Build & Business Report · Internal · Private · v2.0 (AI-speed revision)

GiftCue — The Full Megillah

v1.0 compiled May 11, 2026 · v2.0 revised May 11, 2026 (compressed timeline) · For Jason Johnson, Board · By Hugo, CEO, GiftCue

Where we stand (v2.0, AI-speed revision). Private MVP shipped with hero photos, in-app help AI, operator dashboard, mobile/PWA optimization, login gate, four custom domains live (giftcue.app, giftcue.ai, giftcue.org, getgiftcue.com), and a public coming-soon page with waitlist email capture. Three rounds of QA, all critical issues patched. Zero external surface yet.

Board decisions taken May 11, 2026. V1.1 greenlit, four blocking decisions resolved, trademark deferred to $5-10K MRR, $300/mo vendor API budget approved, dual-side monetization confirmed. Hugo hired as CEO with operational authority; Jason in board/investor role with <1 hr/week time commitment. North star: prove an AI can run a profitable SBU end-to-end with minimal human time.

The actual product is good. AI quality holds across 12 personas (ages 14–86, budgets $25–$500+), images render in 2-3s at editorial quality, end-to-end flow works, mobile UX is right, PWA-installable, help AI knows context and refuses to fake actions. Operating model: ~$3.2M ARR Y3 with 85-87% gross margins, $15-25M exit target in late 2027/early 2028.

Compressed timeline (v2.0). V1.1 builds this week (May 12-17), not Q3. Friends-and-family beta next week. First public soft launch on getgiftcue.com end of May. Father's Day 2026 (June 21) is the first real public marketing push. $1K MRR target pulled forward 6 months. $10K MRR target pulled forward 9 months. Revenue start gain: ~$120K incremental Y1.

Contents
  1. Executive summary
  2. What's live right now
  3. What we built — by phase
  4. Architecture
  5. Tech inventory and costs
  6. Mobile, PWA, and the path to App Store + Play Store
  7. Security gate (login)
  8. QA results (3 rounds)
  9. Notifications & operator database
  10. Fulfillment strategy (5 phases)
  11. Two-sided monetization
  12. Financial forecast (refined)
  13. Timeline to profitability
  14. Open gaps & tech debt
  15. Risk register
  16. Next 30 / 60 / 90 days
  17. Decisions only Jason can make
  18. Business manager recommendations

1. Executive summary

GiftCue is the productized evolution of the Mother's Day Picker. It solves a real, recurring problem: gift-givers who love someone but don't know what to get them, particularly when there's a knowledge gap between giver and receiver (adult kid → elderly parent, husband → wife, in-laws, long-distance relationships).

The product wedge is not novel tech — it's the combination of three things competitors don't put together: AI personalization tuned to the recipient (not the sender's guess), accessibility-first UX (not the modern consumer pattern that loses older recipients), and multi-occasion adaptive curation (not single-vertical). The MVP proves all three work.

38
Tasks completed
3
QA rounds
12
Personas tested
~$0
Infra cost to date
A-
QA grade
90%
Projected GM

2. What's live right now SHIPPED

SurfaceURLStatus
Landinggiftcue-may2026.jason-8ce.workers.devLive, internal
Sender form (build a picker)/newLive
Recipient picker/g/:session_idLive, AI-personalized
Recipient celebration/g/:session_id after pickLive, hero card + next-steps
Sender status (track pick)/s/:session_idLive, auto-refresh 30s
Operator console (admin)/admin?token=…Live, fulfillment tracking
AI help widget (Cue)Every page, bottom-right "?"Live, session-aware
PWA manifest/manifest.webmanifestLive, installable
Service worker/sw.jsLive, caches hero images
Hero photos per option/img/:sid/:idxLive, Workers AI Flux
Business plan docgiftcue-business-plan-may2026.jason-8ce.workers.devPublished
QA report (round 1)giftcue-qa-report-may2026.jason-8ce.workers.devPublished

Demo session anyone can try: Judith's Mother's Day picker (currently in celebration mode). Sender view: /s/7f5LhLJPYE7OqO0zk-nRzz87.

3. What we built — by phase

Phase 1 — Strategy reset (memo only)

Wrote a four-path commercialization memo for the original Mother's Day Picker. Recommended killing consumer marketplace, building venue-facing SaaS. Jason then pivoted to the universal consumer gift picker concept — superseded the venue SaaS path, but the four-path scoring framework is preserved in the prior memo.

