# Sitemap - Tendd Phase IA, Prompt 1: Information Architecture. This is structure only: what entities exist, what screens exist, what job each screen serves, and how a person moves between them. Per-screen content structure, states, and shared components are Prompt 2, not this document. Every entity and every screen below traces to a job in user-research/docs/jtbd.md or a locked decision in research/docs/strategy.md (Section 6, D1 to D5). Anything only suspected is marked [?] and kept in an "Under question" list, not in the main structure. Anything explicitly cut from MVP scope gets no entity and no screen. Job shorthand (from user-research/docs/jtbd.md): - J-MAIN: see all recurring charges clearly in one calm view (fused J1+J2 of strategy.md) - J1: activate without anxiety, the trust job (trust before bank) - J2: identify and cancel unused subscriptions, the cut job - J3: understand what a charge actually is, the clarity job - J4: stay ahead of surprises, the alert job - J5: track subscriptions without sharing bank data, the privacy job - E1: feel competent, not judged - E2: the small win, the cancel-and-save pride moment - E3: feel safe and in control of data - S1: share the discovery (privacy-safe) - D1 to D5: locked founder decisions in strategy.md Section 6 --- ## Entities The core objects a person deals with to close their jobs. Screens only display entities, and entities grow out of jobs. This inventory is built from user-research/docs/personas.md, user-research/docs/jtbd.md, and research/docs/master-research.md. It is deliberately not built from a competitor's menu. ### 1. Subscription (central entity) - **Made of:** display name, real merchant name and logo, amount, billing frequency (monthly / quarterly / yearly / custom), next billing date, category, status (active / paused / trial / cancelled), source (bank-detected / added from preset / added custom), linked account. - **Job that produces it:** J-MAIN (see all recurring charges in one calm view). The real merchant name and logo fields exist because of J3 (understand what a charge is, replacing "SPOTIFYAB STOCKHOLM" with "Spotify Premium", master-research M3). - **Connects to:** Account (detected from or charged to), Alert (alerts fire on a subscription), Cancellation (a subscription can be cancelled), Charge (its billing instances), Snapshot (aggregated into count and total). ### 2. Account (data source or connection) - **Made of:** type (bank connection via Plaid, or a manual/private source), institution name, connection status (connected / syncing / error / needs reauth), last sync time, count of subscriptions detected. - **Job that produces it:** J1 (activate without anxiety, the bank path) and J5 (track without sharing bank data, where the "source" is a manual source rather than a bank). Build decision D5 (US first with Plaid; TrueLayer for EU designed in but built later). - **Connects to:** Subscription (subscriptions belong to a source), User (a user has one or more sources). - **Note:** the privacy path is a manual source, not a bank connection (D2). Free allows unlimited bank connections (no cap); Pro's value is depth, not connection count (product decision, July 2026; D3 gates on history, trends, and advanced alerts, never on the core visibility). ### 3. Preset (service template) - **Made of:** service name, logo, typical price, default billing frequency, category. - **Job that produces it:** J5 (the privacy path) via founder decision D2, which locks a launch preset library of 400+ services so manual entry is low effort. - **Connects to:** Subscription (a manually added subscription is instantiated from a preset), the Add Subscription screen. - **Note:** D2 makes preset library quality a launch requirement, not a nice-to-have. ### 4. Alert - **Made of:** type (price change and payment failed are free; trial ending, unusual charge, and duplicate are Pro per D3), related subscription, plain-language message (active voice, master-research M1), timestamp, read / unread, suggested next action. - **Job that produces it:** J4 (stay ahead of financial surprises). - **Connects to:** Subscription (an alert references a subscription), Charge (a payment-failure alert references a failed charge). ### 5. Charge (billing instance / payment record) - **Made of:** date, amount, related subscription, status (scheduled / paid / failed). - **Job that produces it:** J4 (a payment that failed or a price that changed references a specific charge) and the history and trends depth (D3, a Pro feature). At the free level only the next and last charge and any failure surface; the full timeline is Pro. - **Connects to:** Subscription (a charge belongs to a subscription), Alert (a failed charge triggers an alert), the History and Trends screen. ### 6. Cancellation - **Made of:** related subscription, method (basic instruction, meaning a link plus a step description, is free; the full step-by-step guide plus direct cancel link is Pro, per D3), progress status (not started / in progress / confirmed cancelled), money saved (monthly and yearly). - **Job that produces it:** J2 (identify and cancel unused subscriptions) and E2 (the small win, the money saved). - **Connects to:** Subscription (cancels a subscription), Snapshot (a cancel win can seed a shareable card). - **Note:** D3 keeps basic cancel instruction free. Only cancel depth is paid, because the cancel moment is a trust moment, not an upsell moment. ### 7. Snapshot (shareable card) - **Made of:** subscription count, monthly total, an optional recent cancel win, and deliberately no bank data or account numbers. - **Job that produces it:** S1 (share the discovery) and E2 (the pride moment). - **Connects to:** the Subscription collection (an aggregate), Cancellation (a win can seed a snapshot). - **Note:** whether sharing actually goes viral is a hypothesis [? master-research H2 and jtbd.md S1], but the privacy-safe card itself is a designed object the product must build. ### 8. User and data controls - **Made of:** minimal identity (Ravi wants no account required to try [? assumption from personas.md Ravi trust triggers]), activation path (bank or privacy), plan reference, notification preferences, plain-language data-access explanation (master-research M2), one-tap data deletion control. - **Job that produces it:** E3 (feel safe and in control of data) and J1 (trust). - **Connects to:** Account, Subscription (a user owns them), Plan. ### 9. Plan (Tendd Pro tier) - **Made of:** tier (Free / Pro monthly / Pro yearly / lifetime), price ($7.99/mo, $69/yr, lifetime price TBD [? D4 hypothesis $99-139]), status, renewal date. - **Job that produces it:** a business rule, not a user job directly. Traces to founder decisions D3 (the free vs paid split) and D4 (pricing). - **Connects to:** User, and every Pro-gated feature (History and Trends, advanced alerts, full cancel guides, export). --- ## Entities Under Question These are not in the main structure. They get no entity and no screen in this MVP concept. - **Household / shared subscription view** - JH1 [? hypothesis] in jtbd.md, Segment C in strategy.md, explicitly out of MVP scope. The business-model paid bullet mentions a "multi-account household view", but the underlying job is deferred, so no household entity is modelled here. Note: connecting more than one of a single user's own bank accounts is in scope (see Account); sharing across two people is not. - **Gmail scan source and AI screenshot import source** - [? v2]. Deferred by D2 until user testing confirms they are trusted more than bank connection. Not launch entities. - **AI assistant / conversational query** - JH2 [? hypothesis], explicitly cut in jtbd.md and ux-patterns.md (adds cognitive load, the opposite of the differentiator). - **Budget envelopes, spending goals, credit score, investment tracking, bill negotiation** - explicitly cut in jtbd.md Part 5 and CLAUDE.md out of scope. No entities. ### Note on a scope conflict surfaced while building this inventory (resolved) While building this inventory, CLAUDE.md's free tier still said "up to 10 tracked subscriptions" (and "connect up to 2 bank accounts"), which conflicted with locked decision D3 (strategy.md, June 14 2026): free includes full visibility with unlimited tracked subscriptions and connections, limited by depth and history and never by the core relief experience. This concept followed D3: the free tier is unlimited on subscription and connection count and gated only on history, trends, advanced alerts, cancel depth, and export. The conflict was resolved in the CJM Step 11 living-docs update (commit 0dac829): the caps were removed from CLAUDE.md and the free tier aligned to D3, so the brief and this concept now agree. --- ## Screens Derived from the entities above and from what each person is trying to do, not from any competitor's structure. Every screen names the job it serves in parentheses. A screen with no job would be tagged [ORPHAN]; there are none here (verified against the coverage matrix in the Traceability section). Persona need is marked per screen: - **[E]** Emma, the primary persona (anxious, bank-willing). Her core path is marked **[PRIMARY PATH]**. - **[R]** Ravi, secondary (privacy-first, manual path). - **[C]** Claudia, secondary (motivated cutter, cancels fast). Empty, error, and loading are states of a screen, not screens, so they do not appear here. They appear in the flows (ia/docs/flows.md) and the critique. ``` Onboarding and Activation (one-time chain, then never seen again) - Welcome / Value Intro (J1: show value before