Getting Started — Onboarding & Parent Gate
Use this page to judge adoption risk. If setup feels heavy or confusing, families abandon the app before communication begins. Meadow is trying to reduce that risk aggressively.
Use this page to judge whether defaults, caregiver setup, and child access are introduced safely without undermining communication, access, or clinical flexibility.
60% of AAC devices are abandoned. The first ten minutes predict the next ten years.
Parents who receive an AAC app are often nervous, overwhelmed, and short on time. The moment a setup screen asks them to choose between a dozen options, half of them close the app and never come back.
Meadow’s onboarding principle is radical simplicity: two screens, four choices, under 60 seconds. Everything else sets itself from the age the parent tells us.
“If the app needs explanation, we’re doing something wrong.” Tier 2–3 children should be able to use Meadow immediately after onboarding without a tutorial. The UI is the explanation. The only child who gets a guided intro is Tier 1 — pre-verbal, non-initiating — because those children need the engagement mode before they know what to do.
Onboarding is not just a UI detail. It is one of the first moments that determines whether a family feels Meadow is approachable enough to keep using between therapy sessions.
Left half: the companion. Right half: the form. Four choices total.
The split layout puts a warm face on the left immediately, so the first thing a parent sees isn’t a form — it’s the character their child will meet. The right side is deliberately minimal: name, age, voice, tier. That’s it.
Hi! I’m Meadow!
Let’s set things up so I can help your little one find their words.
Each voice button plays a 2-second sample on tap: “Hi! I’m Meadow!” so the parent can hear what their child will hear before committing.
The three tiers — plain language, no clinical jargon
The tier cards use language parents recognize from observation, not terminology from a diagnostic report. A suggested starting tier is pre-highlighted based on age, but the parent or SLP can choose any tier.
First Words
- Companion intro on first launch (15 sec)
- Core words fire speech directly, no conversation loop
- Maximum celebration intensity
- Longer expectant pause (30–45 seconds)
Word Combinations
- Straight to Kitchen on first launch
- Simple follow-up suggestions after speech
- Standard celebration intensity
- Medium expectant pause (10–15 seconds)
Sentences
- Straight to Kitchen on first launch
- Full SCS conversation loop active
- Moderate, mature celebration
- Short expectant pause (5–10 seconds)
| Age Selection | Suggested Tier | Pre-Highlighted Card |
|---|---|---|
| 12 – 18 months | First Words | Tier 1 |
| 18 – 30 months | Word Combinations | Tier 2 |
| 30 – 48 months | Sentences | Tier 3 |
If a parent selects a tier 2+ steps away from the age suggestion — for example, a 12-month-old on Tier 3 — a gentle, non-blocking note appears: “Most children at this age are still discovering words — does this sound right?” with two options: Continue and Change. This is advisory, never a hard block. The parent always knows their child better than the default.
Brief, protective, no friction.
Screen 2 is one focused thing: protecting settings from little fingers. The copy is plain and honest. There’s a real skip path that defaults to two-finger hold, so parents who find PIN setup annoying have a frictionless way out.
The two options are equivalent in protection for a toddler — a two-finger hold is genuinely inaccessible to a child who doesn’t know the gesture. The PIN is for parents who want an extra layer.
Tapping “Skip” doesn’t skip security — it selects the two-finger hold method. The app is never unprotected. The parent can upgrade to PIN anytime from settings.
Set a Grown-Up Code
This keeps settings safe from little fingers.
You can change it anytime.
Tier 1 gets 15 seconds of warmth. Everyone else gets out of the way.
The best onboarding is no onboarding. For Tier 2 and 3, the moment “Let’s Go” is tapped, the child lands in the Kitchen scene. No tutorial overlay, no guide rails, no interruption. The companion breathes in the corner, available when ready.
Modeling intro moment (~15 seconds)
- 1Companion waves; a brief caregiver-facing message appears: “When your child wants something to eat, tap ‘eat’ on the screen — that’s modeling!”
- 2One gentle spotlight on a single word in the Kitchen scene
- 3Expectant pause — waits up to 45 seconds for any tap
- 4Celebrates the tap, whatever it lands on
- 5Releases into free exploration — no re-entry to tutorial
Straight into the Kitchen
- 1Kitchen scene appears immediately on “Let’s Go”
- 2Companion is visible in bottom-right corner, breathing gently
- 3No tutorial, no overlay, no instruction
- 4Tap any word — it speaks. That’s the whole explanation.
