-record architecture. - [`recommendation-algorithm-v0.1.md`](product/recommendation-algorithm-v0.1.md) and [`algorithm-recommendation.html`](product/algorithm-recommendation.html) — researched initial energy recommendation, safety boundary, confidence model, adaptive-calibration proposal, and validation plan. - [`user-flows.html`](product/user-flows.html) — onboarding, repeat meal, exact search and barcode, voice draft, shared meal, weekly validation, Journal, trust checkpoints, and commercial boundaries. - [`home-screen-direction.md`](product/home-screen-direction.md) — confirmed two-layer home hierarchy, weekly roll-up measures, aggregation rules, and data-integrity constraints. - [`PRESS_RELEASE.md`](product/PRESS_RELEASE.md) — working-backwards narrative and launch promise. - [`benchmark/images/`](research/benchmark/images/) and [`benchmark/design-system-render/`](research/benchmark/design-system-render/) — competitor references and incumbent visual evidence. The repository does not contain a confirmed product name, launch date, spokesperson, real customer testimonial, availability statement, or launch market. Placeholder and illustrative press-release content must not be presented as fact. Market sizes and revenue cases are planning models, not forecasts. The confirmed US price ladder is $4.99 monthly or $29.99 annually for Individual Premium and $49 annually for Duo; regional prices remain open. ## Product Principles 1. **Memory over prediction.** Capture recurring behaviour once and make reuse faster than rediscovery. 2. **Exact when possible, honest when uncertain.** Prefer known foods, weights, sources, and portions; label and review estimates. 3. **Validate before advising.** Protect data integrity before calculating trends or changing a target. 4. **Show progress without judgment.** Celebrate consistency, preserve agency, and keep facts separate from interpretation. 5. **The user owns the record.** Keep history inspectable, correctable, exportable, private, and reversible. ## Accessibility & Inclusion - Guidance must avoid shame, compulsory streaks, diet assumptions, and good/bad food labels. - Users can independently hide or disable calorie visibility, streaks, ratings, insights, or recommendations where supported. - The app must explain why an energy equation requests sex-specific input, keep that value private, and offer a manual target path. - Safety-sensitive users must not receive an automated prescription outside the documented eligibility boundary. - No product-specific accessibility conformance target has yet been confirmed; native iOS and web accessibility requirements remain an explicit open decision.