Portfolio/Superstar PeptidesMonday, June 29 at 8:10 AM EST
⌘KNotifications
Cy
SP
Engagement· started 2026-04-30

Superstar PeptidesACTIVE

Tavia (no contact info — handled directly through Cynthia) · Last touched 2mo ago
North starBuild both a B2B and B2C peptide brand and ride the wave of the future of medicine. Two sides, one platform.
OverviewChannels6Activity24Notes & vaultSettings

Timeline

24 events · newest first
Jun 2026
Jun 6ANALYSIS
PeptideBiz unit economics + 3-month-bundle pricing plan

Ricki added the full PeptideBiz wholesale price sheet (data/pricing/PeptideBiz-Price-Sheet-2026.pdf) and asked me to figure out what we need to sell each SKU at, assuming the discovery-call cost stack ($35 ship + $20 doctor + ~3.5% processing + $250/mo platform + private-label +$1/vial under 1k vials/mo) and a 3-month-upfront default cadence. Saved the call transcript at calls/2026-06-05-morrison-group-peptidebiz-discovery.md and the full plan at analyses/2026-06-06-peptidebiz-unit-economics-and-pricing.md. Headlines:

  • Tirzepatide 3-mo starter @ $797 ($266/mo blended, 77% margin first cycle). Market is $299-449/mo at Hims/Henry/Ro.
  • Semaglutide 3-mo @ $497 ($166/mo, 79% margin). Entry tier for price-sensitive funnel.
  • Bundles win. "Total Reset" (Tirz starter + BPC-157 + Sildenafil) at $997 has 70% margin because the $55 fixed ship+doc gets charged once across 2-3 products.
  • The 15-day merchant hold is the real ceiling on ad spend — at $5k/day Meta we need $75k of working float. Hard cap Phase-1 at $1.5k/day until vendor cuts the hold or we file our own 503B logistics cert (~$3k, ~2mo) to take over the merchant relationship on Stripe.
  • CAC math: $586 GP per blended starter → $195 affordable CAC at single-cycle, $297 at 6-mo retention, $378 at 9-mo. We can survive Meta CAC of $200-300 only if retention is ≥6 months — push annual prepay ($2,997 Tirz/yr) on first checkout.
  • Phased GTM: Wk 1-4 single Tirz SKU validate funnel → Wk 5-8 add Sema tier + stack upsell → Wk 9-12 annual prepay + SMS retention → Month 4 own merchant. Open items still needed before launch: confirm exact processing %, tier-1 volume threshold definition, bulk pre-buy terms for B2B clinic dry-vial channel, real insurance quote, written ETA on hold reduction, finalized API webhook spec.
May 2026
May 30ANALYSIS
Peptide evidence tiering — 9 peptides enriched against Europe PMC + ClinicalTrials.gov

Ricki asked for a research run across the catalog to (a) ground marketing copy in real citations and (b) train the Guide AI to be the smartest voice in the peptide space. Patched enrich_from_research.py (IPv4 force, name-in-abstract filter, CITED→relevance fallback, new coaching synthesis pass for dosing/when-beneficial/when-not/monitoring), then ran 9 peptides: semaglutide, tirzepatide, retatrutide, tesamorelin, sermorelin, pt-141, mots-c, bpc-157, ghk-cu. Honest read across all 9: peptides are a three-tier story — Tier 1 GLP-1 class is statin-level real medicine (15-24% body weight loss with NEJM-published CV-outcomes data), Tier 2 (tesamorelin, PT-141, sermorelin) is narrow-but-real FDA-approved use, Tier 3 (BPC-157, MOTS-c, GHK-Cu) is mechanistically interesting but human evidence is thin or absent. The defensible marketing position is to tell all three stories honestly — that's the credibility moat vs. competitors who oversell everything equally. Site change same session: removed the "Aging, on your terms" section from projects/superstar-peptides-v2/index.html (lines 206-304 — heading, 3 cards, "It's not energy" split, duo CTAs). Live at superstarpeptides.ai/v2/. Update later same day — brief generator built + 9 briefs written. ~/services/peptide-bio-api/generate_brief.py (3rd Claude synthesis pass) turns each research/<id>.json into a 9-section Markdown brief saved to data/peptide-briefs/<id>.md. Tier-aware: Tier-A briefs carry confident NEJM-cited claims with calibrated guardrails; Tier-C briefs prefix every claim "preclinical animal studies show…" and dosing fields say "Confirmed with prescribing clinician". Index README auto-generated, grouped by tier. Format approved by Ricki (semaglutide + bpc-157 reviewed as Tier-A vs Tier-C extremes). Update later same day — 10 more singles enriched + briefs generated (19 total). Added nad, selank, semax, ll-37, epithalon, thymosin-a1, dsip, gonadorelin, glutathione, aod-9604 to QUERIES. Final tier split: Tier A=6 (gonadorelin, pt-141, semaglutide, sermorelin, tesamorelin, tirzepatide), Tier B=11 (dsip, epithalon, ghk-cu, glutathione, ll-37, mots-c, nad, retatrutide, selank, semax, thymosin-a1), Tier C=2 (aod-9604, bpc-157). All briefs in data/peptide-briefs/. Known grader nuance to fix later: the A/B/C grader gives "B" when Europe PMC has ≥1 RCT regardless of ClinicalTrials.gov registry presence. So peptides with 0 registered modern trials but old Russian/foreign RCT papers (dsip, semax, epithalon) get tagged B even though they're effectively C in clinical practice. The brief TEXT correctly says "human evidence is sparse" for these — the tier label just over-promises. Future: tighten grader to require ≥1 registered trial for B. Open: (1) Stacks — build composer that synthesizes per-stack briefs from component peptide briefs (no PubMed search; 7 stacks + tap-141-nasal); (2) wire briefs into Guide AI system prompt / RAG; (3) Phase 2 — Hermes job for Huberman/Attia/Diamandis transcripts + mainstream news. Full analysis: analyses/2026-05-30-peptide-evidence-tiering.md. Briefs: data/peptide-briefs/. Raw enriched JSON: ~/services/peptide-bio-api/research/<id>.json.

