Features Guide

Complete reference of all platform features and capabilities.

---

🎯 Core Features

1. Intelligent PDF Processing

14-Stage Pipeline:

• Product discovery with AI
• Focused extraction (product pages only)
• Semantic text chunking
• Multi-vector embeddings
• Image analysis with OCR
• Product creation with validation
• Metafield extraction
• Quality enhancement

Capabilities:

• ✅ Automatic product identification
• ✅ Metadata extraction (200+ types)
• ✅ Image analysis and OCR
• ✅ Checkpoint recovery on failure
• ✅ Real-time progress tracking
• ✅ Batch processing support

---

2. Multi-Modal Search

Search Types:

Status: ✅ All 6 strategies + 7-vector fusion implemented (100% complete)

1. Semantic Search ✅

• Natural language understanding with MMR diversity
• Understands intent and context
• Lambda: 0.5 (balanced relevance + diversity)
• Response time: <150ms
• Accuracy: 85%+

2. Vector Search ✅

• Pure similarity matching (no diversity)
• Fast lookup using pgvector
• Lambda: 1.0 (pure relevance)
• Response time: <100ms
• Precision: 88%+

3. Multi-Vector Search ✅

• 7-vector fusion: text (1024D, Voyage AI), visual (768D, SigLIP2), understanding (1024D, Voyage AI), color, texture, style, material (768D each, SigLIP2)
• All vectors stored as halfvec (float16) — 50% storage savings, zero accuracy loss
• Understanding embeddings enable spec-based search (dimensions, finishes, properties) from Qwen3-VL structured analysis
• Query-adaptive weight profiles: 7 profiles (product_name, color_finish, specification, texture_pattern, style_aesthetic, material_search, balanced) automatically selected based on query intent
• GPT-4o-mini parses query → selects optimal weights → adjusts fusion scoring per search
• Response time: <200ms
• Best for: Complex queries

4. Hybrid Search ✅

• Combines semantic (70%) + keyword matching (30%)
• PostgreSQL full-text search integration
• Best accuracy for technical terms
• Response time: <180ms
• Accuracy: 90%+

5. Material Search ✅

• Property-based filtering using JSONB
• Filter by material_type, dimensions, slip_resistance, etc.
• GIN index for performance
• Response time: <50ms
• Precision: 95%+

6. Image Search ✅

• Visual similarity using CLIP embeddings
• Accepts image URL or base64
• Find visually similar products
• Response time: <150ms
• Accuracy: 88%+

7. Design Inspiration URL Search ✅

• Paste any design URL (Houzz, Pinterest, Dezeen, ArchDaily, manufacturer sites)
• AI extracts design tokens: colors, materials, textures, styles, room type
• Automatically searches catalog for matching products using 7-vector fusion
• Accessible via Globe button in chat toolbar → InspirationUrlModal
• Results shown as InspirationCard (color swatches + material/style tags) + matched products
• Cost: 1 credit (Firecrawl scrape) + Haiku token cost

8. Explainable Search Spec ✅

• KAI agent provides structured interpretation of every search query
• Decomposes query across 7 dimensions: colors, materials, styles, textures, specs
• Displayed as collapsible SearchSpecCard above product results
• Shows color swatches with hex codes, material/style pill tags, spec details
• Generated by the LLM as part of the tool call — zero extra latency

7. All Strategies Combined ✅

• Runs all 6 strategies in parallel
• Merges and deduplicates results
• Relevance-based ranking
• Response time: <800ms
• Recommended for comprehensive search

Overall Performance:

• Accuracy: 85-95% (strategy-dependent)
• Recall: 90%+
• Precision: 88-95%
• Concurrent queries: 1000+/minute

---

3. Materials Catalog

Features:

• Browse all materials
• Filter by category
• Sort by relevance
• View material details
• See related materials
• Visual preview
• Specifications display

Capabilities:

• ✅ Infinite scroll
• ✅ Advanced filtering
• ✅ Sorting options
• ✅ Search integration
• ✅ Favorites/bookmarks
• ✅ Export functionality

---

4. Product Management

CRUD Operations:

• Create products
• Read product details
• Update metadata
• Delete products
• Bulk operations

Product Information:

• Name and description
• Metafields (200+ types)
• Associated chunks
• Associated images
• Quality scores
• Confidence metrics

---

5. Admin Dashboard

Monitoring:

• Real-time job tracking
• Progress visualization
• Error monitoring
• Performance metrics
• System health

