01 / Visual Mindmap
Boardmix System Map
Interactive mindmap hosted on boardmix.com. Click nodes to explore the full system architecture, API flows, and plan tiers.
Loading boardmix mindmap…
If this doesn't load, the board may require login.
02 / Plan × API Call Matrix
Every endpoint, who gets it, in what order
Locked steps show what each tier does NOT get. Cross-verified by 2 independent agents against live code.
SEEKER · FREE
$0/mo
Basic chart + numerology. Earns trust.
1
GET/entitlements
Check plan
2
POST/calculate/swiss
Full chart (15 planets, 12 houses)
3
POST/name/full
5-tradition numerology + personal year
4
GET/posts
Blog (200+ articles)
5
GET/places/autocomplete
City search
6
POST/chat
AI Coach — LOCKED
7
POST/reading/deep
Deep reading — LOCKED
8
POST/transits
Transits — NOT BUILT
4/10
Free gives away Swiss + chat. Should gate both.
ENTHUSIAST · $19/MO
$19/mo
Daily AI guidance + chat + transits.
1
GET/entitlements
Check plan → enthusiast
2
POST/calculate/swiss
Full natal chart
3
POST/name/full
Numerology + personal year
4
POST/chat
AI Coach (GLM-5.2 + chart context)
5
GET/chat/sessions
Chat history
6
POST/guidance/daily
Daily guidance — NOT BUILT
7
POST/transits
Transits — NOT BUILT
8
POST/reading/deep
Deep reading — LOCKED (Ultra)
5/10
$19 value prop (guidance + transits) not built yet.
ADVANCED · $49/MO
$49/mo
Deeper reports, pattern analysis.
1
GET/entitlements
2
POST/calculate/swiss
+ Moshier cross-check
3
POST/name/full
4
POST/chat
Coach + persona
5
POST/reading/deep
2000-word reading
6
POST/guidance/daily
NOT BUILT
7
POST/transits
90-day forecast NOT BUILT
6/10
Deep reading works. Transits missing.
PROFESSIONAL · $199/MO
$199/mo
Swiss-only, full precision, for coaches.
1
GET/entitlements
2
POST/calculate/swiss
Swiss LOCKED (highest precision)
3
POST/name/full
4
POST/chat
Full persona + history
5
POST/reading/deep
4000+ words (GLM-5.2 monolithic)
6
GET/admin/stats
Platform admin (founder)
7
POST/transits
Full calendar NOT BUILT
8
POST/guidance/daily
NOT BUILT
7/10
Best delivered tier. Missing transits + daily guidance.
03 / Technology Utilization
How well are we using each technology?
Swiss Ephemeris
8/10
15 bodies in 40ms. Arc-minute precision.
Missing: transit calculations (current vs natal positions).
Z.AI GLM-5.2
6/10
Chat + deep reading. Guardrail blocks non-astrology.
Missing: persona persistence, daily guidance, transit interpretation.
SQLite
5/10
Users, charts, posts. Chat sessions in-memory.
Missing: chat persistence, reading storage, analytics.
Caddy 2
9/10
TLS, HTTP/3, static serving, proxy. Sub-5ms.
Missing: rate limiting on API.
React 19 + shadcn/ui
6/10
20 pages, cosmic design. 484KB bundle.
Many pages render empty (transits, horoscope, guidance).
Model Cascade
8/10
GLM-5.2 → MiniMax → Qwen → GPT-4o-mini.
Works. Could add cost tracking + A/B testing.
04 / Promise vs Delivery
What we tell users vs what works
✓
Swiss Ephemeris precision
DELIVERED. 15 bodies, arc-minute. Verified.
✓
Five name numerology traditions
DELIVERED. All 5 compute real values.
✓
Nothing invented, reproducible
DELIVERED. 60+ truth-layer assertions enforce it.
✓
AI Coach grounded in chart
DELIVERED. Chat accepts chart context. Guardrail works.
~
Daily guidance
PARTIAL. Chat can guide, but no automated daily endpoint. Dashboard empty.
~
Deep reading in 48h
PARTIAL. Endpoint works (4000+ words), no payment trigger or queue.
✗
Transits and timing
NOT DELIVERED. No endpoint. Pages render empty.
✗
Daily horoscope
NOT DELIVERED. Static placeholder only.
✗
Pause/cancel in two clicks
NOT DELIVERED. No Stripe integration.
05 / Overall Scores
The honest report card
7
Engine
Swiss + Moshier + patterns + numerology. Missing transits.
5
Product
Chart + chat + admin work. Dashboard, transits, payments missing.
8
AI Integration
GLM-5.2 + cascade + guardrail + deep reading.
9
Infrastructure
Caddy + Docker + TLS + sub-5ms. Zero downtime.
Bottom line: Engine + AI are strong (7-8/10). Product layer is the bottleneck (5/10) — 4 endpoints needed: transits, daily guidance, horoscope, payments. Infrastructure is production-ready (9/10). Cross-verified by 2 independent agents against live code.