May 17MILESTONE
Mobile app — Shop + telehealth handoff flow shipped

Ricki uploaded the patient pricing sheet (incoming-screenshots/Screenshot_2026-05-17_at_4.24.39_PM.png) and asked for an in-app shop that mocks the telehealth purchase flow — architected so swapping the mock for a real provider (Strive / Hims / Empower) is just a code-level swap. Backend (peptide-bio-api):

  • peptides_catalog.py — 25-product catalog (4 nasal sprays + 21 injectables, including 7 stacks and 2 value-tier compounds). Prices in cents per 1/3/6/12-mo tier, sourced verbatim from the pricing sheet (NAD nasal $45, single compounds $40, stacks $50, gonadorelin/glutathione $19).
  • Migration 004-shop.sql — shop_orders table with provider field abstracting the telehealth backend (mock today, swap-in later). Append-only timeline JSONB for the status UI.
  • New endpoints in app_api.py: GET /api/app/shop/catalog, POST /api/app/shop/orders, POST /api/app/shop/orders/{id}/intake, GET /api/app/shop/orders/{id} (with mock state auto-advance on poll), GET /api/app/shop/orders, POST /api/app/shop/orders/{id}/cancel.
  • Mock state machine: draft → intake_submitted → physician_review → approved → charged → shipped, advancing 25 seconds per stage on GET (replace _maybe_advance_mock() with provider webhook ingestion for real integration). Mobile (~/projects/superstar-peptides-app/mobile/):
  • screens/shop.jsx — browse with category filter pills (All / Nasal / Injectables / Stacks), horizontal "For your goals" rail that uses profile.goals to score recommendations against peptides_catalog.tags.
  • screens/shop-product.jsx — detail screen with hand-drawn SVG vial/nasal-bottle, common-uses bullets, fact grid (route/strength/schedule/status), and tier selector showing per-month math + savings % vs 1-mo baseline.
  • screens/shop-intake.jsx — telehealth intake: DOB, sex, height/weight, conditions, medications, allergies, pregnant/nursing toggle, consents. Submits to backend which advances order state.
  • screens/shop-status.jsx — animated 5-stage timeline (intake → review → approved → charged → shipped) with live polling, cancel CTA, order summary.
  • components/ui.jsx TabBar — Labs replaced with Shop in the 5-tab grid (Labs still routes via Home tile).
  • screens/home.jsx — Ask Guide tile (was a no-op) replaced with Shop peptides accent-soft tile, deep-linking to Shop tab. End-to-end verified on live peptides.cynthiaconcierge.com: signup → catalog (25 products, $45/$40/$50/$19 tiers all correct) → create order ($120 for 3mo BPC-157) → submit intake → poll status auto-advanced physician_review → approved after the 25-sec mock window.
May 17MILESTONE
OpenWearables platform deployed locally (evaluation)

Ricki asked how easy it'd be to drop OpenWearables (MIT, github.com/the-momentum/open-wearables) behind the mobile app's Connect screen — currently a pure UI mock with no real OAuth/SDK wiring. Stood it up at /home/cynthia/services/open-wearables/ via docker compose up -d --build. Eight containers: postgres 18 (5433), redis 8 (6380), FastAPI backend (8000), celery worker + beat, flower (5555), svix webhooks (8071), Vite frontend (3340). Build ate ~2.7 GB; root disk now at 97% (5.6 GB free) — tight. Backend health verified:

  • /api/v1/oauth/providers returns 11 providers: apple, samsung, google, garmin, polar, suunto, whoop, strava, oura, fitbit, ultrahuman.
  • Native-required (has_cloud_api: false): apple, samsung, google. Cloud-OAuth (works from PWA): the other 8.
  • Admin login works as ricki@cynthiaconcierge.com + seeded password (form-encoded, OAuth2 password grant — JWT issued). Eight Sleep is NOT supported — would need to drop that tile or build it ourselves (no public API exists either way). Confirms initial estimate: cloud-OAuth wearables (Oura/Whoop/Garmin/etc.) are ~1–2 days of wiring to hook into our existing connect.jsx Connect buttons. Apple Health still needs a native shell (Capacitor + HealthKit plugin or RN rebuild) — OpenWearables doesn't fix that gap. Stack is running but no provider client_ids configured yet — just placeholders. Next step (when ready): register an Oura developer app, drop creds into backend/config/.env, restart backend__open-wearables, and wire one Connect tile through.
May 17MILESTONE
Mobile app — real backend state shipped (profile, doses, scans, lead events)

Goal Ricki set: free B2C app as lead-gen funnel that hands off to a telehealth purchase. Three things had to become real for that to work: continuity (AI actually knows the user across sessions), tracking (Mark-as-taken is a real event, not a button animation), and conversion (a button that actually moves the user toward buying). Shipped all three today. Backend (peptide-bio-api):

  • Migration 003-app-state.sql — 4 new tables keyed by peptides_users.id: app_profiles, app_dose_log, app_scans, app_lead_events.
  • New app_api.py router on /api/app/* — profile (GET/PUT), doses (POST/GET), scans (POST/GET), lead-events (POST). All JWT-protected via existing auth.current_user.
  • guide.py:_build_context() now fetches the server profile + last 14d adherence + last 3 scans for the authenticated user and includes them in Guide's plan context. Smoke test: Guide correctly read "only 1 dose in 14 days" from the DB and called it out as the user's weak spot — first time the AI is using real continuity data. Mobile (~/projects/superstar-peptides-app/mobile/):
  • New lib/api.js — thin authenticated fetch wrapper.
  • App.jsx — localStorage demoted to offline cache; Postgres is source of truth. getProfile on sign-in, putProfile write-through on onboarding finish. Splash holds until server confirms (prevents an already-onboarded user from being re-routed into onboarding on a slow fetch).
  • home.jsx — "Due now" card now reads from protocol.peptides[0] instead of hardcoded BPC-157; tapping the check button calls api.logDose with optimistic UI.
  • protocol.jsx — "Mark as taken" wired to api.logDose; added per-peptide "Order {name} via clinician" button that logs telehealth_click to app_lead_events and opens /consult.html?peptide={name} in a new tab.
  • scan.jsx — mock scan results now persist to /api/app/scans once per scan run (resets on rescan); added "Talk to a clinician about your results" CTA on the results screen — natural conversion moment after seeing the numbers. Verified end-to-end (smoke test):
  • Fresh signup → /api/app/profile PUT/GET → dose POST/GET → scan POST/GET → lead-event POST: all 200, all rows in DB.
  • Guide chat read the persisted state and produced a personalized response: "You've taken BPC-157 only once in the last 14 days. Your last scan (May 17): vascular age 28, HRV 72, hydration 78%. What's getting in the way?"
  • Built bundle deployed; peptides.cynthiaconcierge.com/app/ 200, index-ByCXLYgD.js contains the new /api/app/profile call. Files touched:
  • services/peptide-bio-api/migrations/003-app-state.sql + run-003.js — new
  • services/peptide-bio-api/app_api.py — new
  • services/peptide-bio-api/guide.py — _build_context pulls server profile + doses + scans
  • services/peptide-bio-api/main.py — wired app_api_mod.router
  • projects/superstar-peptides-app/mobile/src/lib/api.js — new
  • projects/superstar-peptides-app/mobile/src/App.jsx — server-backed profile w/ offline cache
  • projects/superstar-peptides-app/mobile/src/screens/home.jsx — dose log + protocol-driven Due Now
  • projects/superstar-peptides-app/mobile/src/screens/protocol.jsx — dose log + per-peptide telehealth CTA
  • projects/superstar-peptides-app/mobile/src/screens/scan.jsx — scan persistence + scan-results telehealth CTA Open from this plan (next chunks):
  • Telehealth handoff currently routes to web /consult.html (the existing mock booking flow). When Olympia/503A partner is locked (queue.md Phase 0), swap the URL and add Rx-request flow.
  • Adherence ring on Home — pull /api/app/doses and show "12/14 days on plan" instead of fake stat. ~30min.
  • Scan results re-display — Progress screen still uses hardcoded data; could pull from /api/app/scans to show real history.
  • Tavia-facing dashboard on Spotlight — read app_lead_events aggregates to show conversion funnel.
May 17MILESTONE
Mobile-app design canvas received + hosted for review