Management:

• User management
• Workspace management
• Document management
• Product management
• Metafield management
• Agent configurations (AI prompts)

Analytics:

• Search analytics
• Processing statistics
• AI model usage
• Cost tracking
• Performance trends

Features:

• ✅ Live progress updates
• ✅ Job history
• ✅ Error logs
• ✅ Performance charts
• ✅ Export reports

---

6. RAG System (Retrieval-Augmented Generation)

Capabilities:

• Document upload and indexing
• Semantic search
• Chat interface
• Multi-document retrieval
• Context-aware responses

Features:

• ✅ Document management
• ✅ Query interface
• ✅ Chat history
• ✅ Source attribution
• ✅ Relevance scoring

---

7. Real-Time Monitoring

Progress Tracking:

• Live progress percentage
• Current stage display
• Estimated time remaining
• Detailed metrics
• Error notifications

Features:

• ✅ Server-Sent Events (SSE)
• ✅ WebSocket support
• ✅ Real-time updates
• ✅ Progress persistence
• ✅ Checkpoint tracking

---

8. Metadata Management

Metafield Types (200+):

• Material composition
• Dimensions (length, width, height)
• Weight and density
• Color and finish
• Texture and pattern
• Application and use case
• Care instructions
• Certifications
• Pricing and availability
• Supplier information

Features:

• ✅ Dynamic metafield creation
• ✅ Type validation
• ✅ Multi-value support
• ✅ Relationship linking
• ✅ Search integration

---

9. Image Management

Capabilities:

• Automatic image extraction
• OCR on material specs
• Quality scoring
• Visual embeddings
• Storage management

Features:

• ✅ Batch processing
• ✅ Quality analysis
• ✅ Metadata extraction
• ✅ Visual search
• ✅ Storage optimization

---

10. Workspace Isolation

Multi-Tenancy:

• User workspaces
• Data isolation
• Row-Level Security (RLS)
• Separate indexes
• Independent quotas

Features:

• ✅ Workspace creation
• ✅ User management
• ✅ Permission control
• ✅ Data privacy
• ✅ Audit logging

---

11. Batch Processing

Capabilities:

• Batch PDF upload
• Batch image analysis
• Batch product creation
• Batch embeddings
• Batch search

Features:

• ✅ Async processing
• ✅ Progress tracking
• ✅ Error handling
• ✅ Result aggregation
• ✅ Performance optimization

---

12. API Gateway

Features:

• Unified API interface
• Request routing
• Response formatting
• Error handling
• Rate limiting
• Authentication

Endpoints:

• 74+ REST endpoints
• 9 categories
• Comprehensive documentation
• OpenAPI schema

---

🔐 Security Features

✅ Authentication:

• Supabase JWT
• MIVAA JWT
• API keys

✅ Authorization:

• Row-Level Security (RLS)
• Role-based access control
• Workspace isolation

✅ Data Protection:

• HTTPS/TLS encryption
• Database encryption
• Secure storage
• Audit logging

✅ API Security:

• Rate limiting
• Input validation
• SQL injection prevention
• CORS configuration

---

📊 Analytics & Reporting

Available Metrics:

• Search analytics
• Processing statistics
• AI model usage
• Cost tracking
• Performance trends
• Error rates
• User activity

Export Options:

• CSV export
• JSON export
• PDF reports
• Custom queries

---

🚀 Performance Features

Optimization:

• ✅ Query caching
• ✅ Result caching
• ✅ Batch processing
• ✅ Lazy loading
• ✅ Pagination
• ✅ Compression

Scalability:

• ✅ Horizontal scaling
• ✅ Load balancing
• ✅ Connection pooling
• ✅ Index optimization
• ✅ Query optimization

---

🔄 Integration Features

External Services:

• Anthropic (Claude Sonnet/Haiku 4.5 + built-in web search for B2B)
• OpenAI (GPT-4o, GPT-4o-mini for query parsing)
• Voyage AI (voyage-3.5 text + understanding embeddings)
• HuggingFace Endpoint (Qwen3-VL 32B vision analysis)
• SigLIP2 via HuggingFace Endpoint (5 visual embedding types, 768D each)
• Replicate (14 interior design generation models)
• WorldLabs Marble (3D Gaussian Splat VR world generation)
• Supabase (Database, auth, storage, edge functions)
• Direct Vector DB RAG (Claude 4.5)

Capabilities:

• ✅ API integration
• ✅ Webhook support
• ✅ Data sync
• ✅ Error handling
• ✅ Retry logic

---

15. Duplicate Detection & Merging

Purpose: Maintain data quality by identifying and consolidating duplicate products

Key Features:

• ✅ Factory-based duplicate detection (CRITICAL: same factory only)
• ✅ Multi-metric similarity scoring (name, description, metadata)
• ✅ Confidence levels (high, medium, low)
• ✅ Product merging with data consolidation
• ✅ Merge history tracking with full audit trail
• ✅ Undo capability with snapshot restoration
• ✅ Batch duplicate scanning
• ✅ Caching for performance

Detection Rules:

• MUST have same factory/manufacturer (primary filter)
• Visual similarity alone does NOT create duplicates
• Different factories = NOT duplicates (even if identical)

Similarity Scoring:

• Name similarity: 50% weight
• Description similarity: 30% weight
• Metadata similarity: 20% weight
• Overall score: 0.0-1.0

Confidence Levels:

• High: 85%+ (very likely duplicate)
• Medium: 70-85% (possible duplicate)
• Low: 55-70% (review needed)

API Endpoints (7 total):

• POST /api/duplicates/detect - Detect for single product
• POST /api/duplicates/batch-detect - Scan workspace
• GET /api/duplicates/cached - Get cached results
• POST /api/duplicates/update-status - Update status
• POST /api/duplicates/merge - Merge products
• POST /api/duplicates/undo-merge - Undo merge
• GET /api/duplicates/merge-history - Get history

Database Tables:

• product_merge_history - Merge operations with snapshots
• duplicate_detection_cache - Pre-computed duplicate pairs

---

16. Quotes System

Purpose: Comprehensive quote management system for customers and admins

Status: ✅ 100% Complete (Production Ready)

Customer Features:

• ✅ Create multiple independent quotes
• ✅ Add products from search, agents, moodboards
• ✅ Custom text-based requests (no products required)
• ✅ Dimensions and area tracking (width, height, sqm)
• ✅ Auto-expiration after 30 days (configurable)
• ✅ Activity tracking extends expiration
• ✅ Submit quote requests
• ✅ View attached extras/upsells
• ✅ Accept/reject individual extras
• ✅ Accept quotes with validation
• ✅ View project timeline (when accepted)

Admin Features:

• ✅ View all quote requests
• ✅ Filter by custom status tags
• ✅ Full-page quote detail view with tabs
• ✅ Assign custom status tags with colors
• ✅ Attach upsells/extras to quotes
• ✅ Monitor customer acceptance of extras
• ✅ Update project timeline progress
• ✅ Add notes to timeline steps

Management Pages:

• ✅ Status Tags Management (/admin/status-tags)

  • Create/edit/delete custom status tags
  • Color picker for tag colors
  • System vs custom tag distinction
  • 6 default system tags (pending, in_progress, quoted, accepted, rejected, expired)

• ✅ Upsells Management (/admin/upsells)

  • Create/edit/delete upsell items
  • Set pricing and descriptions
  • Active/inactive status toggle
  • Display order management

• ✅ Timeline Steps Management (/admin/timeline-steps)

  • Create/edit/delete timeline steps
  • 9 default steps (Quote Accepted → Project Completed)
  • Display order management
  • Active/inactive status

Project Timeline:

• ✅ Auto-initialization when quote accepted
• ✅ 9 predefined steps with descriptions
• ✅ Visual timeline tree with connector lines
• ✅ Status tracking (pending, in_progress, completed, skipped)
• ✅ Admin can update status and add notes
• ✅ Customer read-only view
• ✅ Completed timestamps
• ✅ Color-coded status badges

Quote Acceptance Workflow:

• ✅ Validation requiring all extras to be decided
• ✅ Status change to 'accepted'
• ✅ Automatic timeline initialization
• ✅ Timeline button visibility control

Database Tables (8 total):

• quotes - Core quote data
• quote_items - Quote items with dimensions
• status_tags - Custom status tags
• upsells - Admin-managed upsell items
• quote_upsells - Junction table with acceptance tracking
• timeline_steps - Predefined timeline steps
• quote_timeline - Timeline progress tracking
• system_settings - Platform configuration

API Endpoints: 30+ methods via QuotesService

• Core quote operations (CRUD)
• Status tags management
• Upsells management
• Timeline management
• Expiration tracking

