Deck 8 Veranda · personal network
Let's Chat
Guest-to-guest encrypted comms on the ship — without becoming a Lattice Chat power user first.
Let's Chat is the Goldilocks comms lounge for guests and creators alike. Creators are guests here too. You do not need a provider API key, a repo workspace, or the Lattice Chat allowlist — board with your email inside Let's Chat, then grow a personal network by inviting friends by email.
Why Let's Chat
- Personal network only — you see people you invite (or who accept your invite). Not a global ship roster.
- No harvesting — the center pipe relays encrypted envelopes for seconds, then they expire. No ad profile.
- Fractal EGS encryption — messages derive keys from the EGS frontal constant (Φ ≈ 1.618) on your device. The server sees ciphertext only.
- Goldilocks environment — consent-first. Fair Exchange. Do Not Disturb when you need quiet. Predators never welcome.
- Standalone or Collaborate — open
/lets-chatdirectly, or hop over from Lattice Collaborate on Deck 8 Veranda.
Compared to other lanes
| Let's Chat | Lattice Chat | Big platforms | |
|---|---|---|---|
| Who | Email board · personal network | BYOK agent seats | Anyone |
| Server keeps messages | No (ephemeral relay) | Collaborate DMs optional pipe | Often yes |
| Encryption | EGS fractal · edge decrypt | BYOK to provider | Varies |
| Consent rail | DND · Fair Exchange | Goldilocks valet | Terms of service |
How to board
- Open Let's Chat and enter your email — Come aboard.
- That generates an in-app approval DM to the Purser (valetpru).
- Valetpru opens the chat and taps Approve on that message.
- Tap + and enter a friend's email — an invite email opens with an intro and link to Let's Chat.
When they come aboard with that email, you land in each other's private network. Fair Exchange stays on deck.
Honesty boundary: EGS fractal encryption is catalog-grade AES-GCM on the edge — design language keyed to Φ, not a post-quantum or clinical claim. Voice and video use WebRTC between peers; the relay carries signaling only. Multi-instance deploys may need a shared ephemeral store later — today the pipe is memory-only with TTL.