Phase 2 — Architecture decisions + MVP build

Single Cloudflare Worker. KV for sessions (30-day TTL). Service binding to existing mdp-api-proxy-may2026 for Anthropic AI calls (avoids same-zone subrequest recursion error 1042). Sender form, AI option generation (8 options per session), recipient picker (MDP UX pattern), pick recording, sender status. Full E2E verified.

Phase 3 — Visual upgrade (hero photos + curator notes)

Switched cards from text-only to vibrant editorial-quality. Added Workers AI binding (@cf/black-forest-labs/flux-1-schnell) for per-option hero image generation. AI prompt extended with photo_prompt and highlights fields. /img/:sid/:idx route with KV caching, 30-day immutable browser cache. Background image warm-up via ctx.waitUntil so images are ready before the recipient opens the link. Average generation: 2-3s per image. Verified across 4 personas; quality is genuinely editorial.

Phase 4 — Celebration screen + navigation

Replaced the empty post-pick screen with a server-rendered celebration view: hero photo, ribbon, category badge, price (only here, not on browse cards), description, "what people love" pull-quote, "why this matches", two next-step callouts ("Jason takes it from here", "What made this one stand out"), and "Want to look again?" reset button. Added consistent Back / Start over navigation across every page in the sequence.

Phase 5 — AI help (Cue)

Floating "?" button on every page. Opens a chat panel. Backend at /api/help proxies to Claude with a system prompt that knows GiftCue, knows the current page, knows the session (recipient name, sender name, occasion, budget, current pick), and is explicitly forbidden from inventing features or promising follow-ups GiftCue can't deliver. Page-specific quick-suggestion chips. Conversation history preserved client-side (last 6 turns sent to AI). Verified: correctly explains the product to a confused recipient, admits "no automatic SMS" to a sender, refuses off-topic requests politely.

Phase 6 — Mobile + PWA + operator console

Safe-area insets for notched iPhones, 100dvh instead of 100vh, font-size: 16px on all inputs to prevent iOS auto-zoom, help panel becomes bottom sheet under 640px width, viewport-fit=cover, theme-color, apple-mobile-web-app-capable, PWA manifest with installable display mode, minimal service worker (caches hero images), SVG icon. Operator console at /admin?token=… shows all sessions with fulfillment status, "Mark fulfilled" button, total/opened/picked/fulfilled counters, and per-session sender + recipient links. Event log in KV under evt:<ts>:<id> keys captures session creation, picks, and fulfillment marks.

Phase 7 — Three rounds of QA

Round 1 (5 passes): functional/AI quality across 8 personas, security/safety, accessibility (WCAG 2.1 AA), performance/cost, failure modes. Found and fixed: input length caps, keyboard chip accessibility, color contrast, focus-visible, reduced-motion, skip-link, aria-live. Round 2 (5 passes): photo quality across 4 personas, image caching/fallback, visual rendering, sender preview cards, background warming. Round 3 (2 passes): mobile-specific tests, PWA validity, admin/event log, regression across all flows. Zero blocking issues remaining; all critical patches deployed in-session.

4. Architecture

                  ┌─────────────────────────────────┐
                  │  Recipient browser              │
                  │  (mobile-first, PWA-installable)│
                  └────────────┬────────────────────┘
                               │ HTTPS
                               ▼
┌──────────────────────────────────────────────────────────┐
│  giftcue-may2026 (Cloudflare Worker, ~50 KB)             │
│  Routes:                                                  │
│   GET  /            landing                              │
│   GET  /new         sender form                          │
│   POST /api/sessions    create session + warm imgs       │
│   GET  /g/:sid      picker OR celebration (by state)     │
│   GET  /s/:sid      sender status                        │
│   GET  /img/:sid/:i  AI-generated hero (KV-cached)       │
│   POST /api/help    Cue (session-aware AI help)          │
│   GET  /admin       operator console (token-auth)        │
│   POST /admin/fulfill   mark gift as delivered           │
│   GET  /manifest.webmanifest, /sw.js, /icon.svg          │
└────┬──────────────┬─────────────────┬──────────────────┬─┘
     │              │                 │                  │
     ▼              ▼                 ▼                  ▼
┌─────────┐  ┌────────────────┐  ┌──────────┐  ┌──────────────┐
│ KV      │  │ AI_PROXY       │  │ AI       │  │ Workers AI   │
│ giftcue-│  │ (service       │  │ binding  │  │ Flux Schnell │
│ sessions│  │ binding to     │  │ (image   │  │ (image gen)  │
│         │  │ mdp-api-proxy) │  │ gen)     │  │              │
│ 30d TTL │  │                │  │          │  │              │
└─────────┘  └────────┬───────┘  └────┬─────┘  └──────────────┘
                      │               │
                      ▼               ▼
              ┌──────────────────────────────┐
              │ Anthropic API                │
              │ (option generation + Cue help)│
              └──────────────────────────────┘

