SEEN:app

On track

NextPromote SEEN-5 repetition fix to proxy + profile-nudge fix before 6/20 freeze; SEEN-4 premium/Amaru timing (founder call); SEEN-6/7 design answers.

Waiting onBen review to promote Ryan's sandbox repetition edits into seenapp-proxy-server

Updated 2026-06-16

Summary

Privacy-first AI journaling. seenjournal.app. 250+ cumulative users; clean (dev-filtered) base is ~34 monthly actives, engagement depth ~20m50s avg (eased from 22m37s). Peak ~220 monthly actives during the UNK spring cohort (concluded late May; no active focus group since). Acquisition is essentially all Direct (Organic Search ~5/28d), mostly organic/word-of-mouth — no recruited cohort propping up the base. 100% crash-free, $0 revenue (nothing monetized yet). Premium ($12.99/mo, $129.99/yr / $79 intro) targets ~July 1, gated on the Apple IAP compliance rebuild; it bundles Unlimited Chat, Voice Chat, Custom Avatars, Edit Past Entries, Full Weekly Reflection, Stories, and the Held Space module (Therapy Bridge gate). Amaru proper is the separate B2B track ~30 days later, not a premium feature. Full read: 04 - Products/SEEN-app/02 - Reports/2026-06-15 Firebase Analytics Read.md.

Open

  • SEEN-4: decide premium/Amaru timing relative to the fall semester wave (founder call).
  • SEEN-5: prompt-repetition fix — FIXED + verified in Ryan's W5 sandbox (6/16, "transition question trap" → two SESSION_CONTINUITY edits to fathom-claude.js); NOT yet promoted to seenapp-proxy-server (pending Ben review — Claude API was down). Also open: profile-nudge fix (reflect-assembly.js, ||→&&). AI role clarity in onboarding still ready. Promote both before the 6/20 freeze. Next-pass polish (dropped opening thread, B-side over-claiming) filed to Ryan's dev inbox.
  • SEEN-6: design answer for memory consent/control; loop in Andi. (ready)
  • SEEN-7: written design answer to the validation-risk critique before premium ships; loop in Andi. (ready)

Recent

  • 2026-06-16: SEEN-5 repetition fix reproduced + fixed in Ryan's W5 reflection sandbox and verified clean on a second arc. Root cause = "transition question trap" (model re-asking "what shifted between X and Y?" at successive steps in different words). Two SESSION_CONTINUITY edits to fathom-claude.js. Open: promote to proxy (Ben review) + profile-nudge fix (reflect-assembly.js). Logged via Discord clock-out; vault note at seen_place_llc_dev/02 - Dev Inboxes/Ryan/Archive/6-16-26 - Reflection Sandbox Session.md. Next-pass quality notes filed to Ryan's dev inbox.
  • 2026-06-15: Weekly read. Base flat (34 monthly actives, 21 new). Engagement depth eased 22m37s→20m50s and journaling softened (completed 63→56, steps 308→266) — modest. Still no organic discovery. No active focus group since late May; corrected the stale "focus-group cohort dominates" framing carried in earlier reads. See 2026-06-15 report.
  • 2026-06-10: First clean-era analytics pull (dev-filtered). 34 monthly / 13 weekly actives, 22m37s avg engagement (up from 19m44s), journal_completed 63/28d. W2→W3 retention cliff persists; no organic discovery. Focus-group audience split not yet readable (no backfill; mature ~late June). See 2026-06-10 report.
  • 2026-06-03: SEEN-1 done (Firebase analytics read), SEEN-2 done (focus-group audience published in GA4), SEEN-3 done (dev-build filter active; clean era starts now).

