title: User Journeys

aliases: [Journeys, User Flows]

tags: [3netra/wiki, ux, journeys, diagrams]

product: 3netra Meta Android App

source_repository: 3Netra-ai/3n_meta_android

source_ref: origin/prod

source_revision: b48aada8fd5427b5db0c02945569963c4ae543c8

last_verified: 2026-08-06

| Gesture | Action | Mode | Status |

|---|---|---|---|

| Single tap (right temple) | "What do I see?" | Scene | 🟡 callable method; Home invokes it, but no DAT callback registration was found |

| Double tap (right temple) | "Read what's written" | OCR/Read | 🟡 callable method only |

| Hold 2s (right temple) | "Who is in front of me?" / voice command | Person ID / mic | 🟡 callable method starts Android recognition; no DAT callback found |

| Single tap (left temple) | "What is happening here?" | Activity | ⏳ |

| Double tap (left temple) | "How does that person feel?" | Emotion | 🟡 callable method only |

| Voice "Hey Netra" | Wake word activation | — | ⏳ planned (service removed) |

1. Student/disability support coordinator installs app → Login/Signup (email+password, OTP, or Google)

2. Permission rationale screen (BT, location, camera, mic, notifications) with FERPA/COPPA banner

3. Onboarding: Netra welcome → academic level selection (tunes vocabulary for high school/college) → glasses pairing (Meta View app must be paired first; auto-connect scans for "Meta"/"RB Meta" devices)

4. (Optional) companion selection + nickname — see [[05-Avatars-Companions]]

5. Home: "👋 Hey, {name}! Ready to explore your campus?"

1. Gesture or voice triggers `BluetoothService` headless pipeline (no UI needed)

2. When hardware capture succeeds, the app invokes Supabase `chat-vision`; the Android client currently discards its answer and substitutes fixed success text

3. Scene UI speaks through Android TTS. Speaker routing and Sarvam behavior are not device-tested

4. Every result logged → feeds My Story

Home → 🔍 → `scene_description/Scene` → the student taps a mode chip → capture attempt → fixed `"AI response received"` on a successful Edge Function call → Android TTS + attempted session log. First-load auto-capture is commented out. Mode chips let the student select Scene/Read/Emotions/People.

Home exposes a Classroom quick action, but `SceneDescriptionScreen` has no Classroom chip/prompt branch and falls back to generic scene behavior. `BluetoothService` contains a Classroom/PDF branch, yet no verified UI or gesture route reaches it. The PDF helper uses legacy public external-storage writes and is not a reliable modern-Android export flow.

Instructor/disability support coordinator (PIN-gated) registers classmates & faculty: Friends tab → + → Admin PIN → name → CameraX capture → FaceNet embedding → Supabase.

Student: 👤 Who's here? → FaceNet calculates an embedding and fetches registered embeddings from Supabase for cosine comparison (>0.5). A match is read aloud locally; otherwise the code invokes the Edge Function. It is neither local-only nor verified as instant, and successful Edge calls still resolve to fixed success text in this client.

Chat tab → type question and/or attach gallery image or video → `askAssistant` samples/compresses media → Supabase `chat-vision` → fixed success text on client-side success. Trash clears in-memory history. No response-time target has implementation or test evidence.

Home → 📝 Notes → Start Recording (device speech recognizer) → live transcript pane → Summarize Now → `SupabaseAiService.summarizeLecture`; the current client returns fixed success text rather than decoding the model summary.

My Story tab renders the day's sessions as an academic timeline (mode emoji, time, narration). **🔊 Read My Day Out Loud** compiles "Here is your day so far… That's everything I saw today!" and speaks it — designed for review or sharing.

- **Face library management** (PIN-protected): register/remove classmates & professors

- **Settings:** some preferences persist locally; server settings synchronization is attempted. The high-contrast and reduce-motion preferences are explicitly local-only.

- **Dashboard layer** (`SupabaseDashboardRepository`): code exposes settings/audit calls, but no active educator surface or authorization workflow is verified.

- **Planned:** web Educator Portal, Disability Support Portal summaries, weekly academic reports

Intended: hold temple 2s → `EmergencyAlertManager` (`TRIGGER_SOS` intent) → payload `{student_id, GPS, recent_summaries, status:"active"}` → Supabase `emergency_alerts` → emergency contact notified.

---