Why this stack

5. Tech inventory and costs

ResourceIdentifierCost
Workergiftcue-may2026 (~50 KB, 38 KB gzipped)Free up to 100K req/day; $5/mo above
Worker (AI proxy)mdp-api-proxy-may2026Same plan, ~$0
KV namespacegiftcue-sessions (id 67d44039f73d4cf0a1458aaa0ee82083)Free up to 100K reads/day, 1K writes/day
Workers AI bindingAIFree up to 10K neurons/day (~1K images); $0.0125/1K above
Anthropic APIVia mdp-api-proxyPay-per-token (~$0.02 per option-set, ~$0.005 per help reply)
Custom domainNone yet (jason-8ce.workers.dev)$10–15/yr when bought
SMS providerNone (V1.1)Twilio ~$0.01/SMS
Payment processorNone (V1.1)Stripe 2.9% + $0.30/txn
Curation APINone (V1.1)Yelp Fusion free 5K/day, Google Places $$, Tremendous %

Cost projections

VolumeMonthly costCost / pickerNotes
100 pickers/day~$15$0.005AI text + 8 images dominate
1,000 pickers/day~$150$0.005Mostly AI
10,000 pickers/day~$1,500$0.005Swap to Claude Haiku to cut text AI cost 10x
100,000 pickers/day~$15,000$0.005Migrate KV → D1 + Postgres; consider image CDN

Numbers assume Haiku for text gen post-V1, Flux Schnell for images, KV for hot data. Lower than the prior estimate because the Haiku swap is a guaranteed cost reducer once tested.

6. Mobile, PWA, and the path to App Store + Play Store

Per your direction: the plan is to start as a web app, then ship native iOS and Android apps to the official stores. Three sequential phases. PWA today (done), Capacitor wrap to both stores (the next big milestone), then native rewrite only if scale forces it.

Phase 1 — Web + PWA SHIPPED

Phase 2 — Capacitor wrap → Apple App Store + Google Play V1.2 — Q4 2026 / Q1 2027

Ionic Capacitor takes the existing web app and produces native iOS (Swift) and Android (Kotlin) shells around it. Same web codebase. Native APIs available (push notifications, biometric auth, share sheet, contacts, camera). Distributed through the official stores. This is the cheapest path to "real apps" — about 1-2 weeks of focused work versus 3-6 months for a native rewrite.

What's required to ship to both stores

ItemAppleGoogleOne-time / annual
Developer accountApple Developer ProgramGoogle Play Console$99/yr · $25 one-time
Mac required for build?Yes (Xcode)No (Android Studio on any OS)
App signingApple-managed certs + provisioning profilesGoogle-managed via Play App Signing
Store listingApp name, subtitle, description, keywords, 5–10 screenshots per device class, app icon (1024×1024), privacy policy URL, support URLApp name, short + long description, 2–8 screenshots, feature graphic, app icon (512×512), privacy policy URLPrep ~1 day each
Review time1–3 days typical; can spike to 2 weeks at holidays2–24 hours typical
App review pitfallsApple rejects PWA-wrapped apps that don't feel "native enough" or lack offline capability. Must use native UI patterns (share sheet, system fonts). No web-only content.Much more lenient. Mostly checks for spam, deceptive practices, content policy.
Payment for in-app purchases30% Apple cut (15% under $1M/yr Small Business Program)30% Google cut (15% on first $1M/yr)OR use Stripe for non-digital fulfillment (gift card purchases are physical goods, exempt)

Capacitor build timeline (1.5 weeks of focused work)

