Architecture

Four layers. One platform.

A layered microservices architecture with strict tenant isolation, asynchronous event streaming, and dedicated caching layers.

Layer 1

Knowledge Graph

Attribute-based hotel data, destination intelligence, and a direct tour/package provider registry — not property codes and city search.

Layer 2

Agent Orchestration

Eleven specialized agents spanning the full travel experience — profile, personalization, planning, offers, payments and rewards, travel management, reconciliation, memory, and re-engagement — all on a standardized tool protocol (MCP).

Layer 3

Shared Core Services

Continuous reshopping, dynamic packaging, and loyalty arbitrage — the commodity primitives partners stop having to rebuild themselves.

Layer 4

Distribution & Connectivity

Multi-source aggregation normalizing NDC/GDS endpoints alongside direct hotel CRS/PMS and tour operator feeds.

Core engine capabilities

Shared once. Used by every consumer of the platform.

Capability	What it does	Implementation
Unified Traveler Profile & Vault	Aggregates identity details, passport credentials, loyalty program IDs, seat/room preferences, dietary needs, and payment instruments into an authenticated record.	Zero-knowledge encrypted data store with tenant-level RBAC. Tokenized PII and scoped authorization via OAuth2/OIDC.
Continuous Reshopping & Price Guarantee	Monitors booked airfare and lodging rates, detects price drops, verifies rebooking feasibility against penalty rules, and executes or alerts for re-ticketing.	Asynchronous worker queue executing scheduled rate evaluations against GDS cache and merchant APIs with configurable net-savings triggers.
Dynamic Packaging Engine	Bundles disparate inventory (NDC flights + boutique lodging + local excursions) into a single transaction while maintaining opaque package pricing.	Atomic cart orchestrator handling multi-provider lock-and-commit workflows, split merchant payouts, and single-point payment settlement.
Trip Planning Engine	Multi-agent system generating itineraries by clustering activities, optimizing for time/budget, and aligning with traveler "vibes."	Multi-agent orchestration layer processing constraints against the Knowledge Graph.
Recommendation Engine	Performs attribute-based semantic matching (e.g., remote work ergonomics, acoustic profiles) against inventory.	Graph-based matching engine querying Knowledge Graph embeddings.
Rewards & Card Usage Engine	Orchestration layer evaluating travel costs against points ecosystems, transfer partners, and statement credits.	Valuation engine indexing benchmarks, bonus multipliers, and redemption paths.

Let's talk

Bring the platform into your product.

Whether you're an issuer, an airline, a supplier, a distributor, or a publisher — tell us what you're building and we'll show you exactly which layer of the platform fits.

Request a demo