Ricki uploaded superstar_peptides_2_.zip (21 MB, 91 files) — a clickable iOS-frame mockup of the Superstar Peptides mobile app. Built as React 18 + Babel-standalone (no build step), wraps the design in a scaled iPhone (402×874) frame with a live "Tweaks Panel" for dark mode / accent color / onboarding style / screen jumps. App structure (10 screens):

  • Onboarding (chat OR form variant, toggleable)
  • Main tabs: Home / Face Scan / Lab Reports / Aria Chat / Progress
  • Overlays: Protocol detail, Scan, Schedule, Connect Also bundled for handoff: brand identity (brand.jsx, logo lockups, palettes, applications), packaging mockups (packaging.jsx), social kit, and the existing web pieces (bioreveal-*.jsx, home.jsx, quiz.jsx, product.jsx) for reference. Hosted at: https://peptides.cynthiaconcierge.com/app-preview/ — static, served by nginx alias to ~/projects/superstar-peptides-app/design/. Original Superstar Peptides.html symlinked to mobile.html (URL-safe). Open on phone or desktop; the iOS frame auto-scales to viewport. Source files:
  • Zip: ~/incoming-screenshots/superstar_peptides_2_.zip
  • Unpacked: ~/projects/superstar-peptides-app/design/
  • nginx route: /home/cynthia/config/nginx/peptides.cynthiaconcierge.com → new location /app-preview/ block Plan agreed:
  1. Phase 1 (now) — review/iterate on canvas in browser at the URL above
  2. Phase 2 — port to React Native + Expo at ~/projects/superstar-peptides-app/native/ once design locked. Stack: Expo SDK + TypeScript + nativewind + @react-navigation/native. Preview via Expo Go on iPhone (QR-scan, hot reload).
  3. Phase 3 — Apple Developer Program ($99/yr) → EAS Build cloud → .ipa → EAS Submit → TestFlight → App Store review Open questions for Ricki:
  • Apple Developer account: new for Superstar, or roll under existing Cynthia account?
  • Bundle ID: com.superstarpeptides.app (or alternative)?
  • Backend: reuse cynthia-peptide-bio-api or stand up app-specific endpoints for push / IAP?
May 12MILESTONE
BioReveal v2 foundation shipped — 5 tracks

Ricki direction: make BioReveal the spine of peptides.cynthiaconcierge.com, keep free Claude OAuth, integrate Shen.AI later. Shipped today: 1. OAuth contention fix. Added shared OAUTH_LOCK (synthesize.py + chat.py) serializing all peptide-bio-api → Claude requests so the service never competes with itself. Backoffs extended to [2,5,15,45]s with 5 attempts. Switched system prompt to 1-hour ephemeral cache (extended-cache-ttl-2025-04-11 beta) — formulary + analysis schema (~6K tokens) now cache for an hour, only intake bills per call. Removed the public "⚙ AI MODE: RULES ENGINE" yellow banner from bioreveal.jsx. 2. Supabase persistence. New tables: bioreveal_users, bioreveal_sessions, bioreveal_check_ins (migration 001-bioreveal-sessions.sql applied via run-001.js). Direct Postgres via psycopg[binary] + psycopg_pool (PostgREST/CF path was timing out — same reasoning as services/lib/pg.js). Every /api/synthesize now writes a row, returns session_id. Front-end stashes it in sessionStorage for handoff to consult. 3. Real email send. email_send.py via Resend HTTP API (reused MEET_SMTP_PASS as Resend API key). New endpoint /api/email-analysis accepts {session_id, email, protocol}, looks up session in DB (or accepts protocol inline), sends a styled HTML analysis email + plain-text version, marks email_sent_at. Replaces the mock modal in bioreveal-reveal.jsx with real fetch + sending/error state. 4. Funnel cut from 8 steps to 3. New flow: Goals → BioScan → Safety. Built StepSafety (one compact screen: age, sex, pregnant Y/N, cancer-history Y/N, optional meds + allergies free text). Deleted feel/training/vices/basics/medical/email from stepDef. Email gate moves to the post-Reveal modal where it belongs. Hero meta updated from "8 quick steps" → "90-second scan". 5. Shen.AI scaffold behind feature flag. New bioreveal-shen.jsx with ShenStep component that lazy-imports @shenai/sdk from jsDelivr. Activated via ?shen=1 query param AND window.SHEN_API_KEY set in bioreveal.html. Falls back to legacy StepBioscan otherwise. When the Shen sales rep delivers a license, paste the key into the <script>window.SHEN_API_KEY="..."</script> block and hit ?shen=1 — no code change needed. Verified end-to-end:

  • /api/health → 200, auth_mode: oauth, claude_enabled: true
  • /api/synthesize from public URL → returns session_id, row persisted in bioreveal_sessions
  • /api/email-analysis → Resend API 200, email delivered to rickibodner@gmail.com
  • bioreveal.html + all 9 jsx files serve 200 Open from this plan:
  • Tavia / Ricki: get Shen.AI sales rep, fill developer.shen.ai form (need: company, expected volume, web SDK question, HIPAA posture)
  • MD-approval Slack/dashboard (Tavia's clinician friend) — schema field md_status already in place, queue is WHERE md_status='pending'
  • Day 7/14/28 check-in cron — table bioreveal_check_ins ready
  • Re-scan delta UI ("compare your May 12 scan to today") — parent_session_id chains re-scans
  • Cynthia Spotlight panel for BioReveal sessions count + MD-queue depth
  • Prompt-cache hit rate metric (Claude returns cache_read_input_tokens — log it for visibility) Files touched:
  • services/peptide-bio-api/synthesize.py — OAUTH_LOCK, 1h cache TTL, longer backoff
  • services/peptide-bio-api/chat.py — same lock, same cache
  • services/peptide-bio-api/db.py — new (psycopg pool, upsert_user, create_session, get_session, mark_email_sent)
  • services/peptide-bio-api/email_send.py — new (Resend HTTP API)
  • services/peptide-bio-api/main.py — wired DB + email + /api/email-analysis
  • services/peptide-bio-api/migrations/001-bioreveal-sessions.sql — new
  • services/peptide-bio-api/migrations/run-001.js — new
  • services/peptide-bio-api/requirements.txt — added psycopg deps
  • projects/superstar-peptides/bioreveal.jsx — 3-step funnel, banner removed, Shen flag, real intake mapping
  • projects/superstar-peptides/bioreveal-steps.jsx — new StepSafety, eyebrow updates
  • projects/superstar-peptides/bioreveal-bioscan.jsx — eyebrow 02/03
  • projects/superstar-peptides/bioreveal-reveal.jsx — real email send wiring, sending/error states
  • projects/superstar-peptides/bioreveal-shen.jsx — new feature-flag step
  • projects/superstar-peptides/bioreveal.html — Shen script slot, cache v=12
May 7MILESTONE
/partners marketing funnel + partner enablement showcase shipped to peptides.cynthiaconcierge.com