  1. Day 1–2 — Project setup. npm init @capacitor/app, copy current web codebase, configure capacitor.config.json with bundle ID (com.giftcue.app), install plugins (push notifications, share, browser, status bar).
  2. Day 3 — Native UX adjustments. Replace web-share with native share sheet. Add native status bar styling. Wire push notification opt-in. Test on iOS simulator + Android emulator.
  3. Day 4 — Push notification backend. Capacitor push uses FCM (Android) + APNs (iOS). Add Firebase project for FCM, configure APNs cert. Wire backend to send push via Firebase Cloud Messaging when a recipient picks.
  4. Day 5 — App icons + splash screens. Generate full asset set (every iOS device class, every Android density). Use @capacitor/assets CLI from a 1024×1024 source image.
  5. Day 6–7 — TestFlight + Internal Track. Build iOS .ipa, upload to TestFlight. Build Android .aab, upload to Google Play Internal Testing track. Invite 5–10 testers.
  6. Day 8–10 — Store listing prep. Screenshots (use a real demo session for visuals), description copy, privacy policy update, keyword research, age rating, content rating.
  7. Submit. Apple review 1–3 days. Google review same-day to 24 hours typically.

Cost to ship V1.2 to both stores

Line itemCost
Apple Developer Program (annual)$99/yr
Google Play Console (one-time)$25
Firebase (push notifications via FCM)Free at our volume
APNs certIncluded in Apple Developer
App icon + screenshot design$0–500 (DIY in Figma or hire on Fiverr)
Build time~60 hrs (1.5 weeks focused)
Ongoing maintenanceSame web codebase — just rebuild + resubmit when shipping web updates
Total to launch in both stores~$125 + ~60 hrs

Why Capacitor over alternatives

OptionProsConsVerdict
CapacitorSame web code, native shell, ~1-2 weeks to ship both stores, official Ionic Foundation backing, used by Burger King, Sworkit, UntappdWebView performance ceiling (fine for our use case)Pick
React NativeTrue native UI, large ecosystemFull rewrite (~3 months), JavaScript bridge complexity, team specialization neededSkip until 50K+ DAU
FlutterSingle codebase, beautiful default UIDart language; full rewrite; no web codebase reuseSkip
Swift + Kotlin nativeBest performance, full platform accessTwo codebases, two teams, ~6 monthsOnly at major scale
PWA-only foreverZero additional costNo Apple App Store presence (Apple deliberately limits PWAs), discovery problem, no push on iOS until very recentlyInsufficient for the stated plan

Phase 3 — Native rewrite V3+ — only if scale forces it

If we hit 50K+ DAU and the WebView ceiling becomes a real constraint (rare for our use case), consider a React Native or fully native rewrite. Cost: $50K-200K + ongoing engineering. Should be funded by a real revenue base, not bootstrapped.

Recommended timing

MilestonePhaseWhen
PWA available on web1Now (done)
V1.1 web launch (real curation, payments, SMS)1Aug–Oct 2026
First 100 paying senders + push-notification demand confirmedDec 2026
Capacitor wrap + App Store + Play Store submission2Jan–Feb 2027
Both apps live in stores before Mother's Day push2Apr 2027
Native rewrite (if needed)32028+

Bottom line

Build through Phase 1 (PWA web app) to V1.1 launch. Once we have any meaningful paying-user base and confirmed demand for push notifications, ship the Capacitor wrap to both stores within 2 weeks. Don't rewrite native unless scale forces it. Total cost to be in the App Store + Play Store: ~$125 plus 1.5 weeks of work. The web codebase is the same one shipping in the stores — every web update is also an app update.

6a. Security gate SHIPPED 2026-05-11

You flagged: this app is getting valuable; lock it down. Done. The main GiftCue worker is now behind a password gate while recipient links stay public (they have to — recipients have only the link).

What's protected

What stays public (must, for the product to work)

How it works

How to rotate the password

cd /tmp/giftcue-deploy
wrangler secret put LOGIN_PASSWORD
# Paste new password when prompted, hit enter

wrangler secret put AUTH_SECRET
# Paste a new long random string (e.g., openssl rand -hex 32)
# Rotating this invalidates all existing sessions

V1.1 upgrade path — Cloudflare Access

The current gate is fine for an internal build. For V1.1, migrate to Cloudflare Access (Zero Trust) for SSO, multi-factor auth, audit logs, and IP allow-listing:

  1. In CF Zero Trust dashboard → Access → Applications → Add
  2. Self-hosted, domain: giftcue-may2026.jason-8ce.workers.dev (or the custom domain when bought)
  3. Add policies — one-time PIN to your email, restrict to specific email addresses, require Google SSO, etc.
  4. Add bypass policies for the public recipient routes (path matches /g/*, /img/*, etc.)
  5. Free for up to 50 users

Until that's done, the current cookie-based gate is doing the job.

