GEMS Maturity Audit — 2026-05-15

> Status: 5 GEMS incubated 2026-05-13. User said "unimpressive + immature".

> This audit identifies what's missing and proposes a 2-week maturity sprint.

---

Current state (S3 inventory)

|---|---|---|---|---|---|

| heartprint | /apps/heartprint | index.html · audio.mp3 · manifest | ✓ | ✓ | — |

| griefroom | /apps/griefroom | index.html · 3 audio files · manifest | ✓ | ✓ | — |

| dreamfeed | /apps/dreamfeed | index.html · audio.mp3 · manifest | ✓ | ✓ | — |

| pulsepoem | /apps/pulsepoem | index.html · poem-2026-05-13.mp3 · manifest | ✓ | ✓ | — |

| silent | /apps/silent | index.html · manifest · sw.js | — | ✓ | ✓ |

Files were ALL generated 2026-05-13 at ~20:00 UTC. None has been re-generated since.

Symptoms ("unimpressive + immature")

Without reviewing each app live, the most common immaturity signals across rapid-incubated apps are:

1. Single-shot content: each app has the day-1 generation but no daily refresh — content goes stale fast

2. No empty / loading / error states: pages assume happy path

3. No persistent state: users can't save, return, share their session

4. No social loop: no share button, no public archive of past pieces

5. No clear value prop above the fold: visitors hit it without knowing what they get

6. Mobile layout untested: looks fine on desktop, broken on phone

7. No analytics: zero feedback on what's resonating

Maturity sprint — 2 weeks, 5 ships

Week 1: Foundation

Day 1–2: One-page maturity checklist applied to each GEM

Above-the-fold value prop (1 sentence + 1 visual)
Mobile layout pass (375px width baseline)
Empty + loading + error states
Share button (Web Share API + clipboard fallback)
Past archive list (last 14 entries)

Day 3–4: Daily content refresh cron per GEM

titan-gems-daily-refresh (already has scaffolding; needs wire-up)
Each GEM gets a daily fresh artifact (poem, dream, snippet) generated by the existing Veo3+Imagen+TTS Symphony pipeline
Cache + idempotency: never regenerate same date

Day 5: Analytics

Plausible.io or self-hosted GoatCounter
One event per GEM: view, play, share, archive_browse

Week 2: Loop

Day 6–7: Email capture per GEM

Existing titan-email-capture Lambda already supports per-app tagging
"Get a new [poem/dream/snippet] every morning" → daily SES drip

Day 8–9: Public archive page per GEM

/apps/<gem>/archive/ lists past 30 days with thumbnail + date
SEO: each archive entry gets its own URL for shareable links

Day 10: A/B test the headline

Two variants of the hero copy
Plausible event tracking decides winner after 7 days

Week 3 onwards: Product-market signal

If after the maturity sprint a single GEM shows:

> 100 unique visitors/day
> 5% share rate
> 10 email signups/week

→ Double down on that one. Kill the others. Build it into a full PRISM-tier studio.

---

Specific per-GEM critique (from filename + size signals)

| GEM | Quick read | First fix |

|---|---|---|

| heartprint | 21 KB HTML, 482 KB audio. Probably a single-page heart-rate-to-music visualizer. Likely lacks a why-care hook. | Add a 5-second auto-play demo above the input |

| griefroom | 21 KB HTML, 3 audio files (108+76+88 KB). Audio split implies a multi-stage flow — uncommon UX, may confuse. | Combine into one 5-min ambient track; add a "stay for 5 min" timer |

| dreamfeed | 22 KB HTML, 307 KB audio. Likely a dream interpretation app. Will fail without daily new content. | Wire to daily Veo3 video; archive past 14 dreams browsable |

| pulsepoem | 25 KB HTML (largest), per-date poem MP3 (91 KB). Already has dated content — closest to mature. | Public RSS feed + Twitter auto-post the daily poem |

| silent | 16 KB HTML + service worker (.sw.js). PWA already enabled — only GEM with offline mode. | Most "ready" of the 5; focus first refinement here |

---

Order of operations

1. Pulsepoem first — closest to mature, has dated content

2. Silent second — PWA-ready, lowest friction to ship

3. Dreamfeed third — needs the daily-refresh wire-up

4. Heartprint + Griefroom — last; both need UX rethink

---

Owner

This is queued under #gems-overhaul for the next focused work block.

Estimated effort: 15–20 hr for the full 2-week sprint.

Recommended kick-off: after Beehiiv migration ships (since the newsletters

will start surfacing the GEMs to a larger audience and we want them tight first).