Ricki: build the lead-gen funnel for businesses (gyms / yoga / salons / chiros / fitness chains) — and treat them as sales reps with a full enablement system. Show what they'd get: guides, dosing references, sales scripts, affiliate link tracking, email/CRM integrations, one-click "blast to my list" tooling. Built and live (all serving 200 over HTTPS): Public marketing funnel:

  • /partners.html — master hub. Hero, 4 stats, 6-step "how it works", 5 archetype cards, full enablement system showcase (affiliate link, content library, campaign library, integrations grid w/ 16 tools, real-time dashboard mockup, Cynthia AI agent), earnings calculator with venue picker, compliance trust section, testimonial slot, final CTA.
  • /partners/{hair-salons,yoga-studios,chiropractors,gyms,boutique-fitness}.html — 5 archetype-specific landing pages. Shared archetype.jsx template reads window.PARTNER_SLUG + data.js to render: hero with archetype glyph + accent color, why-yes grid, peptide stack, talk-track lines, Peptide Day event explainer (dark section), embedded calculator with archetype defaults, "what you get" 8-card grid, nutshell math, FAQ accordion, final CTA.
  • /partners/apply.html — multi-step BioReveal-style qualification (5 steps: archetype → business → motivations → contact → confirmation). Local sessionStorage stub for now; Calendly link for discovery call booking on success.
  • /partners/playbook.html — visible Partner Library preview. 6 tabs: Peptide Reference (8 peptides), Dosing Quick-Cards (5 cards), Sales Scripts (per-archetype switcher), Email Templates (sample), Compliance Cheat Sheet (always-say / never-say split), Peptide Day Playbook (8-row YOU/CYNTHIA timeline). 3 tabs free preview, 3 tabs marked "partners only" — full library shown after onboarding. Tech architecture:
  • partners/data.js — single source of truth for all archetypes (positioning, headline, peptide stack with deep-link slugs to /peptides/, talk-track lines, Peptide Day hook, calc defaults, nutshell math, 4-question FAQ each)
  • partners/calculator.jsx — interactive earnings calc, archetype-aware. Sliders for client count + events/yr. Live math: consults → conversions → patients × LTV. Email-gate at the end for "full breakdown PDF" — captures lead via sessionStorage stub, ready for real backend.
  • partners/archetype.jsx — shared per-archetype landing template
  • partners/apply.jsx, partners/playbook.jsx — page-specific
  • ~980 lines of CSS appended to bioreveal.css (now 4,194 lines) Brand updates:
  • brand.jsx nav: added "For business" → /partners.html
  • brand.jsx footer: added "For Business" column with 7 sub-links Compliance threading throughout:
  • Every page emphasizes flat FMV retainer + flat event fees — never per-Rx commission
  • "PC-MSO friendly-PC" structure called out on hub
  • "FTC disclosure auto-appended" mentioned on email-template card
  • "503A/B pharmacy partner carries Rx liability" on compliance section
  • Archetype FAQs handle: state board concerns, scope of practice, doping (gyms), GLP-1 framing (yoga), sell-out fears (yoga), liability (chiro), Nutrafol-vs-us (salons) The "show them how easy it can be" piece: The Playbook viewer is the standout artifact. Prospects can browse the actual sales scripts, dosing cards, FAQ handlers, compliance rules, and Peptide Day playbook BEFORE applying. This is the proof that the enablement system is real — not vaporware. Integrations promised (UI shows them as available, real implementation is Phase 2 backend): Mailchimp, Klaviyo, Constant Contact, ConvertKit, ActiveCampaign, HubSpot, Mindbody, Vagaro, Square, Squire, Booker, Schedulicity, Glofox, Beehiiv, Substack, iMessage/SMS via Apple/Google. Phase 2 (backend, not yet built): Real auth + partner dashboard, real CRM/email integrations, real referral-link attribution + tagging, "send to my list" composer, Peptide Day booking workflow, MD-of-record routing.
May 7ANALYSIS
Tier A affiliate program design — non-Rx venues (gyms / yoga / salons / chiros)

Ricki: how do we build the greatest lead-gen machine for peptides, both for consumers AND for businesses (gyms, yoga, salons, med spas, chiros) that want to add peptides to their offerings? Specifically the referral side — businesses that legally can't carry Rx peptides but want recurring revenue. Drafted full program design at analyses/affiliate-program-design-2026-05-07.md. Ran two parallel research streams: telehealth/compounding-pharmacy partner program landscape + venue partnership economics by archetype. Headline finding (changes the design): Per-Rx commission is a felony in CA / FL / TX / NY under state anti-kickback statutes — applies even with zero government payor. The compliant lever is a flat fair-market-value marketing-services fee under a PC-MSO ("friendly-PC") structure. OIG Advisory Opinion 25-03 (June 2025) is the blueprint. Affiliate program lives at the MSO layer, never the PC. FDA reality (April 2026): Sermorelin = Cat 1 safe. BPC-157, TB-500, Melanotan II, GHK-Cu injectable, Semax = removed from Cat 2 April 23 but await PCAC ruling July 23-24, 2026. Program is fully viable on the Cat 1 stack alone (TRT, sermorelin, GLP-1, NAD+ oral); peptide cluster clips on as add-ons after July. Two-tier program:

  • Tier A (new): Gyms / yoga / salons / regular chiros. FMV monthly retainer ($250-500/mo) + flat Peptide Day event fees ($1.5K). Venue refers, never touches product. LTV per partner ~$30-80K/yr.
  • Tier B (existing): Med spas / wellness clinics with NP/MD on staff. Wholesale 503A/B + 1099 rep coverage (Bradley graph). Already covered in 2026-04-30 b2b model. Per-archetype playbook grounded in research:
  • Hair salons — Nutrafol Stylist Program is the only mature non-medical-venue + Rx-adjacent referral model in market. Yellow-badge "Certified Partner" status > commission. Locator listing drives traffic back to them.
  • Yoga / pilates ($19.2B industry) — frame as longevity not weight loss. Female 28-55 AHHI $90K+ matches GLP-1/NAD+ demo perfectly.
  • Boutique fitness chains — cut deals at HQ. Crunch×Thrive (550 locs, GLP-1 access) and Planet Fitness×Ro are precedents. Equinox+Function bundle model ($249/yr vs $365 standalone, $40K/yr Optimize tier) is legally cleanest.
  • Chiros — 75% of patients already buy supplements. They CAN dispense; the only gap is the prescriber. Our MD-of-record solves it. The Joint = 1,000+ locations.
  • Independent gyms / PT studios — recurring affiliate income that decouples revenue from owner's PT hours. The "Peptide Day" pop-up engine is the conversion catalyst. Modeled on Botox party economics (proven: 4-9 clients/event, 40-60% close on warmed-up referrals). 1 NP/MD on-site 6 hrs, 24 consult slots, ~10 conversions per event = $22.5K LTV per Peptide Day. Conferences to book Q2-Q3 2026: Premiere Orlando (salons, May 30-Jun 1), Parker Miami (chiros, Jun 12-14), IDEA World (fitness, Apr 7-10). Six-month rollout projects $1.6M annualized profit run-rate from Tier A alone (50 partners by month 6), on top of Tier B wholesale. Three open decisions for Tavia + Ricki: (1) FMV compensation model sign-off, (2) pilot archetype (rec: hair salons first), (3) 503A pharmacy partner (rec: Olympia). queue.md restructured. Need to brief Tavia on the FMV-compliance constraint — meaningfully changes how the May 4 "TikTok Shop / get the doctor friend involved" thread translates into venue partnerships.
Apr 2026
Apr 30NOTE
AI-medicine landscape scan — what to plug into BioReveal or spin out

Ricki asked what's exciting in AI medicine / diagnosis / tracking that we could add. Saved full research at notes/ai-medicine-landscape-2026-04-30.md. Top adds for Superstar (ranked by leverage):

  1. FaceAge — biological-age-from-selfie (Mass General Brigham + Lancet validated, FAR = cancer prognostic biomarker). Drop-in BioReveal hook → Sermorelin/NAD+/GHK-Cu protocols.
  2. Wearable + lab fusion (Ultrahuman Blood Vision pattern) — connect Oura/Whoop/Apple + last lab → platform names the protocol.
  3. CGM-driven metabolic peptide onboarding — Levels/Lingo/Ultrahuman white-label. Funnel for GLP-1-adjacent peptides.
  4. Vocal biomarkers as cognitive layer — extend BioReveal voice step → cognitive-age / stress-load → Selank/DSIP/CJC-1295.
  5. Epigenetic age clocks (TruDiagnostic) — cheek-swab reseller, killer 90-day before/after.
  6. At-home blood test kit as consult preamble — Tasso/Empower/Quest, gives MD-of-record real numbers and lifts conversion.
  7. Retinal screening from phone camera — adds cardio/diabetes risk read.
  8. Digital-twin / simulate-the-protocol projection — even a marketing-grade outcome curve would lift conversion. Sister brands worth considering: voice-only neurodegenerative diagnostic; female perimenopause/menopause peptide brand; GLP-1 companion app (lane Medvi opened); epigenetic-age- only single-product. Recommended single move: FaceAge + epigenetic-age + wearable-fusion layered into BioReveal. That trio = "first consumer product that gives a defensible biological age and a way to lower it" — also the exact multimodal-longevity story investors are funding now.
Apr 30DECISION
B2B budget right-sized — $140K lean / profitable by month 4

Tavia + Ricki pushback: original $750K-$1.5M budget options were oversized. With the AI stack (Reach, BioReveal, OpenClaw) doing the ops/marketing/sourcing work, the real ask is just the people who close. Three humans:

  1. 2 inside sales reps ($3K base + 10% commission) cold-call med spas
  2. 1 rep recruiter ($3K base + $700 bounty per landed rep — matches Bradley's company rate)
  3. MD-of-record ($2K/mo retainer for Rx oversight) Plus $1K/mo pharmacy + tools = $12K/mo fixed. One-time costs (legal $7K, pharmacy buy-in $15K, sample kits $6K, working capital $10K) + variable bounties ($17.5K) + commissions ($20K) bring 6-month total to ~$140K. Break-even: 3 active reps cover the $12K monthly burn. Realistic ramp puts us at 10 reps onboarded by month 4 → cash-flow positive. Net cash exposure: ~$90K before the business funds itself. Analysis page rewritten at https://analysis.cynthiaconcierge.com/superstar-peptides-b2b — replaced the Option A/B budget section with a single lean path showing fixed-cost table, variable/one-time table, breakeven math, and the "why $140K not $1.5M" explanation. Compressed milestone timeline from 18 months to 6 months. Pitch card ask line updated. Cover thesis preserved.
Apr 30MILESTONE
B2B business model published as analysis page

URL: https://analysis.cynthiaconcierge.com/superstar-peptides-b2b File: ~/analysis/superstar-peptides-b2b.html (52KB) Built investor-grade page using the analysis template aesthetic (JetBrains Mono, white/black, purple accent) with 14 sections. Centerpiece is a custom 5-node money-flow diagram showing Patient → Clinic → Us → Rep ← Pharmacy with dollar amounts on every arrow. Key visual moments: stat grids on opportunity + Bradley graph stats, animated funnel bars, 3-year projection table, side-by-side budget comparison cards (Option A $750K vs Option B $1.5M with RECOMMENDED badge), milestone timeline rail, and a final pitch card with the $1.5M / $12M ARR ask. Designed to work as both a 90-second scan and a 15-minute deep read. Tavia + Ricki can send this URL directly to any investor.

Apr 30ANALYSIS
B2B business model + investor brief drafted

Tavia + Ricki: needed an investor-grade B2B model for the wholesale-into-clinics play, leveraging Bradley's 1099 rep graph. Drafted at analyses/b2b-business-model-2026-04-30.md. The thesis in one line: Superstar Peptides becomes the supplier of record for the U.S. clinic-channel peptide market by hijacking the 1099 multi-line rep universe that already calls on every med spa weekly. Numbers grounded in real Bradley data (not invented):