| Setting | Default Value | How it’s derived |
|---|---|---|
| Word labels (text under pictures) | Off | Pre-literate audience navigates by picture |
| All scenes available | Yes | No gating principle — open access |
| Routine schedule | Age-appropriate | Age selection drives bedtime, nap defaults |
| Expectant pause duration | Longer for younger | Age & tier selection |
| Celebration style | Full (age-scaled) | Age & tier selection. Options: Full (default, bigger for younger) / Subtle (gentle feedback only) / Off (no celebrations). Configurable by SLP or parent to align with therapeutic philosophy. |
| Access method | Standard touch | No hold delay, no scanning mode |
| Speech rate | 0.85× (slightly slow) | Parentese TTS principle |
| Sound effects | On | Engagement & feedback default |
| “I love you” | Always available | Permanent — not configurable by design |
| Default view | Scene | Primary interface; grid is fallback |
| Button size | Standard (60pt) | AAC minimum touch target requirement |
| Select on release | Off | Tap-down fires immediately for responsiveness |
| SCS conversation mode | Full | Sentence templates + follow-up bubbles active. SLP can dial to Light (templates only) or Off (direct speech). D21 |
| Carrier phrases (T2) | On | Fill-in-the-blank scaffolding for Tier 2. Off in T1 (too advanced) and T3 (prevents dependency). D24 |
| ASL signing | On (T1) / Off (T2–T3) | Signing demos shown during Speak With Me and on long-press. Tier-gated, SLP can override. D19 |
| Speak With Me (Companion) | On | Proactive engagement mode. SLP can turn off if child should only respond to self-initiated communication. |
| Feelings tray | On | Avatar feelings tray accessible on tap. SLP can turn off if not a therapy target. |
| Vocabulary per scene | All words shown | SLP or parent can hide individual words per scene to reduce overwhelm or match therapy targets. |
| Routine glow | On | Time-aware scene highlighting. Suggestion only — child can always navigate freely. |
| Age | Suggested Tier | Pause Duration | Routine Defaults | Celebration |
|---|---|---|---|---|
| 12–18 months | Tier 1 | 30–45 seconds | Early bedtime 7pm, 2 naps | Big, simple |
| 18–24 months | Tier 1–2 | 15–30 seconds | Bedtime 7:30pm, 1 nap | Big |
| 24–36 months | Tier 2 | 10–15 seconds | Bedtime 8pm, quiet time | Medium |
| 36–48 months | Tier 2–3 | 5–10 seconds | Bedtime 8:30pm, no nap | Moderate |
Two-layer progressive security. Invisible to the child, obvious to the parent.
The settings gear lives in the top-left corner, visually subtle. A toddler tapping it nothing happens. A parent holding two fingers on it for 3 seconds gets in. The gesture is discoverable in about 10 seconds for any adult who’s looked for it; it’s invisible to a child who hasn’t.
Layer 1 — Two-Finger Hold
The settings gear is small and deliberately low-contrast. One-finger taps: nothing happens, no error. Two-finger hold for 3 seconds: the gear spins slowly as visual feedback, then settings open.
This gesture is outside the motor vocabulary of a toddler. No PIN required, no explanation to the child, no visible lock icon.
Layer 2 — PIN (Optional Upgrade)
After the two-finger hold succeeds, if PIN is enabled, a 4-digit entry screen appears. Large number pad, no keyboard, no autocomplete.
Three failed attempts: 30-second cooldown, then retry. No email recovery (we don’t collect emails). Forgot the code? Reinstalling resets it — documented plainly in the About section.
🔒 Protected — requires gate
- Child profile (name, age, voice, tier)
- All display settings
- Access & motor settings (scanning, hold duration, button size)
- Vocabulary controls (hide/show words, favorites)
- Routine schedule
- Insight & Usage statistics (read-only)
- My Words — custom vocabulary (names, family, favorites) NEW v2
- Parent gate settings (PIN, auto-lock timeout)
- Terms of service, privacy, reset
🟢 Never gated — always accessible
- Tapping any word — it always speaks
- Navigating between scenes and rooms
- Feelings tray — emotional expression
- “I love you” — always available, permanent
- Companion — child discovers and taps freely
- All room and scene navigation
Session unlock: once the gate is passed, settings stay open for the current session (configurable: 5 / 10 / 30 min / never auto-lock). Settings entry: the settings screen slides in from the side — it is not a modal overlaying the scene, so there is no jarring interruption if a child is nearby. Child visibility: there is no visible lock icon, no “settings locked” message — the gear is simply unresponsive to one-finger taps, which is normal enough that a toddler ignores it.
Organized by parent mental model — every feature has a toggle.
Settings are grouped the way a parent thinks about their child, not the way a developer thinks about configuration. Ten groups, each answering a distinct parental question. iOS-native feel — familiar, not surprising. Design principle: every feature the app offers can be turned on, off, or adjusted by a parent or SLP. This gives clinicians full control over what the child sees and hears, matching the device to the therapy plan — not the other way around.
