Kornerway
Partner Portal · v1.1
Invalid credentials
Partner Portal
System health · Pilot metrics · Production analytics
Infrastructure
System health
⬡
Railway API
Checking…
—
⬡
Database
Checking…
⬡
LLM Pipeline
Checking…
Data Volume
Ingestion metrics
Reviews
Total ingested
Reviewers
—
Unique profiles
Items
—
Profiled venues
Domains
—
Active verticals
Tokens MTD
—
Ingest pipeline
Est. Cost MTD
—
Ingest only · USD
Token cost estimated at $4/M blended (input + output). /recommend calls not metered here.
Readiness
Pilot readiness checklist
✓Backend: 125/125 unit tests passingDone
✓Railway production deployment liveDone
✓pgvector HNSW indexes in place (<100ms target)Done
✓Restaurant domain: 16 attributes + context weights seededDone
✓Hotel domain: 16 attributes seededDone
○POST /ingest — Traditional Chinese end-to-end validatedPending
○GET /recommend — context parameter verified (romantic, business, family, solo)Pending
○dining_occasion column persisted to DBPending
○All endpoints stress-tested (1,000 reviews, 100 concurrent recs)Pending
○Pipeline + Decoder static/dynamic toggle deployedPending
○Railway Starter → Pro upgrade before first partner goes livePending
○FunNow demo — Pilot the Pilot (friend test) completePending
POC Validation Baseline
Engine performance — validated in POC
+234%
Precision Over RFM
401 synthetic households · 366 real HKJC promos · POC v2 M4
0.92–0.95
Cross-Domain Cosine
800 synthetic reviewers · 8 archetype ground truth · POC v1
$0.00212
Per-Event Cost
7,858 events decomposed · M7-validated
10-dim
Universal Taste Vector
M6-validated luxury dimension set
Numbers above reflect POC validation on synthetic populations with real catalog data where noted. Pilot KPIs — divergence, CTR lift, rank displacement — will be measured on partner data during the pilot phase.
Timeline
Pilot programme status
Pre-pilot
Setup
API integration · review ingestion · baseline established
Week 1
Attribute Report
Per-attribute score breakdown for top 20 venues
Week 2
Rank Divergence
Star ranking vs. Kornerway ranking analysis
Week 4
Final Report
CTR lift · commercial proposal · go/no-go decision
Divergence Analysis
Taste divergence by restaurant
Divergence score — star ranking vs. taste-matched ranking
Ranking Comparison
Star ranking vs. Kornerway · sample user
Restaurant Cuisine Star Rank Kornerway Rank Movement Match Score
Mingchú 明廚 Taiwanese 7 1 ↑ +6 94%
Wisteria 紫藤廬 Tea House 12 2 ↑ +10 91%
Addiction Aquatic Seafood Market 3 3 — 0 88%
RAW Modern Taiwanese 1 6 ↓ −5 71%
Din Tai Fung 鼎泰豐 Shanghainese 2 9 ↓ −7 64%
Sample user profile: high food_quality (9.1), high cultural_authenticity (8.8), low noise_tolerance (3.2), low service_speed weight. Taste persona: quiet authentic local.
Live Trend
Weekly CTR lift — pilot data
◎
Pilot data appears here
Updated weekly once partner integration is live. CTR lift target ≥10% per LOI success gate; actual lift measured on partner A/B data.
Live API
Live recommendation test
API Usage
Call volume — month to date
Total Calls MTD
—
All endpoints
/ingest Calls
—
Review processing
/recommend Calls
—
Recommendation queries
Uptime
—
30-day rolling
Avg Response
—
/recommend p50
Error Rate
—
5xx responses
Production stats load once first partner is live. Showing pre-pilot baseline.
Commercial
Pricing tiers
Pilot
$1,000 one-time
  • 4-week structured validation
  • Attribute decomposition report
  • Rank divergence analysis
  • Final report + commercial proposal
  • $500 credited to first subscription
Starter
$500 / month
  • 50,000 API calls / month
  • 1 geography
  • Restaurant + hotel domains
  • Standard SLA (99.5%)
Growth
$1,000 / month
  • 150,000 API calls / month
  • 2 geographies
  • All domains
  • Enhanced SLA (99.9%)
  • Priority support
Enterprise
$2,500 / month
  • 500,000 API calls / month
  • Unlimited geographies
  • All domains + custom
  • 99.9% SLA + credits
  • Dedicated support
  • Geography exclusivity option
Geography add-on: +$300/month per additional country · Annual billing: ~16.7% discount
Rollout
Geography status
🇹🇼
Taiwan
Wave 1 · Target
🇯🇵
Japan
Wave 2 · 12–18 mo
🇸🇬
Singapore
Wave 3 · Phase 3
🇰🇷
Korea
Post Series A
SLA
Service level commitments
Metric Starter Growth Enterprise Current
Uptime 99.5% 99.9% 99.9% Live
/recommend response < 200ms < 150ms < 100ms —
Data deletion on exit 30 days 30 days 30 days Contractual
Exit notice 7 days 7 days 7 days LOI clause
Geography exclusivity — — 18 months max On request