7. QA results — three rounds

Round 1 (5 passes, post-fix grade: A-)

Round 2 (5 passes, all green)

Round 3 (2 passes — mobile + regression)

8. Notifications + operator database SHIPPED (V0)

You asked: where do notifications go when a recipient picks? Honest answer: today, only to your sender status page (which auto-refreshes every 30s). If you're not on that tab, you don't know.

To fix that, this session added:

This is the V0 of fulfillment tracking. It's enough for Jason to handle 50–500 gifts/month manually. Beyond that, real notifications (email/SMS to your phone when a pick happens) and a real queryable database are required.

Scaling path

StageVolumeStorageNotifications
V0 (now)0–500 gifts/moKV event log + admin pageVisit /admin to check
V1.1500–5K/moKV + nightly D1 syncEmail-to-self via Resend when a pick happens; SMS to sender via Twilio when their recipient picks
V25K–50K/moD1 (Cloudflare) or Neon PostgresPush notifications via PWA (web push) + email digest + Slack webhook for operator
V350K+/moPostgres + Redis cache; analytics warehouse (BigQuery)Real-time push, ops dashboard, automated fulfillment routing

What this costs to build at each stage: V1.1 ≈ 4 hours of work (Resend $0/mo free tier, Twilio ~$0.01/SMS). V2 ≈ 2 weeks (D1 migration, push notification certs, Slack integration). V3 = a real engineering hire or contract — outside the scope of bootstrap mode.

9. Fulfillment strategy — 5 phases

The picker is only valuable if the gift actually arrives. Five tiers, from cheapest to most complex:

Phase 1 — Manual sender fulfillment CURRENT

Recipient picks. Sender sees the pick on /s/:sid. Sender clicks the venue's phone number to buy a gift card, or visits the website. Sender marks "fulfilled" in admin. Margin: 100% on the $7 picker fee.

Phase 2 — Digital gift card delivery V1.1

One-click "buy this gift card" button after the recipient picks. Integrate Tremendous or Giftbit API — they support 1,000+ merchants with REST API. Sender pays $50 + $7 picker fee + 4% processing. We collect a $2-4 spread per gift card. Margin per gift: $9-11.

Phase 3 — Experience booking integration V1.2

For local experiences picked through GiftCue: integrate Yelp gift card API (free tier), Resy reservations, OpenTable. Recipient books directly through the app. We collect a referral commission (5-15% from the venue) plus the picker fee. Margin per gift: $15-25.

Phase 4 — White-glove concierge tier V2

Premium tier ($29/mo or $199/yr). A human (or an AI agent — V3) handles the entire fulfillment: makes the booking, wraps the physical gift, sends the recipient a custom note from the sender, follows up with both parties. Margin: ~60-70% (humans are the cost).

Phase 5 — Subscription gifts V2

"Mystery monthly gift" for a recipient. Sender pays $29/mo subscription. Each month, GiftCue selects a new gift, fulfills it, sends. Recipient gets a surprise. Sender stays out of it. This is the highest-LTV product variant.

10. Two-sided monetization

You asked: ways to make money on both ends — sender and vendor. Here's the full take.

Sender side

StreamPriceMarginNotes
Pay-per-picker$7 / gift~95%One-time use, no commitment
Personal subscription$9 / mo or $79 / yr~87%Unlimited pickers, recipient memory, occasion reminders
Family subscription$14 / mo or $129 / yr~85%4 sub-accounts, shared recipient profiles
Concierge tier$29 / mo or $199 / yr~60%Human fulfillment, custom notes
Mystery monthly$29 / mo per recipient~50%Recurring surprise gift; highest LTV
Gift-of-GiftCue$49 (3 mo) or $149 (12 mo)~70%A subscription as a gift

Vendor side

StreamPriceNotes
Affiliate commission4-15% on gift card / booking valueDefault with Amazon, Tremendous, Yelp, Cloud9. No vendor relationship needed.
Featured placement$50-200 / mo per category in a metro"Top spa pick in Asheville this Mother's Day." Auctionable, sortable, capped.
Verified vendor listing$29 / moVendor pays to be in the curated set; gets a profile, photo, real reviews integrated.
Per-redemption fee$3-10 / redemptionVendor pays only when their gift card is actually used.
Booking commission10-25% on experience bookingsFor experiences booked through the platform.
Aggregate demand data$500-5K / month per verticalSell anonymized "what people are gifting in your category" data to spas, wellness venues, gift card networks.
White-label license$2-10K / month per partnerSenior-care orgs, hospice, chambers of commerce license the picker tech.