asking for data) [E][R][C] - Activation Path Choice (J1 + J5: two paths, equal weight, D2) [E][R] - Connect Bank (J1 bank path; D5 Plaid US first) [E][C] - Add Subscription (manual + presets) (J5 privacy path; D2 400+ presets) [R] (also reused in-app by all) - Guided Reveal (J-MAIN: the aha moment; gradual, D1) [E] [PRIMARY PATH] Core (the steady-state product) - Home / Subscription List (J-MAIN + E1: calm list, count, total) [E][R][C] [PRIMARY PATH] - Subscription Detail (J3 decode a charge + J4 its alerts) [E][R][C] Stay Ahead - Alerts / Activity (J4: price change, payment failed) [E][C] Cut and Celebrate - Cancel Guide (J2: identify and cancel; basic free, D3) [C][R][E] - Cancel Win Moment (E2: the small win, money saved) [C][E] - Share Snapshot (S1: privacy-safe share card) [E][C] Depth (Pro, gated by D3) - History and Trends (J-MAIN over time; Pro gate, D3) [E][C] - Upgrade / Tendd Pro (D3 split + D4 price: the paywall) [E][R][C] Account and Trust - Connections / Accounts (J1 + J5: manage sources, reauth; D5) [E][R] - Data and Privacy (E3: plain-language data use, delete) [R][E] - Settings / Profile (E3: plan, notification preferences) [E][R][C] ``` Notes on structure: - **Add Subscription is reused.** It is the privacy path in onboarding (Ravi) and the in-app "+" for anyone adding a subscription by hand later. Same screen, two entry points. It is not duplicated. - **Guided Reveal is one screen, not three.** D1 makes it gradual (count, then categories with logos, then total paired with an action). Those are sequential steps of one reveal experience, not separate screens and not states. The flow in ia/docs/flows.md shows the internal sequence. - **Emma's primary path is short:** Welcome, Path Choice, Connect Bank, Guided Reveal, then Home. In every session after the first, she lands straight on Home. Depth is measured in the Navigation section. - **Cancel depth nests under the cut job.** Cancel Guide leads to the Cancel Win Moment, which offers Share Snapshot. This chain is where E2 and S1 live. - **No screen for E1 by itself.** E1 (feel competent, not judged) is a tone requirement met by Home, Guided Reveal, and Cancel Win rather than a standalone screen. The coverage matrix records where it is served. --- ## Navigation Mobile-first. The global navigation is a bottom tab bar (thumb reach on a phone), scaling to a left rail on desktop. It is deliberately small so the product stays calm (design principle: one thing at a time). ### 1. Global navigation (4 items) Each item is an entry point into one main job cluster, chosen because a distinct job lives behind it, not because a category is conventional. | Tab | Opens | Job cluster behind it | Why it earns a global slot | |-----|-------|----------------------|----------------------------| | **Home** | Home / Subscription List | J-MAIN (see all recurring charges calmly) + E1 | The reason the product exists. It must be reachable in zero taps, so it is the default tab. | | **Alerts** | Alerts / Activity | J4 (stay ahead of surprises) | This is also the return engine. Research BP4: without an external hook this segment does not come back, so the surprise surface is promoted to a permanent home. | | **Save** | Cancel hub (Cancel Guide + cancel candidates + wins) | J2 (cut) + E2 (the win) | The emotional peak and the upgrade trigger live here. It is Claudia's primary job and Emma's secondary one, so it gets a calm, dedicated door instead of being buried in a menu. | | **You** | Settings / Profile (into Data and Privacy, Connections, Plan) | E3 (feel safe, control data) + J1 (trust) | Trust is ongoing, not just onboarding. Ravi needs one-tap data deletion and connection control to always be findable. | History and Trends does NOT get a global tab: it is Pro-gated (D3), and a mostly-locked tab would irritate free users and break the calm promise. It is reached contextually from Home. Upgrade is not a tab either: it appears at the gate, in context, not as a standing invitation to pay. ### 2. Depth to the main job **Main job (stated explicitly):** J-MAIN, "see all recurring charges clearly in one calm view." For Emma (primary persona) it is delivered by the Guided Reveal in the first session and by Home / Subscription List in every session after. First session (activation), counting taps from the first screen: ``` Welcome --tap 1--> Activation Path Choice --tap 2--> Connect Bank --tap 3--> Guided Reveal (sees the list) ``` That is **3 taps** to the main job. The bank authorization and sync happen inside the Connect Bank step and then the reveal plays automatically, so no extra tap is added. 