---

17. VR World Generation

Purpose: Generate explorable 3D worlds from interior design images using WorldLabs Marble API

Status: ✅ Complete (Production Ready)

Key Features:

• ✅ One-click "Generate VR" on any interior design image
• ✅ 3D Gaussian Splat rendering via Spark.js
• ✅ Orbit controls (drag to rotate, scroll to zoom)
• ✅ First-person navigation (WASD + mouse look + Shift for speed)
• ✅ Navigation mode toggle (Orbit / Walk)
• ✅ 3 quality levels (Draft 100k, Standard 500k, Full)
• ✅ Fullscreen mode
• ✅ Adaptive status polling with loading animations
• ✅ Credit-based pricing with refund on failure
• ✅ Persists across chat sessions

Models:

• marble-0.1-mini: 50 credits, ~30-45 seconds
• marble-0.1-plus: 200 credits, ~5 minutes

Output Assets:

• SPZ files (3 quality levels — Gaussian Splats)
• Collider GLB (collision mesh for navigation)
• Panorama image (360 render)
• Thumbnail (preview)
• AI-generated caption

Components:

• Edge Function: generate-vr-world
• Frontend: WorldViewer.tsx (Spark.js renderer)
• Service: vrWorldService.ts
• Database: vr_worlds table

Documentation: vr-world-generation.md

---

📱 User Interface Features

Frontend Components:

• Materials catalog
• Search hub
• Admin dashboard
• Real-time monitoring
• Progress visualization
• Error handling
• Responsive design
• Quote management system
• Project timeline viewer

User Experience:

• ✅ Intuitive navigation
• ✅ Fast loading
• ✅ Real-time updates
• ✅ Error messages
• ✅ Help documentation
• ✅ Accessibility
• ✅ Tab-based interfaces
• ✅ Modal workflows
• ✅ Visual feedback

---

---

18. Flow Engine — Workflow Automation

Purpose: Visual drag-and-drop automation builder for multi-step workflows

Status: ✅ Complete (Production Ready)

Documentation: flow-engine.md

• ✅ xyflow graph-based visual editor at /admin/flows
• ✅ 3 node types: trigger, condition, action
• ✅ Trigger types: manual, cron schedule, webhook, platform event
• ✅ Condition types: if_else, switch, filter, delay
• ✅ Action types: send_sms, send_email, http_request, create_notification, send_quote
• ✅ {{template}} variable resolution from execution context
• ✅ Dry-run / test mode (no actions fire)
• ✅ Scheduled execution via flow-scheduler-cron (pg_cron, every minute)
• ✅ External webhook triggers via flow-webhook
• ✅ Execution logs in flow_runs table

---

19. Background Agents

Purpose: Autonomous scheduled or event-driven agents running workspace tasks

Status: ✅ Complete (Production Ready)

Documentation: background-agents.md

Agent Types:

• ✅ kai-task — Runs arbitrary KAI tool calls autonomously
• ✅ product-enrichment — Fills missing product metadata via Claude + web search
• ✅ material-tagger — Auto-classifies products with material tags
• ✅ social-analytics-sync — Syncs engagement metrics from Late.dev
• ✅ social-insights-sync — Syncs account-level insights from Late.dev
• ✅ factory-enrichment — Enriches factory records via Apollo/Firecrawl/Hunter.io

Infrastructure:

• ✅ background-agent-runner edge function (universal executor)
• ✅ agent-scheduler-cron (pg_cron every minute)
• ✅ auto-recovery-cron (stuck run detection every 5 minutes, up to 3 retries)
• ✅ Chain triggers (child agents auto-dispatch on parent completion)
• ✅ Delegation to Python backend for tasks >25s (DelegateToMivaaError)
• ✅ Real-time logs via agent_run_logs Supabase subscription
• ✅ Admin UI at /admin/background-agents

---

20. Interior Video Generation

Purpose: AI-generated interior design videos for walkthroughs, product spotlights, and social content

Status: ✅ Complete (Production Ready)

Documentation: interior-video-generation.md

Model	Credits	Best For
Veo-2 (Google)	30	Cinematic walkthroughs, floorplan flythroughs
Kling v3.0	20	Product spotlights, before/after, social reels
Wan 2.1 i2v 720p	12	Budget general-purpose
Runway Gen4 Turbo	40	Premium quality