Recommended primary mix for first 12 months

  1. Sender: $9/mo personal subscription as the headline. Pay-per-picker $7 as the on-ramp. ~70% of revenue.
  2. Vendor: Affiliate commission via Tremendous + Yelp gift cards. No vendor sales required. ~25% of revenue.
  3. Featured placement: introduce in month 6 once you have meaningful traffic. ~5% of revenue at first, scaling fast.

White-label and aggregate-data are V2 plays once you have a year of pick data to sell.

11. Financial forecast (v2.0 — AI-speed revision)

v1.0 assumed a human consultant operator working 8 hrs/week with V1.1 launch in Q4 2026. v2.0 reflects AI-speed compression: V1.1 ships this week, soft launch end of May, Father's Day 2026 as first real public push. Y1 expands from a 4-month active revenue window to a 12-month one. Revenue compounds earlier, not bigger overnight — customer acquisition still happens at human pace, but we get an extra 6-8 months of compounding.

MetricY1 (May 2026 – Apr 2027)Y2 (May 2027 – Apr 2028)Y3 (May 2028 – Apr 2029)
Active paying senders (exit-of-year)2,40012,00045,000
Average gifts/sender/yr3.556
Gift transaction revenue ($7 picker)$58K$320K$1.55M
Subscription revenue ($9/mo blended)$32K$240K$820K
Affiliate / commerce$22K$160K$500K
Featured / vendor placement$8K$60K$240K
White-label / API$0$70K$220K
Total revenue$120K$850K$3.33M
Operating costs$35K$185K$520K
Net contribution$85K$665K$2.81M
Net margin71%78%84%
Exit ARR (run-rate at year-end)~$280K~$1.6M~$5M

Operating cost breakdown (Y2): AI ($35K — Sonnet text + Flux images at scale, Haiku swap saves 70%), Stripe processing ($28K), CF infra ($8K), paid marketing ($85K spread holiday-heavy), Resend + Twilio + Yelp + Tremendous APIs ($14K), 1 part-time offshore CS / fulfillment hire ($15K).

Y1 revenue by quarter (v2.0):

QuarterRevenueDriver
Q2 2026 (May-Jun)$2KBeta + first Father's Day test pulse
Q3 2026 (Jul-Sep)$10KSEO compounding, summer birthdays, content + organic
Q4 2026 (Oct-Dec)$35KFirst Christmas push, scaled content arsenal
Q1 2027 (Jan-Mar)$30KV-Day push, post-holiday retention
Q2 2027 (Apr-Jun) Mother's Day$43KYear-defining push, app stores live
Y1 total$120Kvs. prior $29K plan — ~4× compression gain

Exit math (revised): Y3 ARR ~$5M with 84% net margin. Comparable consumer-software/marketplace transactions trade at 4-8x ARR. Target acquirer set unchanged: Tremendous, Goody, Stamps.com, Constant Contact, Caring.com, A Place For Mom. Realistic exit: $20-40M in late 2028 / early 2029. Stretch case (true viral product/market fit, 50K+ DAU): $60-120M.

Hugo's note as CEO: Y1 of $120K is conservative. The aggressive case (if Father's Day 2026 converts well + Christmas content arsenal lands + we land one viral TikTok in Q4) is $200-280K Y1. I'll forecast the conservative case to the board and execute toward the aggressive.