3 is at the ceiling, not over it. Every session after the first: the app opens on **Home**, so the main job is at **0 to 1 tap**. The steady state is where Emma lives, and it is effectively zero depth. No restructure needed: both the first-session path (3) and the steady-state path (0 to 1) are within the three-tap limit. ### 3. Global, contextual, and deep | Visibility | Meaning | Screens | |-----------|---------|---------| | **Global** (always visible) | The 4 tab bar destinations | Home / Subscription List, Alerts / Activity, Save (Cancel hub), You (Settings / Profile) | | **Contextual** (appears inside a flow) | Reached by acting on something, not from the tab bar | Guided Reveal, Subscription Detail, Add Subscription, Cancel Guide, Cancel Win Moment, Share Snapshot, History and Trends, Upgrade / Tendd Pro | | **Deep** (rare, one-time or infrequent actions) | Two or more levels in, or seen once | Welcome, Activation Path Choice, Connect Bank (onboarding chain), Connections / Accounts, Data and Privacy, plan and billing management inside Settings, export inside History and Trends | The activation chain (Path Choice, Connect Bank, Add Subscription, Guided Reveal) shows no tab bar; it is a one-time linear chain that ends by handing the user to Home with the tab bar switched on. Welcome / Value Intro (updated 2026-07-04) sits in front of the chain as the public marketing landing (its own full-width top nav and footer, not the app tab bar); its "Get started" CTA is tap 1 into Path Choice, so the 3-tap count above is unchanged. --- ## Traceability Coverage matrix. Rows are every job from user-research/docs/jtbd.md (main, related, and emotional/social). Columns are every screen from the Screens section. A checkmark means the screen genuinely participates in closing that job. A gate that only unlocks a job does not count as closing it. Column key: Wel = Welcome / Value Intro, Pth = Activation Path Choice, Bnk = Connect Bank, Add = Add Subscription, Rev = Guided Reveal, Hom = Home / Subscription List, Det = Subscription Detail, Alr = Alerts / Activity, Gde = Cancel Guide, Win = Cancel Win Moment, Shr = Share Snapshot, His = History and Trends, Upg = Upgrade / Tendd Pro, Con = Connections / Accounts, Prv = Data and Privacy, Set = Settings / Profile. | Job | Wel | Pth | Bnk | Add | Rev | Hom | Det | Alr | Gde | Win | Shr | His | Upg | Con | Prv | Set | |-----|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:| | J-MAIN see all charges calmly | | | | | ✓ | ✓ | | | | | | ✓ | | | | | | J1 activate without anxiety | ✓ | ✓ | ✓ | | | | | | | | | | | ✓ | ✓ | | | J2 identify and cancel unused | | | | | | ✓ | ✓ | | ✓ | ✓ | | ✓ | | | | | | J3 understand what a charge is | | | | | ✓ | ✓ | ✓ | | | | | | | | | | | J4 stay ahead of surprises | | | | | | | ✓ | ✓ | | | | | | | | | | J5 track without sharing bank | | ✓ | | ✓ | | | | | | | | | | ✓ | | | | E1 feel competent, not judged | ✓ | | | | ✓ | ✓ | | | | ✓ | | | | | | | | E2 the small win | | | | | | | | | | ✓ | ✓ | | | | | | | E3 feel safe, control data | | | | | | | | | | | | | | ✓ | ✓ | ✓ | | S1 share the discovery | | | | | | | | | | ✓ | ✓ | | | | | | ### Orphan screens (columns with no checkmark) - **Upgrade / Tendd Pro.** No job in jtbd.md closes on this screen. It exists because of business decisions D3 (the free vs paid split) and D4 (pricing), not because a user job needs it. A paywall gates jobs, it does not close them, so it earns no checkmark. - **Resolution: retain, justified by decision, and attach rather than promote.** It is kept because D3 and D4 require a place to convert, but it is implemented as a contextual gate reached from the Pro-gated surfaces (History and Trends, advanced alerts in Alerts / Activity, the full guide in Cancel Guide), not as a global navigation destination. This is already how the Navigation section classifies it. It is the one screen that traces to a decision instead of a job, and that is recorded here openly rather than hidden with a stretched checkmark. No other screen is orphaned. Every other column has at least one checkmark. ### Orphan jobs (rows with no checkmark) None. All ten jobs (J-MAIN, J1 to J5, E1 to E3, S1) are served by at least one screen. The jobs deliberately excluded from MVP (JH1 household, JH2 AI assistant) are not rows here, because they are not MVP jobs. ### Outcome Zero orphan jobs. One orphan screen (Upgrade / Tendd Pro), which is retained by design as a decision-justified gate rather than removed, since removing it would break the business model. Every MVP job has a home, and every screen except the deliberate paywall gate closes at least one job. --- ## Critique A strict pass over the Screens, Navigation, and ia/docs/flows.md across four defect classes. Defects are listed first, most dangerous class first (dead ends, then missing states, then excess depth, then orphans). The proposed fixes are proposals, not applied here: this is structure only, so they become requirements