For room scenes, three controls stay independent: interaction tier decides what the child can do, scene density decides how many objects are visible in that scene, and word labels decide whether text is shown with those objects. None of the three should silently change the others.
My Child
- text Name
- picker Age (bubble picker)
- preview Voice (Boy / Girl + audio)
- card Interaction tier (same 3 cards) — controls scene complexity, not object count
My Words
- list My People (child, family, friends, pets)
- list My Things (comfort object, toys, favorites)
- list My Places (school, park, grandma’s)
- action Add wildcard (+) per category
- camera Photo import from camera roll
- text Pronunciation hint (optional)
Vocabulary
- per-scene Scene density: Low / Medium / High / Custom NEW
- per-scene Hide/show words per scene NEW
- toggle Word labels on/off — display support only, independent of tier and density D12
- toggle Favorites marking
- selector Default view (Scene / Grid)
- info “I love you” — always on, not toggleable
Communication
- selector SCS conversation mode: Full / Light / Off D21
- toggle Carrier phrases on/off (T2 only) D24
- toggle Feelings tray on/off NEW
Learning & Engagement
- toggle Speak With Me (Companion) on/off NEW
- toggle ASL signing on/off D19
- selector Celebrations: Full / Subtle / Off D20
- slider Expectant pause duration override NEW
- toggle Routine glow (time-aware highlighting) NEW
Access & Motor
- slider Hold duration (0–1000ms)
- toggle Select on release
- selector Scanning (off / auto / manual)
- selector Button size (60 / 80 / 100pt)
Sound
- slider Volume cap
- slider Speech rate + preview
- toggle Sound effects on/off
Routines
- time Breakfast
- time Lunch & Snack
- time Dinner
- time Bath & Bedtime
Parent Gate
- info Current method indicator
- action Enable / Change PIN
- selector Auto-lock: 5/10/30 min / never
About
- info App version
- view Terms of Service
- view Privacy Policy
- action Reset all settings
Design principle: every feature is a toggle
SLPs customize AAC devices heavily for each child’s therapy plan. A child working on single-word requests doesn’t need the conversation engine. A child with sensory sensitivities needs celebrations off. A child who isn’t ready for signing shouldn’t see signing demos. The app must let clinicians strip it down to exactly what the child needs, then layer features on as the child progresses. Every engine in the architecture supports being enabled or disabled independently — the settings surface exposes that control. Items tagged D## trace to confirmed decisions from Alyse Royba’s SLP review (May 2026). Items tagged NEW are additions identified during M2 engine architecture review.
Multiple child profiles · SLP export / progress reports · Vocabulary pack downloads · Progressive onboarding prompts
Usage statistics that stay on the device. Built for the triad that needs them.
Insurance funding requires utilization evidence during the eval phase. SLPs need usage records for billing documentation. Parents want to see if their child is actually using the device between sessions. Meadow provides all three — without sending a single byte off the iPad.
The Insight & Usage screen
Located behind the parent gate as a sibling to Settings. After unlocking, the parent sees a segmented control: Settings | Insight & Usage. The screen is entirely read-only — nothing here affects the child’s experience. The only interactive element is the time range picker.
Word taps — every time the child selects a word and TTS fires.
Message sequences — the ordered chain of taps within each interaction (e.g., I → want → banana → speak), not just individual word counts. Reconstructs the actual messages the child produced — the format insurance evaluators need to see.
Communication scenarios — each interaction is auto-classified by pragmatic function: express wants/needs, request object, answer yes/no, label object, request help, answer wh-question, use 2–3 word sentence. Classification derives from our existing vocabulary tags and SCS context — no manual entry required.
Scene changes — navigating between routine scenes. Maps to functional use evidence (can the child independently navigate the device).
SCS interactions — follow-up selections in conversation loops (Tier 2–3).
Speak With Me — companion engagement sessions started or completed.
Raw events are rolled up into daily summaries automatically — one row per day with counts and frequency maps. No names, no device IDs, no session tokens.
There is no export mechanism in v1. Data physically cannot leave the device unless the parent screenshots the screen.
No network calls. Zero.
Meadow’s event logger is designed against AbleNet’s insurance data collection format for QuickTalker Freestyle funding requests. The auto-tracked fields — message content, communication scenarios, functional navigation, and scene context — map directly to the data points insurance firms evaluate during the trial period. Three fields that require clinical observation (communication partner, level of cueing, type of cueing) are candidates for a future SLP session log feature where the therapist records who was present and how much support the child needed. The app auto-fills everything else.
SLP session log (communication partner, cueing level/type) · Vocabulary growth charts over time · SCS completion rates · Speak With Me engagement trends · Session duration tracking
What Meadow remembers about your child
This is not a technical section — it’s a plain-English account of what the app stores on-device and why. All data lives locally in SwiftData. No cloud, no accounts, no analytics. Reinstalling the app resets everything.