Recent

  • 2026-06-16SEEN-5 prompt-repetition fix reproduced + fixed in Ryan's W5 reflection sandbox (root cause "transition question trap"; two SESSION_CONTINUITY edits to fathom-claude.js), verified clean on a second arc. Not yet promoted to the proxy — pending Ben review (Claude API down). Profile-nudge bug identified, not fixed (reflect-assembly.js). Next-pass quality notes (dropped opening thread, B-side over-claiming) filed to Ryan's dev inbox. Status capture dropped on the POTD Scratchpad for Theo. - 2026-06-03 2148: Nicolai executed the console plan with Ben authorizing: dev-traffic filter Active, UNK Focus Group audience published. Audience preview showed the cohort at ~76 percent of last-28-day sessions; organic base smaller than topline. SEEN-2 and SEEN-3 closed. - 2026-06-03 2139: Claude Code deployed the dev-analytics code fix and delivered the console action plan to 01 - Inboxes/Ben. SEEN-2 and SEEN-3 reduced to console steps. - 2026-06-03 2123: cross-cohort synthesis (Fall 2025 + Spring 2026) written; SEEN-6, SEEN-7 added from Fall-only signals. - 2026-06-03 2038: Spring 2026 focus group synthesis promoted into 04 - Products/SEEN-app; SEEN-5 added from its High priority fixes. - 2026-06-03 2035: recorded first user story (Shibly, Bangladesh) and first lifetime premium grant in 04 - Products/SEEN-app/User Stories.md. Entitlement must be honored when premium ships. - 2026-06-03 2031: live Firebase review with Ben. 12-month arc shows semester-cohort rise and fall; organic floor ~33 MAU, deep engagement. SEEN-1 done; SEEN-2/3/4 added. - 2026-06-03 2352: scaffold created.
Read more — full task list (seen-app-tasks.md)

Tasks

  • [x] SEEN-1: confirm current user count and headline metrics. (done 2026-06-03, Firebase Analytics Read in 04 - Products/SEEN-app)
  • [x] SEEN-2: UNK Focus Group audience published in GA4 (Access Tier = focus_group, at any point in time, max duration). Done 2026-06-03 2148.
  • [x] SEEN-3: dev-build analytics: code fix deployed (rig) AND "Exclude development builds" data filter ACTIVE in GA4. Done 2026-06-03 2148. Historical data stays mixed; clean era starts now.
  • [ ] SEEN-4: decide premium/Amaru timing relative to the fall semester wave. (founder call)
  • [~] SEEN-5: focus group High priority fixes before premium launch. Prompt-repetition: fixed + verified in Ryan's W5 sandbox 6/16; pending promote to seenapp-proxy-server (Ben review) + profile-nudge fix (reflect-assembly.js ||→&&) — both before 6/20 freeze. AI role clarity in onboarding still ready. Backed by BOTH cohorts.
  • [ ] SEEN-6: design answer for memory consent/control: Fall data shows cross-session memory delights some and unsettles others (unwanted resurfacing of painful topics). Loop in Andi. (ready)
  • [ ] SEEN-7: written design answer to the validation-risk critique ("dangerously affirm bad values") before premium ships. Loop in Andi. (ready)

Updates

  • 2026-06-03 2352: scaffold created.

  • 2026-06-03 2031: live Firebase review with Ben. 12-month arc shows semester-cohort rise and fall; organic floor ~33 MAU, deep engagement. SEEN-1 done; SEEN-2/3/4 added.

  • 2026-06-03 2035: recorded first user story (Shibly, Bangladesh) and first lifetime premium grant in 04 - Products/SEEN-app/User Stories.md. Entitlement must be honored when premium ships.

  • 2026-06-03 2038: Spring 2026 focus group synthesis promoted into 04 - Products/SEEN-app; SEEN-5 added from its High priority fixes.

  • 2026-06-03 2123: cross-cohort synthesis (Fall 2025 + Spring 2026) written; SEEN-6, SEEN-7 added from Fall-only signals.

  • 2026-06-03 2139: Claude Code deployed the dev-analytics code fix and delivered the console action plan to 01 - Inboxes/Ben. SEEN-2 and SEEN-3 reduced to console steps.

  • 2026-06-03 2148: Nicolai executed the console plan with Ben authorizing: dev-traffic filter Active, UNK Focus Group audience published. Audience preview showed the cohort at ~76 percent of last-28-day sessions; organic base smaller than topline. SEEN-2 and SEEN-3 closed.

  • 2026-06-16: SEEN-5 prompt-repetition fix reproduced + fixed in Ryan's W5 reflection sandbox (root cause "transition question trap"; two SESSION_CONTINUITY edits to fathom-claude.js), verified clean on a second arc. Not yet promoted to the proxy — pending Ben review (Claude API down). Profile-nudge bug identified, not fixed (reflect-assembly.js). Next-pass quality notes (dropped opening thread, B-side over-claiming) filed to Ryan's dev inbox. Status capture dropped on the POTD Scratchpad for Theo.