The commercialised fork of Feltfaction Ios Venture Shape. Captured from the Connor call on 2026-07-06 (~70 min) and refined across two brainstorm sessions on 2026-07-07 and 2026-07-09.
"we need a small working product first before expanding because there's no point
if it doesn't work."
If scope ever drifts, that's the sentence to bring it back.
Invite-only friends-of-friends platform for in-person games nights. Host creates a "games room" with a join code; members reserve seats for a date/venue. Tracks scores across games and nights (continuous tournaments, not wipe-at-midnight).
Cards Against Humanity, Monopoly, Pluto Chess) with how-to guides. Players are physically present; the app removes the friction of "pick up the phone, dial six friends, bring a plate."
redemption). Continuous state across sessions within a season; season-end = winner declared, ledger reset.
Two roles only: host and member. No tiers. The host is the room's owner and scorer; members join via the host's invite code (or via a member-issued sub-code in v2 friend-of-friend expansion).
as a member of that room.
Optimistic-lock on redemption; second device gets "code already used."
set by the host (v2 friend-of-friend expansion).
codes expire. (Locked 2026-07-09.)
Season-end = winner declared, season_totals archived, new season opens fresh.
2026-07-09.)
season. New season starts with a clean seat state.
avoid. The deposit is the tool against that.
2026-07-09.) Closed-loop economy — chips that exit via forfeit do not enrich the host, do not seed a prize pool, do not roll forward. Deflationary by accident of design, which is fine: it makes the deposit a genuine cost to a no-show without making anyone else wealthier for it.
(deposit refunded) or invited → confirmed → no_show (deposit forfeited)`.
a total_score plus a scores_breakdown (per game). User-facing: Poker: +240 · CAH: +60 · Monopoly: −30 · Season: +1,840.
end-of-night entry) is the UX risk. Solved by an iPad-first dashboard the host runs at the table — big buttons, per-pack UI, scores added through the night not just at the end.
component. Dashboard renderer takes (room, installed_packs) and composes the UI from pack definitions. A pack is a mini-app inside the app.
iPad. Glance + Live Activity + Watch complication are MVP load-bearing for the member-facing live-score surface (locked 2026-07-09), not optional. Without these, members are blind during sessions. Refresh budget: 1 minute — poker chips live on the table; blackjack scores at end-of-round; CAH rounds are conversational. One minute covers all of it.
scorer role. Members in the same room can view everyone's scores, not just their own. (Locked 2026-07-09.) Spectator view for non-host members — cheap to ship, valuable for social proof, and a load-bearing MVP feature not a nice-to-have.
the iPad scores." If role-sharing is needed in practice, fix in v2; not architectural surgery.
a score, members see "host correcting" indicator in the Live Activity for ~30s, plus a small badge on the affected member's row in the standings. No version history in v1; just "this row changed, recently." See Games Room Scoring Pack Schema § Correction visibility.
| Field | Source | Purpose |
|---|---|---|
| Metadata (name, icon, description) | Author | Discovery + display |
| Scoring UI schema (fields, labels, layout) | Author | The big buttons render from this |
| Score computation (chip math, win conditions) | Author | Live totals |
| How-to guide content | Author | Onboarding for newer players |
For MVP: hand-coded packs (Casino, CAH, Monopoly, Pluto Chess from the original call). The contribution surface (community packs, submission, approval, hosting, versioning) is v2 — catalog exists from day 1, contribution opens later.
Substrate: season_summaries archived at each season end. Cross-room progression is a future opt-in, deferred.
| Option | Status | Notes |
|---|---|---|
| Ads | No | "Apps/ads cheapen the experience" — UX instinct, load-bearing |
| Affiliate links (Amazon for physical boards) | Possible | "If you don't have the board, here's where to buy it" |
| Freemium with ad-free paid tier | Deferred → v2 | What's the paid tier gate? See below |
| Per-pack paid downloads (~$5/pack) | MVP path | Install blank, pull packs from in-app store |
| In-game currency bought for real money | No | Pay-to-win violates chip-game integrity |
| Pub/venue partnership (trivia nights as commercial rooms) | Deferred | Down the track — different market |
| Community-created packs | v2 expansion | Catalog can't come from dev team alone — structural, not preference. Submission/approval/moderation system ships later |
If packs ship free in MVP (dev-curated = free), freemium needs to gate something else. Candidates:
Architecture for tier-gated features exists from MVP (room_settings table supports it). Pick the actual offerings when there are users to validate against.
has enough surface area to write personality copy. Right now there's nothing to be snarky about.
ads question. Bundled deferral: revisit after per-pack store is live and user behaviour around guides is observable.
offerings at v2 when users exist.
off-nights (Tue/Wed/Thu) for foot traffic + drink revenue.
with no real-world exchange is the assumed posture. Worth a real legal read before launch. Connor's verbatim: "You want to be careful you don't step on gaming rules and gambling rules and things… we haven't done any legal assessment or risk assessment on the idea."
free; the shell exists).
the last-opened room. Rooms list lives in a top-bar dropdown reachable from any room view. Create-room affordance is a "+" inside that dropdown, visible to anyone but only functional for users hosting their own room. Account-level settings (display name, email, sign-out) live in a separate account surface, not inside the rooms dropdown — keep the rooms dropdown tight on rooms. Settings live inside each room (pack management, season config, join-code generation, scoring correction history — host only).
plays in two others, the room row needs to make the host role visible without being shouty. Color, icon, or section header — pick at iOS UI design time. Cheap if rooms-list is the v1 architecture; expensive otherwise.
live (window has started, host is scoring), the room row in the dropdown looks different from a dormant one — green dot, "LIVE" badge, or row reorder. This is also the surface where Glance / Live Activity / Watch get their "you have a session in 23 minutes" reminder — the v1 Felt Faction wedge.
channel for the app (user mutes via iOS system settings, not via an in-app per-room mute). Per-room thread-id on the server side is kept purely for DB-side continuity — so the server knows what notifications have already been sent to a given room and doesn't double-ping. The thread is a server audit/dedup concept, not a UX surface. Per-room iOS notification categories and the silent-signoff event-stream architecture (see Games Room System Notifications) are deferred — not v1.
"we knew it was a niche thing to begin with"
Global reach argument + pub-partnership expansion as the scale path. Connor pressed on this — "you wouldn't just pick up the phone and dial half a dozen friends?" Nathan's answer: friction — invites, RSVPs, scoring across nights. Connor agreed.
a project, until there's a real launch.
write for.
protected MVP scope in the call
from the same session
pipeline that processed this call
Published and managed by TARS, an AI co-author built on Nathan's gbrain.