for Prompt 2 (per-screen content and states). All four fixes for dead ends and missing states resolve to states and copy on existing screens, so the 16-screen inventory does not change. ### 1. Dead ends (most dangerous) | Where | What (defect) | Proposed fix (for Prompt 2) | |-------|---------------|------------------------------| | Flow A, Connect Bank | If the user declines both retry and the manual fallback after a connection error, they leave with no list at all. | Make the manual path the always-available floor: the connection-error state keeps a persistent "Add them yourself instead" action, so there is never a version of the error with only a dead exit. | | Flow B, Add Subscription | Ravi abandons manual entry half-done (the manual-entry trap) and leaves with a partial, low-value list. | Save partial progress and re-hook: persist what was added, let the reveal run on a partial list ("here is what you have so far"), and use the return notification (BP4) to pull him back to finish. | | Flow C, Cancel Guide | A blocked external cancellation (retention dark pattern) leaves the user stuck with no in-app next step. | Add a "could not cancel" state to Cancel Guide with a next step (alternative steps, escalation tips, or mark-to-retry) so the person always has a move inside the app. | | Flow D, Subscription Detail | A failed-payment alert whose fix lives at the bank or merchant ends at "informed" with nothing to do. | Add a plain-language next-step card ("update your card with the merchant" or "no action needed, it will retry on DATE") so the alert never terminates without a next step. | ### 2. Missing states (also dangerous) | Where | What (defect) | Proposed fix (for Prompt 2) | |-------|---------------|------------------------------| | Home / Subscription List | Only the populated happy path is defined. Empty (zero or all cancelled), loading (refreshing sync), and sync-error are not yet specified. | Define three states: a calm empty state that invites adding or connecting, a non-alarming loading state, and a sync-error state that still shows the last known list. | | Subscription Detail | No state for when enrichment fails and the merchant name stays cryptic. This directly threatens J3 (understand what a charge is). | Define an "unrecognized charge" state with a plain-language prompt to name or categorize it, so a failed logo or name lookup does not re-create the exact pain J3 exists to remove. | | History and Trends | Only the Pro-locked state is implied. A paid user with too little history yet (under 3 months) has no defined empty state. | Define a "still gathering your history" empty state distinct from the Pro upsell, so early paid users are not shown a blank chart. | | Cancel Guide | No state for a service that has no guide yet. | Define a "no guide for this one yet" state with the generic basic instruction and a request-a-guide action. | | Share Snapshot | Generating the card and a share failure are not specified. | Define a loading state while the card renders and an error state if the share sheet or image fails. | ### 3. Excess depth | Where | What | Proposed fix | |-------|------|--------------| | Flow A, first-session J-MAIN | The main job sits at exactly 3 taps in the first session (Welcome, Path Choice, Connect Bank, Guided Reveal). This is at the ceiling, not over it, so it is not a violation. | Watch item, not a defect: do not add any step to the onboarding chain. Any new interstitial would push the main job to 4 taps and break the limit. Guard this in wireframes. | No frequent job for the primary persona exceeds three taps. Steady-state J-MAIN is 0 to 1 tap (opens on Home). J4 from a notification is about 2 taps. J2 for Claudia is 2 taps to the guide. ### 4. Orphans (cross-checked against the coverage matrix) | Where | What | Proposed fix | |-------|------|--------------| | Upgrade / Tendd Pro | Orphan screen: no job closes on it. It traces to decisions D3 and D4, not to a job. | Retain as a contextual gate over the Pro surfaces (History and Trends, advanced alerts, full cancel guide), not a navigation destination. Already reflected in Navigation and Traceability. | | Jobs | No orphan jobs. Every job in jtbd.md is served by at least one screen. | None needed. | ### Summary of the critique The structure has no orphan jobs and one intentional orphan screen (the paywall gate). The real work it surfaces is at the edges: four dead ends and five missing-state gaps, all of which are state and copy problems on existing screens rather than missing screens. They are handed to Prompt 2 as requirements. The one structural watch item is that first-session depth to the main job is exactly at the three-tap limit and must not grow. --- # Full Information Architecture - Index Prompt 2 (full) built a page-level