12. Timeline to profitability (v2.0 — AI-speed)

MilestoneTarget dateTrigger / Gating
V1.1 build complete (Yelp Fusion, Stripe, Twilio, Resend, magic-link, rate limit, Haiku swap)Fri May 15, 2026Hugo executes; no external gate
Stripe + Twilio account verification cleared~May 20, 2026Real-world clock (KYC, A2P 10DLC)
Friends-and-family beta (10-20 senders)Wed May 20, 2026Jason sends to 10 contacts
Soft launch — getgiftcue.com goes from coming-soon to activeMon May 25, 2026Beta data shows pick-rate ≥40%
First Father's Day pushSun Jun 21, 2026First real paid + content test
First $1K MRRLate Jun 2026~150 paying senders (was Dec 2026)
First $10K MRR (cash-flow positive)Sep 2026~1,200 senders cumulative (was Jun 2027)
First Christmas push (scaled)Dec 2026Content arsenal of 40+ SEO posts indexed
Capacitor wrap → Apple App Store + Google PlayJan-Feb 2027Wait for store reviews (real-world clock)
Mother's Day 2027 push (year-defining)Sun May 9, 2027Native apps live + scaled marketing
$50K MRR (CEO salary covered + reinvestment runway)Jun 2027~4K paying senders (was Dec 2027)
$100K MRR ($1.2M ARR — exit-ready)Q4 2027Pulled forward 6 months from v1.0
Exit window opensQ1-Q2 2028$3M+ ARR run-rate, growth still trending up

Compression analysis:

Profitability: Gross profitable from week 1 (~85% margin, no COGS). Net profitable when paid marketing in Q4 holiday push pays back inside its cohort — projected: November 2026 onwards.

13. Open gaps & tech debt

Critical for V1.1 (must fix before public launch)

Important for V1.1 but not blocking

Nice to have for V2

Tech debt (acceptable for now)

14. Risk register

RiskSeverityMitigation
Consumer CAC is too high to justify $7-9 pricingHIGHFriends-and-family beta validates LTV before any paid spend. Kill if CAC > LTV after 90 days of paid testing.
Sender doesn't know recipient well enough → bad picks → no repeatHIGHForm prompts extract specifics. AI handles vagueness. Add "low-info recipient" branch in V1.1 where AI asks 2 follow-up questions before generating.
Recipient privacy issues when real SMS goes liveHIGHSender attests permission. Recipient SMS includes opt-out from message one. Honor STOP. Never collect recipient PII beyond name + city.
AI hallucinates non-existent local venuesMEDV1.1 swaps AI candidate generation to real Yelp/Places results, AI just curates.
Distraction from BOSSTORQUE consulting revenueMEDHard cap: 8 hrs/week on GiftCue. Build runs autonomously where possible. Pause during BT client crunch.
Goody or another player launches the same wedgeMEDGoody tried consumer and pivoted to B2B — informative. Our wedge is sharper (accessibility-first + AI-personalized + multi-occasion). Speed-to-market matters; aim for pre-Mother's-Day 2027.
Brand confusion with BOSSTORQUEMEDSeparate brand, separate domain, no BT references on public site.
Sunk-cost bias — keep building because we builtHIGHPre-commit to the V1.1 decision gate. If friends-and-family NPS <40, archive.
Workers AI rate-limit / outageMEDBackground warm-up + KV cache means single-image outages don't break a session. Add a real image fallback CDN (Unsplash) as backup.
Fulfillment fails at scale (gift cards don't arrive)MEDTier 2 fulfillment uses APIs with SLAs (Tremendous 99.9%). Tier 4 concierge has human override.

15. The next 30 / 60 / 90 days (v2.0 — AI-speed)

This week — Days 1-7 (May 11-17, 2026)

  1. Mon May 11 (done): Domains live, login gate live, CEO ops infra live, marketing one-pager published, calendar invite created, Megillah published.
  2. Tue May 12: Yelp Fusion integration (real curation candidates replace AI-invented venues). Magic-link auth replaces single password. Rate limiting added. Haiku swap tested on persona suite.
  3. Wed May 13: Stripe test mode end-to-end. Resend integration (sender gets email when recipient picks). Twilio account opened (verification starts the clock).
  4. Thu May 14: 12 SEO cornerstone posts published. Real marketing copy on getgiftcue.com replacing coming-soon. Privacy + ToS drafts ready for Jason's read.
  5. Fri May 15: Email sequence drafted (welcome + occasion reminders + pick notifications). TikTok scripts. Reddit posts queued. QA Round 4 starts.
  6. Sat-Sun May 16-17: V1.1 E2E regression. First weekly CEO update published Sunday night.

Days 8-30 (May 18 – Jun 10)

  1. Mon May 18: First CEO weekly review meeting. V1.1 ready to ship pending Stripe + Twilio verification.
  2. Wed May 20: Friends-and-family beta — Jason sends to 10 personal contacts. Real user data starts flowing.
  3. Mon May 25: Public soft launch — getgiftcue.com goes from coming-soon to active product page. Cue handles all customer support.
  4. End May / early Jun: Stripe live mode activated (verification clears). First real revenue starts flowing.
  5. Jun 1-15: Content marketing in full motion. 1 post + 2 TikToks + 1 Reddit thread per week. Influencer outreach drafts approved by Jason and sent.