• ✅ Auto model selection by video_type
• ✅ Async polling pattern for long renders (>55s → returns job_id)
• ✅ Aspect ratios: 16:9, 9:16, 1:1
• ✅ Separate generate-social-video wrapper for social-specific flow

---

21. Virtual Staging

Purpose: Transform empty room photos into furnished interior renders

Status: ✅ Complete (Production Ready)

Documentation: virtual-staging.md

• ✅ Replicate proplabs/virtual-staging model (~56s)
• ✅ 20 credits per generation
• ✅ Room types: living_room, bedroom, dining_room, kitchen, bathroom, home_office, kids_room, outdoor
• ✅ Styles: modern, scandinavian, industrial, traditional, bohemian, minimalist, luxury, rustic
• ✅ Optional free-text furniture_items override
• ✅ Accessible as KAI agent tool

---

22. Region Editing (Inpainting)

Purpose: Paint a zone on a room image and replace it with AI-generated content

Status: ✅ Complete (Production Ready)

Documentation: segmentation-inpainting.md

• ✅ RegionEditCanvas.tsx — brush paint / erase mask tool
• ✅ generate-region-edit edge function — Grok Aurora masked inpainting (20cr)
• ✅ SAM 2 via Replicate for pixel-perfect mask generation (/api/segment/mask)
• ✅ Pillow bbox/ellipse fallback (instant, zero cost)
• ✅ AnyDoor for reference-image product placement
• ✅ FLUX Fill Pro for text-guided replacement

---

23. Gemini Interior Generation

Purpose: Gemini-powered interior image generation with four distinct modes

Status: ✅ Complete (Production Ready)

API Docs: api/generate-interior-gemini-api.md

Mode	Description	Credits
text-to-image	Prompt → new room render	6/15
image-edit	Image + instruction → edited image	6/15
floor-plan-render	Floor plan → photorealistic perspective	6/15
floor-plan-text	Text description → 2D floor plan	6/15

• ✅ Two-step style-transfer pipeline (inspiration → design spec → edit)
• ✅ Models: gemini-3.1-flash-image-preview (6cr) and gemini-3-pro-image-preview (15cr)

---

24. Social Media Suite

Purpose: Generate, schedule, and publish social media content across 8 platforms via Late.dev

Status: ✅ Complete (Production Ready)

Documentation: social-media-system.md | API: api/late-social-api.md

Content Generation:

• ✅ Caption generation (3 variants) for Instagram, Facebook, LinkedIn, TikTok, Pinterest, YouTube, Twitter, Threads — 2 credits
• ✅ Image generation: xAI Aurora (10cr), Gemini Imagen (5cr), FLUX Dev (6cr) — auto-routed by image type
• ✅ Video generation: Kling 1.6 Pro (15cr), Veo-2 (30cr)

Publishing (Late.dev):

• ✅ OAuth account connect/disconnect for all 8 platforms
• ✅ Publish immediately or schedule for a future datetime
• ✅ No credit cost — uses workspace's Late.dev subscription
• ✅ Analytics sync: post engagement + account insights
• ✅ Best posting time recommendations

Admin UI: /admin/social-media/accounts

---

25. Billing & Credits System

Purpose: Stripe-powered subscription plans and credit purchases

Status: ✅ Complete (Production Ready)

Documentation: billing-credits-system.md | API: api/crm-stripe-api.md

• ✅ Subscription plans with monthly/annual billing (subscription_plans table)
• ✅ Credit packages for top-up purchases (credit_packages table)
• ✅ Stripe Checkout sessions for both subscriptions and credit purchases
• ✅ Stripe Customer Portal for self-service plan/payment changes
• ✅ Webhook processing for payment events (stripe-webhooks)
• ✅ Per-workspace credit balance tracking
• ✅ Upfront credit deduction before AI calls
• ✅ Frontend: /billing/credits and /billing/subscriptions

---

26. CRM System

Purpose: Customer relationship management — contacts, companies, platform users

Status: ✅ Complete (Production Ready)

Documentation: crm-system.md

• ✅ Contact management: CRUD, tagging, notes, activity log
• ✅ Company management: CRUD, contact associations
• ✅ User management: role changes, workspace membership
• ✅ Contact ↔ platform user linking (linked_user_id)
• ✅ Integration with messaging campaigns
• ✅ Integration with quotes system
• ✅ Role-based access: admin/owner = full CRUD, manager = read+write, factory = read own
• ✅ Admin UI: /admin/crm, /admin/crm/companies/:id, /admin/crm/contacts/:id