Child Profile
- Name — shown in greeting
- Age in months — drives defaults
- Voice — boy / girl TTS selection
- Interaction tier — 1, 2, or 3
- Parent gate method — hold or PIN
- PIN hash — nil if skipped
- Auto-lock timeout — minutes
- Onboarding completed — flag
App Settings
- Word labels — on / off (global display toggle)
- Default view — scene or grid
- Hold duration — 0–1000ms
- Select on release — on / off
- Scanning mode — off / auto / manual
- Button size — standard / large / XL
- Volume cap — 0.0–1.0
- Speech rate — default 0.85×
- Sound effects — on / off
Scene Preferences
- Scene ID — kitchen, bathroom, bedroom, etc.
- Scene density — low / medium / high / custom
- Custom priority cutoff — optional advanced override
Routine Schedule
- Breakfast — age-derived default
- Lunch — age-derived default
- Snack — age-derived default
- Dinner — age-derived default
- Bath — age-derived default
- Bedtime — age-derived default
Word Visibility
- Word ID — reference to vocabulary
- Hidden — parent can hide words
- Favorited — parent can pin words
Interaction Event
- Timestamp — when the interaction occurred
- Scene — active scene/routine at time of event
- Tap sequence — ordered word IDs (e.g., I → want → banana)
- Scenario type — auto-classified: request, express want, label, yes/no, wh-answer, help, multi-word
- Event type — word tap / SCS follow-up / scene change / Speak With Me
Daily Summary
- Date — calendar day
- Total interactions — all events
- Word tap count — words spoken
- Scene change count — navigation
- Word frequencies — word → tap count
- Scene frequencies — scene → usage count
- Scenario frequencies — scenario type → count
- Multi-word utterance count — messages with 2+ words
- Unique words used — distinct count
Scene Overflow
- Scene ID — which scene this belongs to
- Word ID — contextually relevant word
- Sort order — display order in grid
All data is stored on-device only using SwiftData. No cloud sync, no analytics, no accounts. There is nothing to breach, leak, or subpoena. If a parent reinstalls the app, everything resets — this is documented clearly in the About section under “Forgot your code?” The profile shape is designed for future multi-profile support without requiring a schema migration.
Clinical questions before implementation
- Are the age-derived defaults appropriate as a starting point for the three tiers?
- Is the Tier 1 companion intro helpful, and are Tiers 2-3 appropriately unobstructed?
- Is the parent gate strong enough without creating caregiver friction?
- Are any critical caregiver controls missing from first-run or immediate settings access?
- Are the Insight & Usage metrics (word frequency, scene usage, days active) useful for SLP documentation and insurance reporting?
Four implications for the M2 build
Age-derived defaults are a first-class feature
The defaults system is not a footnote — it’s the primary mechanism that makes sub-60-second onboarding possible. The engine needs a DefaultsEngine that takes age-in-months and returns a complete AppSettings struct. These defaults surface in settings (so parents see what was chosen for them) and are editable at any time.
Scene-aware grid, not a flat vocabulary dump
When a child opens grid view from the Kitchen, they see Kitchen overflow words first, then core words. From the Bathroom, they see Bathroom overflow words first. The grid is contextual. This requires a scene context signal to flow from the navigation engine into the vocabulary service at render time.
Numbers gap in core vocabulary
Numbers are missing. At 12–48 months, children are actively learning 1–10, their age (“I’m 2!”), and quantities (“more,” “all gone,” “two cookies”). This needs to be added to the M1-002 vocabulary update before M2 ships. Numbers belong in core vocabulary, not scene overflow.
My Words in V1 — unified Apple TTS voice
Custom vocabulary (“My Words”) ships in V1 with ~42 personalizable slots. The child’s communication voice uses Apple premium TTS (AVSpeechSynthesizer) for all words — both bundled vocabulary and custom names. This eliminates the voice-split problem: “Mama wants juice” sounds like one person, not a patchwork of ElevenLabs and Apple voices. ElevenLabs is used for the companion voice only. Custom words integrate into scenes, edges, grid, and the sentence engine as first-class vocabulary.
The Meadow design system
Onboarding is the entry point. These pages describe what the parent and child encounter next.
Interaction Design
The compass frame, core word architecture, scene-aware navigation, and the three-tier behavior system that onboarding configures.
The Companion
The character Tier 1 children meet on first launch. Engagement mode, signing demonstrations, and behavioral rules.
Developmental Level
The clinical grounding behind the three tiers and why the 12–48 month window is the right target for this intervention.
Vocabulary & Symbols
The core word architecture, scene overflow vocabulary, and symbol strategy that the child encounters from their first tap.