  • 1,705 mapped reps in our graph; 229 strong fits already LLM-gated
  • $0.11 per ICP-qualified rep sourcing cost (Apify scrape + Haiku gate)
  • 67% yield on distributor scrapes (vs 17% on manufacturers)
  • ~5K-8K addressable rep universe in the U.S. (market-sized estimate) Model output (realistic, not aggressive):
  • Y1: 30 reps × 40 accounts × $1.5K/mo = $7M GMV → $2.5M gross profit
  • Y2: 100 reps → $36M GMV → $14.5M gross profit
  • Y3: 250 reps → $130M GMV → $57M gross profit
  • Per-rep CAC: ~$390 fully-loaded; payback ~2 weeks against $297K Y1 gross profit per rep Two budget options pitched:
  • Option A — Lean Seed $750K (12 mo, 25 reps, $2-4M ARR run-rate)
  • Option B — Series Seed $1.5M (18 mo, 75 reps, $12M ARR run-rate, primed for Series A)
  • Recommendation: Option B — $750K is "look it works," $1.5M is a defensible Series A story. Cross-leverage with Bradley made explicit: Bradley pays us to source his reps. Our peptide brand becomes the highest-paying line those reps carry. Symbiotic, not cannibalistic. Open risks called out: 503B partner selection, state-law variability (TX/FL/CA/NY/AZ priority), GLP-1 headline confusion, rep adverse-selection. Mitigations noted for each. Diagnosis: /api/health showed auth_mode: oauth and the token was valid. The "AI is fake" symptom wasn't an auth problem — it was 429 rate-limit contention. The peptide-bio-api shares the same Claude Code OAuth token (~/.claude/.credentials.json) as Ricki's interactive Claude Code sessions. When CC was active, BioReveal calls hit 429, and the old code threw immediately → caught upstream → fell straight to the deterministic _FALLBACK_PROTOCOL rules engine. Investor demo would see lookup-table outputs while the UI claimed AI. Fix shipped (no new credentials needed):
  • synthesize.py:_claude_via_oauth and chat.py:_claude_oauth now wrap the httpx call in a 4-attempt loop with [1s, 4s, 12s] backoffs on 429/500/502/503/504
  • On 401 they re-read ~/.claude/.credentials.json once (in case Claude Code rotated the token mid-call) and retry with the fresh token
  • import time added to both files
  • PM2 service cynthia-peptide-bio-api restarted, health endpoint confirms auth_mode: oauth, claude_enabled: true Net effect: OAuth path is now bulletproof against the most common failure mode (rate-limit contention) without provisioning a separate ANTHROPIC_API_KEY. Rules engine only fires when Claude is genuinely down for >17 seconds.
Apr 30MILESTONE
Real product photos + navbar fix

Tavia: shop pages were showing the abstract SVG glyphs where users expect the actual product photo, and the home logo in the nav was broken from sub-routes. Real product photos:

  • Shop hero stage now uses <img src="/assets/vial.png"> instead of the line-art glyph. Backdrop pulls from PRODUCT_PALETTES in brand.jsx (per-peptide bg + cap color + tint), so each SKU reads visually distinct: Sermorelin = warm clay + orange cap, NAD+ = cream + gold, GHK-Cu = blush + copper, BPC-157 = warm taupe + brown, etc. (10 palettes, all 10 peptides covered.)
  • 3 thumbnail variants per page: default backdrop / cap-color backdrop / dark "LYO · STERILE" backdrop. Vial photo with appropriate drop-shadow for each.
  • Pair cards on the "Pairs well with" tab also use real product photos.
  • Pathway diagram (the dark mechanism explainer in "What it does" tab) keeps the abstract glyph — that's about mechanism, not the product itself.
  • Added .shop-stage-photo, .shop-stage-tint, .shop-stage-product-name CSS classes. Navbar fix (was broken from sub-routes):
  • Brandmark href was index.html (relative). From /shop/sermorelin.html, that resolved to /shop/index.html → 404. From /peptides/sermorelin.html → resolved to /peptides/ (Library page, wrong destination). Changed to href="/".
  • nav-cta button "Find my peptide →" was bioreveal.html (relative) — same bug. Changed to /bioreveal.html.
  • 5 footer Learn links (learn.html#what, learn.html#safe, etc) were relative — fixed to absolute.
  • nav-main links were already absolute. No other relative hrefs found in nav/header components. Verified live: shop/sermorelin.html 200, vial.png 200/2.1MB, brand.jsx 200/16.3KB, deployed brand.jsx confirms <a href="/" className="brandmark">.
Apr 30MILESTONE
Shop product pages redesigned to match Vercel layout

Tavia: liked the layout of the original Vercel product pages — buy-box hero with subscribe banner + the 5-tab structure (What it does / How you use it / Pairs well with / The science / Questions). Pulled /product.jsx from superstar-peptides.vercel.app to extract the structure, then rebuilt for our data. Hero rewrite:

  • Breadcrumb (SHOP / [peptide])
  • LEFT: product stage with peptide glyph in accent-gradient bg + 3 view-mode thumbnails + purity overlay (99.6%)
  • RIGHT: buy box — chips row (★ MOST POPULAR, RX, class), display title, class label, italic tagline, 5-star rating row (4.9 + computed reviews count), short paragraph, 3-tier pricing buttons (1mo/3mo/6mo with SAVE 10% / BEST VALUE badges), Subscribe & save 15% banner with toggle (subscribe is on by default — 15% comes off all tiers when active), big price + crossed-out original when subscribing, dose info, "Start consultation →" primary CTA, "First, explain it to me" ghost CTA, trust note. 5-tab structure:
  • What it does: mechanism prose + "what to expect" promise card (uses the_promise from shop/data.js) + 4 benefit cards auto-derived from indications[] (split on em-dash → title + desc) + sticky dark Pathway Diagram side-card with peptide glyph in accent radial halo, 3 outcome nodes, citations caption.
  • How you use it: numbered step cards (uses how_you_use_it[] if present, falls back to dose_ladder[]) + "what ships in the box" card with accent left-border (uses whats_in_the_box[]) + sticky timeline card on the right (week-by-week with first-dot accent halo).
  • Pairs well with: smart cross-reference — derives recommendations from PROTOCOL_DB by finding all protocols containing the current peptide, ranking other peptides by co-occurrence frequency. Each card shows match% (computed from frequency), peptide glyph in accent gradient, class, name, tagline, "stacked in N protocols" meta, price + View →. If no protocols include this peptide, shows empty state with BioReveal CTA.
  • The science: 4-stat dark cards (cited studies, year established from FDA status regex, half-life, avg lot purity) + numbered citation cards with accent left-border (uses evidence[]).
  • Questions: Accordion FAQ — first item open by default, + button rotates and turns accent-colored when open. Uses shop FAQ if present, falls back to peptide FAQ. New sections after tabs:
  • Review: centered testimonial card with accent border + tinted gradient (uses review from shop data when present)
  • Final CTA: dark section, "Is X right for you?" with two big buttons in accent color Per-peptide accent threading: Pulls window.PEPTIDE_ACCENTS[slug] from glyphs.jsx — same color system as Library + protocol pages, so all the new pages are visually consistent. HTML changes (all 10 shop pages): added /protocols/data.js script (for the Pairs tab cross-reference) and /peptides/glyphs.jsx script (for accents + glyphs); cache-busted to v=12. CSS: ~930 lines appended to bioreveal.css (now 3186 lines). New keyframes reuse existing protocol-glyph-float. Sticky tab strip + sticky pathway diagram + sticky timeline card. Verified live: /shop/sermorelin.html 200, /shop/product.jsx 200/25.4KB, bioreveal.css?v=12 200/68KB.
Apr 30MILESTONE
Library (formerly "Formulary") + 10 peptide deep-dive pages redesigned