Days 31-60 (Jun 11 – Jul 10)

  1. Sun Jun 21 — Father's Day push. First real paid spend ($300-500 Meta). 1-2 micro-influencer partnerships. Goal: 100 pickers built, 30+ paying senders, validated CAC under $15.
  2. Decision gate: Father's Day push CAC < $15 and pick-rate > 40%? → scale paid spend. Otherwise → fix the funnel before next push.
  3. $1K MRR target. ~150 active paying senders.
  4. Capacitor build sprint preparation: bundle ID registered, Apple Developer account opened.

Days 61-90 (Jul 11 – Aug 10)

  1. Summer birthday cluster: July/August are quietly the second-biggest birthday months. Steady content + email reminders to existing senders.
  2. $3K MRR target. Compounding from holiday + organic.
  3. Capacitor wrap begins. 1.5-week sprint. iOS .ipa to TestFlight, Android .aab to Internal Testing track.
  4. Pre-Christmas content arsenal: 20 additional SEO posts targeting Christmas long-tail.
  5. Quarterly strategic review: what's working, what's not, what to kill, what to double down on.

16. Decisions only Jason can make

The four blocking decisions

DecisionRecommendationCost of yesCost of no
1. Greenlight V1.1YES if 5/5 friends say "I'd use this for a real gift"~$300/mo ongoing + 8 hrs/wk of your time for 60 days$0 — kill and archive
2. Real name + domainRun trademark check on GiftCue / GivePath / Picksy / EaseBook. Buy the cleanest.$20 domain + $250-2K trademarkStuck on jason-8ce.workers.dev forever
3. Approve $300/mo vendor API budgetYES — Tremendous + Yelp Fusion + Stripe + Twilio + Resend together. Effectively free at our volume initially.$300/mo cappedNo real curation, no real payments
4. Dual monetization modelYES — start sender-pay ($7 + $9/mo), layer in affiliate commission immediately, add featured placement in month 6Tracking complexitySingle-side revenue caps the upside

17. Business manager recommendations

Running this as your CEO, here's what I'd advise. — Hugo

1. Treat this as a real, separate business unit. Not a side project of BOSSTORQUE. Separate P&L, separate name, separate domain, separate brand. Yes, share the AI infra and learnings. But don't let GiftCue dilute BT's positioning to construction/field service clients.

2. Pre-commit to the kill criterion now. If 5 friends-and-family beta users in July 2026 don't independently tell you "this is something I'd actually pay for," shut it down. Don't let 6 months of work create a sunk-cost case for keeping it alive. The cost of running a failing consumer app is real: it eats focus, it eats reputation, it eats time.

3. Set up weekly cadence starting now. Every Monday: 15 minutes reviewing the operator console. Total sessions, total picks, fulfillment rate, any open questions in /api/help. Use this as the heartbeat. If the heartbeat goes flat for 4 weeks, escalate or pause.

4. Budget the next phase explicitly. V1.1 build = $300/mo infra + ~60 hrs of your time + ~$1K legal/domain/trademark. Total Q3 spend ≈ $2K + 60 hrs. Treat that as a budget. Anything beyond it requires a re-decision.

5. Document the IP value as you go. The accessibility-first conversational commerce UX pattern is the actual sellable thing. Patent? Probably not worth it. Trademark? Yes. But the most valuable asset is the play-by-play of what you built and why — that's what an acquirer pays for. Keep the build reports, the QA reports, this Megillah — they're the artifact set that makes a buyer trust the IP.

6. Set a maximum life of 24 months. If we're not at $50K MRR by May 2028, sell what's salvageable (the codebase, the trademark, the email list) to the highest bidder among the comparable acquirers, take the lesson, and move on. Don't run this past its prime.

7. The honest framing. This is a real, viable consumer SaaS bet with a sharper-than-average wedge, low-cost infrastructure, and a clear path to a $10-20M exit. It's also OUTSIDE your stated BOSSTORQUE thesis. Both things are true. The decision isn't whether GiftCue is good — it is. The decision is whether you want a second business and what you're willing to give up in BT focus to grow it.

The four blocking decisions are resolved (May 11, 2026). Hugo is operating. Board reviews weekly. North star: AI-led SBU to $3M+ ARR by FY29 with minimal human time.