---

27. AI Pricing Auto-Updater

Purpose: Keep AI model pricing data current with automated weekly syncs

Status: ✅ Complete (Production Ready)

• ✅ Weekly cron via ai-pricing-updater edge function
• ✅ Per-model auto_update_enabled flag
• ✅ Fallback prices embedded for Anthropic, OpenAI, Voyage AI
• ✅ Full audit log of all price changes
• ✅ Manual trigger via POST request

---

---

28. AR Material Preview

Purpose: View materials applied to real-world surfaces using AR, or preview them as 3D textured swatches with interactive rotation

Status: ✅ Complete (Production Ready)

Documentation: ar-material-preview.md

Key Features:

• ✅ WebXR AR on Android Chrome (camera-based surface detection)
• ✅ 3D swatch viewer fallback on iOS and desktop
• ✅ QR code handoff from desktop to mobile (/ar/:productId)
• ✅ PBR map generation via Replicate (albedo, normal, roughness, metalness)
• ✅ Device capability detection via useARSupport() hook
• ✅ Credit cost: 8 credits for PBR generation, viewing is free

Components: src/components/features/ar/ — ARPreviewModal, ViewInARButton, ARPage, useARSupport Edge Function: generate-pbr-maps/index.ts

---

29. Lighting Simulation

Purpose: Preview how materials look under different lighting conditions with real-time 3D rendering and AI-powered room re-lighting

Status: ✅ Complete (Production Ready)

Documentation: lighting-simulation.md

Layer 1 (AI Room Re-lighting):

• ✅ "Lighting Variants" dropdown on ProgressiveImageGrid
• ✅ 6 presets: Golden Hour, Bright Midday, Soft Overcast, Warm Evening, Night, Dramatic Spots
• ✅ Gemini AI edit with lighting-specific prompt + spatial lock clause
• ✅ 6 credits per variant

Layer 2 (3D PBR Viewer):

• ✅ React Three Fiber scene with PBR material rendering
• ✅ Time-of-day slider (6AM - 9PM) with play/pause animation
• ✅ Room orientation (N/E/S/W compass buttons)
• ✅ Surface type selection (Wall, Floor, Column, Curved)
• ✅ Material tuning sliders (roughness, clearcoat)
• ✅ Screenshot + download

Components: src/components/features/lighting/ — MaterialLightingViewer, LightingPreviewModal, lightingPresets, useSunPosition

---

30. Pinterest Integration

Purpose: Import Pinterest pins into moodboards with automatic material matching from the catalog

Status: ✅ Complete (Production Ready)

Documentation: pinterest-integration.md

Phase 1 (No OAuth):

• ✅ Paste pin URL → oEmbed extraction → import to moodboard
• ✅ Auto-matching via MIVAA visual search (7-vector fusion)
• ✅ No API key required

Phase 2 (OAuth):

• ✅ Connect Pinterest account → browse boards → select pins → bulk import
• ✅ OAuth tokens stored in social_accounts table
• ✅ Auto-refresh on token expiry (30-day access, 365-day refresh)

Components: PinterestImportModal on MoodBoardDetailPage header Edge Functions: pinterest-import, pinterest-oauth Service: src/services/pinterestService.ts

---

31. Manufacturer Analytics (Enhanced)

Purpose: Comprehensive product interaction tracking and analytics dashboard for manufacturers/factories

Status: ✅ Complete (Production Ready)

Documentation: manufacturer-analytics.md

Event Tracking:

• ✅ Batched fire-and-forget tracking (flush every 5s or 20 events)
• ✅ 6 event types: product_view, product_save, product_quote, product_search_impression, product_search_click, product_compare
• ✅ IntersectionObserver for view tracking (50% visibility threshold)

Dashboard Sections:

• ✅ KPI row (8 metrics), Radar chart, Rating distribution
• ✅ Activity over time (8 weeks), Top products
• ✅ Quote pipeline, Conversion funnel
• ✅ Geographic Demand (Pro tier)
• ✅ Designer Engagement by Profession (Pro tier)
• ✅ Competitive Positioning (Pro tier)

Tiered Access: free | pro | enterprise — Geographic, designer, and competitive sections gated behind Pro

Service: src/services/manufacturerAnalyticsService.ts Dashboard: src/components/analytics/MyFactoryTab/MyFactoryTab.tsx

---

Last Updated: April 2026 Version: 3.6.0 Status: Production