Platform Core (Cross-Vertical Infrastructure)

Shared infrastructure supporting all verticals.

LOE: ~40-60 hours

---

Epic: Unified Data Models ✓

Plan: Define shared base types (Location, Contact) used across all verticals, enforce schema consistency, and build a versioning and migration system.

Architectural Context: Every vertical shares common entity shapes. Unified data models prevent fragmentation and enable cross-vertical search, entity resolution, and itinerary construction. Schema versioning ensures backward compatibility as the data model evolves.

Build Brief · Build Summary

Tasks

• Define shared base types (Location, Contact) used across all verticals
• Implement cross-vertical schema consistency
• Build schema versioning and migration system

---

Epic: Cross-Vertical Search

Plan: Implement a search_destination tool that queries all verticals for a given destination, with ranking, filtering, and result merging.

Architectural Context: Cross-vertical search is the primary way Research Agents find information across domains. The tool accepts a destination and returns merged results from all verticals, ranked by relevance and filtered by the agent's constraints.

Tasks

• Implement search_destination tool that searches all verticals for a destination
• Build ranking and filtering across vertical types
• Implement cross-vertical result merging

---

Epic: Entity Resolution

Plan: Build an EntityResolver that matches entities across verticals by name, website, phone, and email, then deduplicates companies appearing in multiple verticals.

Architectural Context: The same restaurant, hotel, or experience may appear in multiple verticals with slightly different data. Entity resolution merges these into a single canonical entity, preventing duplicate results and enabling cross-vertical context.

Tasks

• Build EntityResolver that matches entities across verticals by name + website + phone + email
• Implement deduplication of companies appearing in multiple verticals

---

Epic: Cross-Vertical Itinerary Construction

Plan: Implement a suggest_itinerary tool that generates multi-vertical itineraries based on destination, duration, and traveler profile.

Architectural Context: Itinerary construction combines results from multiple verticals (dining, experiences, transportation, hotels) into a cohesive trip plan. The tool ranks and recommends options based on the traveler's preferences and constraints.

Tasks

• Implement suggest_itinerary tool (destination, duration, traveler profile → multi-vertical itinerary)
• Build itinerary ranking and recommendation engine

---

Epic: Platform Infrastructure

Plan: Build the foundational infrastructure for the Intelligence platform: containerization, CI/CD, caching, authentication, and notifications.

Architectural Context: This epic establishes the operational foundation that all verticals depend on. Docker + CI/CD ensures consistent deployment. Redis caching reduces redundant API calls. JWT auth secures MCP server access. Notifications enable alerting on data quality issues.

Tasks

• Implement Docker + CI/CD pipeline
• Build monitoring and alerting
• Implement caching layer (Redis)
• Build user authentication (JWT)
• Implement notification system (email/SMS/webhooks)