Most GraphQL Gateway and general GraphQL discussions focus on customer-facing applications—aggregating microservices behind a single endpoint for web and mobile clients, enabling frontend teams ...

Trivago's platform is a central hub for technology enthusiasts and travel industry professionals, offering insights into travel technology, data analytics, and online booking platforms. Through articles, case studies, and tech talks, Trivago offers insights into building scalable and personalized travel search engines. Developers can learn about data science techniques, recommendation algorithms, and real-time data processing to improve user experience in the travel industry.

Tech at trivago

Trivago built an internal GraphQL gateway to unify 15+ fragmented internal APIs—REST endpoints, databases, gRPC services—into a single queryable interface powering their admin tooling. Using GraphQL Mesh for schema stitching and transforms, the gateway centralizes authentication, authorization, caching, and audit logging without requiring upstream services to change. Six years in, the team shares key lessons: the gateway became a mirror revealing hidden data relationships and organizational gaps, a flat graph works better than deeply merged entities for internal use, and the lack of a schema registry from day one caused ongoing pain. The post also explores how the unified, well-typed, access-controlled graph could serve as a foundation for AI agents using protocols like MCP.

Unifying Internal APIs: A Different Use Case for GraphQL ...

The Solution: GraphQL and Gateway Architecture

What’s Next: A Foundation for AI Agents?