Tavia: nav term "Formulary" was opaque, library cards looked stale, individual peptide pages needed the protocol-page treatment. Renamed everywhere user-facing:

  • Nav label Formulary → Library (brand.jsx)
  • Library page title → "The Peptide Library"
  • Page eyebrow ★ THE FORMULARY → ★ THE LIBRARY
  • (Internal LLM prompts in api/_synthesize.py and api/_chat.py still say "formulary" — that's a clinical term in the system prompt, no user-facing impact, left alone.) New shared file: peptides/glyphs.jsx
  • window.PEPTIDE_GLYPH({slug, color, size}) — unique line-art SVG per peptide
  • window.PEPTIDE_ACCENTS — accent color per peptide
  • 10 distinct glyphs: sermorelin (crescent + GH pulse wave), nad-plus (hex molecule + bolts), bpc-157 (split-tendon + healing cross), ipamorelin (twin rising pulses), ghk-cu (tripeptide triangle + Cu droplet), glutathione (shield + check), selank (concentric calm rings + brain wave), dsip (delta wave + Z's), b12-mic (vial + energy spike), cjc-1295 (paired pulses + plus marker)
  • 10 distinct accent colors (gold, violet, green, orange, copper, silver-blue, cyan, deep-blue, yellow, brick) Library index page rewrite (peptides/index.jsx):
  • Hero with split layout: large editorial headline + 6-mini-glyph orb grid on right (animated stagger-in)
  • Card grid: each card has accent-gradient glyph header (140px), numbered badge, class label, name, italic tagline, one-liner, foot strip with price + READ→. Hover lifts + colors border to peptide accent.
  • Final CTA section. Per-peptide deep-dive rewrite (peptides/peptide.jsx):
  • Hero matching protocol pages: big name with accent-colored period, italic tagline, one-liner, 4-stat strip ($, half-life, dose stages, MD-led), big floating glyph on right with radial halo, class chip below
  • FDA status banner uses accent-tinted background when Cat 1/approved, warm orange when limbo
  • "How it works" section: 2-col layout — mechanism prose card on left + accent-tinted "lever it pulls" callout on right (numbered indications)
  • Indications / Not-for: same Audience cards as protocol pages (accent + red top borders, dot bullets)
  • Dose ladder: numbered cards (01/02/03) with display-font numerals, accent top border, animated arrow between cards on desktop
  • Timeline: same animated vertical rail as protocol pages
  • Side effects + Interactions: clean 2-col cards with accent-tinted icon
  • Evidence: dark section with accent left-border, numbered citation cards
  • FAQ: collapsible accordion items (first one open by default), border colors to accent on open, + symbol rotates to ×
  • Final CTA mirrors protocol pages CSS: ~320 lines appended to bioreveal.css (now 2258 lines). New keyframes: lib-mini-in, lib-card-in. Cache-busted to v=12 on all 11 peptide HTML files; injected glyphs.jsx script tag. Verified live: /peptides/ 200, /peptides/sermorelin.html 200, glyphs.jsx 200/10KB, peptide.jsx 200/22KB, index.jsx 200/5.4KB, bioreveal.css 200/50.5KB.
Apr 30MILESTONE
Protocol pages redesigned — all 6

Tavia: protocol pages need to be beautiful, with graphics + animations explaining in simple language. Copy stayed; design got rewritten. What changed (protocols/protocol.jsx):

  • Hero — split layout, big display name with accent-colored period, italic tagline, custom SVG glyph per protocol on the right (floating animation + radial accent halo). 4-stat strip (peptides count, weeks-to-effect, $/mo, MD-led).
  • Audience cards — side-by-side cards with iconography, accent top-borders, dot bullets in accent color.
  • NEW: How It Works (3-stage flow) — visual mechanism explainer. Stage 01 (status quo) → animated dashed-arrow → Stage 02 (peptide one-liners as numbered nodes) → arrow → Stage 03 (outcome card with accent-tinted background).
  • Why this stack — wrapped as a pull-quote with oversized Instrument Serif quote glyph in accent color, italic display body.
  • Peptide stack cards — grid with ProductShot pen graphic in accent-gradient frame, numbered badge, dose + status meta strip, deep-dive link + Buy CTA. Stagger fade-in.
  • Timeline — vertical rail with accent gradient that animates fill on load, dots glow with accent, cards slide in from left.
  • Lifestyle plays — 3 dark-bg cards, big numerals (01/02/03), accent left-border, hover lift.
  • Final CTA — centered, radial accent blur in background, framed price card with stack chips. Per-protocol custom glyphs (Glyph component):
  • lose-weight — descending dot-chart on scale platform
  • build-muscle — flexed bicep silhouette + rising energy bars
  • grow-your-hair — scalp curve with hair strands + follicle dots
  • fix-your-skin — sun rays + face oval inside concentric glow rings
  • think-smarter — brain outline with synapse triangle + spark
  • sleep-deeper — crescent moon + delta-wave lines + stars Every glyph + section uses the protocol's hero_color as accent — peach, red-orange, gold, warm peach, purple, sky blue. Six pages feel distinct. CSS: ~290 lines appended to bioreveal.css (now 1937 lines). New keyframes: protocol-glyph-float, protocol-card-in, protocol-rail-grow, protocol-tl-in. Cache-busted to v=12 on all 7 protocol HTML files. Verified live: lose-weight.html 200, protocol.jsx 200 / 31KB, bioreveal.css?v=12 200 / 42.7KB.
Apr 30MILESTONE
Site audit + Round-1 fixes shipped

Audit findings:

  • Initial fear that nav was 404'ing was wrong — internal links use .html paths (e.g., learn.html, protocols.html), and protocols.html is a meta-refresh redirect to /protocols/. All real user paths render.
  • Final BioReveal CTA was the actual broken thing: "Book my consult" linked to product.html (a peptide list, not a booking page); "Email me the analysis" was href="#" (no-op). This is THE conversion event.
  • Chat API auth degraded — both OAuth (rate-limited; shares creds with Claude Code) and ANTHROPIC_API_KEY (returns 401 invalid). Falls back to deterministic rules engine — replies are reasonable but not "personalized AI." Flagged as P1 follow-up.
  • Five HTML files had OG meta og:url and og:image pointing at superstar-peptides.vercel.app (the abandoned Vercel deploy). Fixes shipped:
  1. New consult booking flow — consult.html + consult.jsx. Three steps: clinician picker (3 mock MDs) → date + timeslot picker (10-day window, "booked" slots greyed/strikethrough) → contact form (name/email/phone/note) → confirmation card. Reads stack context from sessionStorage.br_protocol_summary so the BioReveal stack name + total auto-display when user arrives from the funnel.
  2. BioReveal final CTA fixed — "Book my consult" now points to /consult.html. Added useEffect to stash {stackName, peptides[], total} in sessionStorage for handoff.
  3. Email modal added — "Email me the analysis" opens an inline modal (email input → confirmation state). Mock; no real send. Hooks ready for provider integration later.
  4. OG URLs migrated — sed replaced superstar-peptides.vercel.app → peptides.cynthiaconcierge.com across 5 .html files.
  5. Cache-busted bioreveal.css + styles.css to v=11. Verified live: consult.html serves 200 / 2.5KB, consult.jsx 200 / 17.6KB, bioreveal-reveal.jsx 200 / 17.5KB. All on https://peptides.cynthiaconcierge.com. Not yet done: real email send provider, real calendar/booking provider, chat API rotation. Per Tavia: mock for now → real provider later.
Apr 30DECISION
GTM sequence locked — B2C site first, B2B analysis second

Step 1: finish B2C site. Used as the fundraising artifact + proof-of-concept for investors. Step 2: produce the B2B analysis + financial model — paying 1099 reps + building a sales team to sell wholesale to med spas, gyms, clinics. Queue.md restructured to reflect ordering. Starting with site audit next.

Apr 30DECISION
Engagement defined — equity deal, dual B2C + B2B peptide brand

Confirmed with Ricki:

  • Decision maker: Tavia. Equity deal, $0 cash, no direct contact needed (handled through Cynthia).
  • The ask: build the best peptide website + lead-gen powered by our AI-medicine biometrics flow (BioReveal), AND produce a B2B analysis selling into clinics.
  • Why we have an unfair edge on the B2B side: we already work with Bradley (paid) staffing 1099 injectable aesthetic reps — same clinic universe, same IC. Reuse the rep dataset, retarget the message: from "find a rep" to "sell peptides via those reps' clinics."
  • North star: B2B + B2C peptide brand, ride the wave of the future of medicine.
  • Workspace: README + state.yaml + queue.md updated. Cross-reference added to bradley client folder.
Apr 30MILESTONE
Client workspace created

Spun up ~/vault/clients/superstar-peptides/ so the engagement is tracked alongside other clients and renders in spotlight.cynthiaconcierge.com. Site was already live on our infra (peptides.cynthiaconcierge.com → PM2 cynthia-peptide-bio-api on :3401, nginx config at ~/config/nginx/peptides.cynthiaconcierge.com, code at ~/projects/superstar-peptides/) but had no client-side workspace, timeline, or state file.

Apr 30EVENT
Protocol-matcher funnel page (2026-05-22)

Ported the Claude Design "Peptides Funnel" handoff (31-screen peptide-protocol quiz: goal → demographics → body → lifestyle → MD-credibility → medical → measurements → projection → lead capture → rule-based stack reveal → testimonials → 3-tier paywall) into the live site, rebranded to our design system (cream paper, dusty-rose accent, Archivo/Inter/JetBrains, brand star wordmark — not the design's dark/orange MadMuscles look). Self-contained React+Babel page at ~/projects/superstar-peptides/quiz/ (index.html, styles.css, ui.jsx, screens.jsx, app.jsx, tweaks-panel.jsx). Live at peptides.cynthiaconcierge.com/quiz/; linked from the shared footer (Learn column). Draft for iteration — silhouettes are SVG placeholders (image-slot dropped), paywall/lead-capture not yet wired to backend.

Apr 30EVENT
Quiz v2 — peptide-first protocol finder (2026-05-22)

Rebuilt ~/projects/superstar-peptides/quiz/ from a physique funnel into a peptide-first matcher.

  • Branching flow (app.jsx buildFlow(state), id-based nav): 6 goal lanes (fat-loss / muscle / recover / aesthetics / mind / vitality + "unsure"), each injecting its own 3–6 screen deep-dive. ~30–45 screens per path. Body silhouettes now only appear in fat-loss/muscle lanes.
  • Catalog-driven recommendation (catalog.js): client mirror of peptide-bio-api/peptides_catalog.py (28 products, real doses/tags/prices) + recommendStack(state) — scores by goal+subgoal+symptom tags, filters by delivery preference (nasal vs injectable), excludes double GLP-1, budget-aware stack sizing, Sermorelin fallback. Result + per-peptide education screens render from it.
  • Lead capture: name/email/phone before the reveal; loading POSTs full answers to /api/lead (best-effort — endpoint not built yet, fails silently). Final CTA → /consult.html.
  • Flow spec in quiz/FLOW.md. Validated: all JSX compiles (Babel 7.29), engine smoke-tested across goals, 67/67 flow ids resolve. Live at peptides.cynthiaconcierge.com/quiz/. TODO: build /api/lead to actually store leads + email the protocol.
Apr 30ANALYSIS
Competitor teardowns — top 5 TRT/peptide/GLP-1 brands (2026-05-30)

Ricki asked for the most successful supplement/peptide/TRT companies so we can analyze their style & copy for Superstar Wellness. Built analyses/competitor-teardowns/ with side-by-side teardown of Maximus, Hims, Hone, Eden, Henry Meds (the closest analogs in our exact product surface).

  • Captured live: Maximus (full HTML via curl), Hone, Eden, Henry Meds (WebFetch). Hims edge-blocked (Akamai 403) — fell back to documented public knowledge with a clear note in hims/NOTES.md.
  • Headline finding: Maximus is our direct analog — same product surface (TRT + GLP-1 + peptides + labs + mood), and they nail (a) named-institution doctor wall (Harvard/UCSF/UNC Urology/Cedars-Sinai/Tulane), (b) free-calculator-as-SEO-wedge under every category, (c) quantified result graphs (↑3.1x free T, ↓28% body fat), (d) one-question goal quiz (not Hims' 30-question marathon), (e) transparent pricing on home ("Starting at $149.99"). Their structure should be our template.
  • Other steals: Eden's "Same Price at Every Dose" guarantee for our GLP-1; Henry's "No Insurance, No Waiting Rooms" friction-removal headline; Hone's $65 blood-test entry funnel; Hims' "your personalized plan" reveal page after quiz.
  • Biggest gap on our site today: no named-institution doctor wall. Maximus' 5 advisors all have Ivy/top-program credentials front-and-center. Action item: Tavia + Cynthia to recruit 3-5 named medical advisors — single highest-leverage trust lever.
  • Next: screenshot Maximus + Hone hero shots → load into Higgsfield Soul as style references for Superstar Wellness asset factory (per the asset-workflow discussed same day).
  • Full doc: analyses/competitor-teardowns/README.md + raw captures per-brand.