Call notes — Data Model / CMS (2026-06-29)¶
Primary source. Verbatim summary pasted from the call; distilled into 02-product-domain.md, 04-mvp-scope.md, 06-content-migration-risk.md. Transcript (Granola): https://notes.granola.ai/t/0715888a-d1b4-4f41-8390-909f7bf841f5
Note: "Payload" appears below as an example CMS alongside "Sanity, Payload, etc." The decision leans Wagtail (ADR-0001); Payload here is illustrative, not the chosen tool.
Content Architecture and UX Concepts¶
- App centers on a "portal" metaphor: splash screen animates through a door into a master homepage
- Master homepage is the first touchpoint (not theme-filtered), featuring curated content, journeys, and a top passage
- Two distinct content structuring mechanisms:
- Themes: thematic/mood-based filters (grief, joy, myth, courage, etc.); switching theme recolors the whole UI
- Journeys: editorially curated reading lists/courses (e.g., "Seven Days with CS Lewis", Narnia); can span multiple themes
- Passage types at the top of the homepage:
- From a letter: links to the full letter (short enough to hold in-app)
- From a book: links to a "Work" landing page with title, quote, and buy CTA (no e-reader)
- AI interpretations attached to each piece of content; editor QA'd but AI-generated
- Displayed as a prompt/lens before reading (e.g., "What if loneliness was a dog?")
- Applies to the chapter if content is in a journey; applies to the standalone piece if not
- Reflections tied to context:
- Standalone piece: reflection attached to that piece
- Journey chapter: reflection attached at chapter completion
- Soul Map = visual representation of bookmarks and reflections; out of scope for Phase 1
CMS Approach and Content Migration¶
- CMS is the centerpiece: the app is essentially a styled content engine
- Strong preference for an existing CMS (Sanity, Payload, etc.) over building from scratch
- Needs: taxonomy/tag relationships, hierarchical content, reverse lookups, document matching
- Content migration is the biggest risk and sequencing challenge:
- No raw content received yet (unknown format: PDF, JSON, rows?)
- AI pipeline planned to process corpus: extract passages, generate interpretations, apply tags
- Preference for a script-based approach over pure AI to avoid hallucinated quotes
- Content entry explicitly excluded from Fueled's statement of work
- Phase 1 CMS priorities:
- Basic CMS setup with Payload
- Passage defined with body type, content type (letter, poem, nonfiction), source, theme tags, draft/review/published workflow stages, image attachment
- Tag types for passages and articles
- Direct data access via API
Phase 1 Scope and Next Steps¶
- Authentication: simple email/password login; Apple ID or Google ID as stretch; guest mode for unauthenticated users
- Basic user profile and reflection storage
- API-side authentication pipeline to be started this week
- Podcast = just a receptacle for future content; not built yet
- Client needs to approve CMS recommendation before build starts (potential hosting/licensing costs not yet communicated)
- 14-week timeline is extremely tight; realistic only with a content subset to start
- Demo deferred: Lina to share a URL or reschedule for Wednesday
Next Steps¶
- Prepare CMS recommendation for client approval — client signs off before build begins; hosting/licensing costs may not yet be on their radar.
- Define content subset for Phase 1 migration — agree a manageable set of letters/passages to seed the CMS before curation or journey-building.
- Start authentication and basic user profile on the API side this week (Mayank) — email/password login, guest mode, reflection storage.
- Reschedule demo to Wednesday (Lina) — share a URL or confirm Wednesday slot.