Each row writes into the same signalsStore via /api/event-intelligence/signals, with a distinct attributionSource so dashboards can trace which surface captured what. SDK references are linked on each row; we cite Meta, Apple, and Google docs directly so the path is auditable.
Pinch + intent preset (5 intents per world). Compass heading + GPS via paired phone. Sideloadable today via Meta AI app → Developer Mode.
POST /api/event-intelligence/signals · attributionSource: rayban-pinch
Native iOS/Android companion via the Meta Wearables Device Access Toolkit. MWDATCamera for photos, HFP for 8 kHz mono mic. 2-3 days to ship to TestFlight.
POST /api/event-intelligence/signals · attributionSource: rayban-dat-camera | rayban-dat-mic
SDK announced at I/O 2025, general availability end of 2026. Wired to our roadmap; no code path today.
—
Open SDK; ~1 week to wire a custom AugmentOS app that posts to our signals API. Small device base (<10K).
POST /api/event-intelligence/signals · attributionSource: augmentos
Requires a native visionOS app target. Not worth pre-launch given device base; tracked for the v2 worlds path.
—
'Hey Siri, save that to Trovii' → custom AppIntent in a tiny iOS app. Surfaces in Siri, Spotlight, Action Button, Apple Intelligence summaries. TestFlight in 2-3 days.
POST /api/event-intelligence/signals · attributionSource: ios-app-intent
Custom com.trovii.SAVE_SIGNAL capability declared in shortcuts.xml. Routes Gemini contextual chips + 'Hey Google, save to Trovii'. Internal track in 2-3 days.
POST /api/event-intelligence/signals · attributionSource: android-app-action
User builds a Shortcut that calls our URL scheme. Works today on any iPhone without an app install (Shortcuts ships built-in). ~2 hours to author.
trovii://capture?text=…&scenario=… → POST signals
Hold-to-speak in any world page (/basel/my-week, /unrivaled/my-world, /cannes/passport, /swissnex/invisible-infrastructure). MediaRecorder + Whisper transcription.
POST /api/event-intelligence/signals · attributionSource: voice
File input with `capture=environment`. Routes the photo through the signal envelope; downstream CV is wired per-world (gallery interest, place node detection).
POST /api/event-intelligence/signals · attributionSource: camera
Per-world intent presets (5 per world). One tap = one signal. Zero permissions required — works on every browser today.
POST /api/event-intelligence/signals · attributionSource: tap
Any place node or activation surface has a QR; iOS Camera auto-decodes → deep-link into the right world with `?ref=` so attribution flows.
GET → in-app deep-link → tap capture
Operator pass with embedded NFC. Tap-to-add posts a synthetic signal (place, time, scenario). 1-2 days per operator to issue + deploy.
POST /api/event-intelligence/signals · attributionSource: nfc-tile
User forwards a gallery email / hotel confirmation to capture@trovii.com. Inbound parser extracts entities and POSTs the signal. ~3 days to wire SES.
Email inbound → POST signals · attributionSource: email-forward
Existing /api/telegram bot relays user voice notes + text into the signalsStore via the same envelope. Live for the inner-circle group.
Bot webhook → POST signals · attributionSource: telegram-bot
Host apps/trovii-next/ray-ban-app/ as a static site at glasses.trovii.com. In Meta AI app: Devices → Display Glasses → App connections → Web apps → Add.
Enable Developer Mode: Settings → App Info → tap App version 5×.
Drop Swift sources from apps/trovii-next/ios-intent-shell/ into a fresh iOS 16.4+ app. Enable Siri capability. Archive → TestFlight.
Phrase: “Hey Siri, save that to Trovii.”
Drop XML + Kotlin sources from apps/trovii-next/android-action-shell/ into a fresh Empty Activity project. Internal-track in Play Console.
Phrase: “Hey Google, save that to Trovii.”