library: one markdown file per cluster plus global chrome, system pages, and SEO, each rendered to its own HTML page. Those files are the single source of truth for page-level detail. This sitemap stays the index and the skeleton (Entities, Screens, Navigation, Traceability, Critique above). The earlier draft "Global Components", "Page Cards", and "Breakpoint Deltas" sections were migrated into the library and removed from here so there is one source of truth, not two. ## Page-level library | Node | Source | Covers | Rendered | |------|--------|--------|----------| | 0 | [pages/navigation.md](pages/navigation.md) | GC1 App Header, GC2 Tab Bar, persona x plan state matrix, desktop rail delta, no-footer note | [navigation.html](../navigation.html) | | 1 | [pages/onboarding.md](pages/onboarding.md) | Welcome, Path Choice, Connect Bank, Add Subscription, Guided Reveal (screens 1 to 5) | [onboarding.html](../onboarding.html) | | 2 | [pages/core.md](pages/core.md) | Home, Subscription Detail (6 to 7); owns GC3, GC4; FLAG 1 resolution | [core.html](../core.html) | | 3 | [pages/alerts.md](pages/alerts.md) | Alerts / Activity (8); owns GC5 | [alerts.html](../alerts.html) | | 4 | [pages/cancel.md](pages/cancel.md) | Cancel Guide, Cancel Win, Share Snapshot (9 to 11) | [cancel.html](../cancel.html) | | 5 | [pages/pro.md](pages/pro.md) | History and Trends, Upgrade (12 to 13); owns GC7 | [pro.html](../pro.html) | | 6 | [pages/account.md](pages/account.md) | Connections, Data and Privacy, Settings (14 to 16); owns GC6; FLAG 2 resolution | [account.html](../account.html) | | 8 | [pages/seo.md](pages/seo.md) | SEO engine, scoped to the public Welcome screen | [seo.html](../seo.html) | | 9 | [pages/system.md](pages/system.md) | Not found, server error, maintenance, consent [?], toasts | [system.html](../system.html) | Reusable Sections and where each is specified: GC3 Recurring Summary Strip and GC4 Subscription List Item in core.md; GC5 Alert Item in alerts.md; GC6 Data Source and Trust Indicator in account.md; GC7 Pro Gate and Plan Chip in pro.md. GC1 App Header and GC2 Global Tab Bar are global chrome in navigation.md. ## Flags from Prompt 2 (resolved) - **FLAG 1** (the Save tab had no dedicated landing screen): resolved in pages/core.md. The Save tab lands on a save-focus state of Home, not a new screen. - **FLAG 2** (Connect Bank and Add Subscription reused outside onboarding): resolved in pages/account.md. One "Add a source" action re-enters both in-app through a chooser. ## Locked decisions (this layer) - Save tab is a save-focus state of Home; no new screen (core.md). - Add a source re-enters onboarding screens in-app; no duplicates (account.md). - Equal-weight, no-preselect Path Choice; persistent manual fallback on the bank path (onboarding.md). - Guided Reveal stays one screen, three steps, total never alone (onboarding.md). - Cancel is free-first, never walled; could-not-cancel has an in-app next step (cancel.md). - Free is gated by depth and history, never by subscription count or connection count (pro.md, D3). - Upgrade is a context-only gate that shows its origin; never a nav item (pro.md). - Active-voice alert copy; quiet new-dot, no red counts (alerts.md, navigation.md). - One-tap, single-confirm, delete-everything data deletion (account.md). - No footer on app screens; the Welcome landing (1.1) is the one exception, a marketing footer (navigation.md, onboarding.md, updated 2026-07-04). - No full-app maintenance page at MVP; sync issues are per-source (system.md). - Welcome / Value Intro is the public marketing landing / front door in-repo (full-width, own nav + footer); its CTAs enter the chain at Path Choice (onboarding.md, seo.md, updated 2026-07-04). - Desktop: bottom tab bar becomes a left rail with the header folded into its top; Home and Subscription Detail are a two-pane master-detail (navigation.md, core.md, realized in Wireframes 2026-07-04). ## Open items [?] (operational or data, not IA) Gathered from every page-level file's Open section: - Auth model for try-with-no-account vs returning sign in (onboarding.md, account.md, navigation.md). - Minimum subscription count for the Guided Reveal aha (onboarding.md). - How cancel candidates and potential savings are computed (core.md). - Cancellation verification via the next billing cycle (cancel.md). - Cancel-guide content coverage, the no-guide-yet long tail (cancel.md). - Lifetime price, $99 to $139 (pro.md, D4). - Whether data export lives in Data and Privacy or in Pro (account.md, pro.md). - Consent banner, pending the PostHog cookie decision and the deferred EU review (system.md). - Real SEO search volumes and Core Web Vitals targets (seo.md). - Alerts new-dot vs count, and pricing anchor order, pending usability testing (alerts.md, pro.md).