May 2025 Summaries

A developer can simplify their React API logic by using Connectors for REST APIs and GraphQL. By defining a GraphQL schema, they can unify data from multiple sources in a single query without custom resolvers or backend logic. The process involves creating an Apollo Studio account, installing Rover CLI, generating a command to create the graph, and configuring it with the `rover init` command. The developer then creates a file called `earthquake.graphql` to define the schema, which includes types for earthquakes, properties, and geometry. A query is defined to fetch recent earthquakes with location details using Nominatim's reverse geocoder. The GraphQL query is then integrated into the React app by replacing the old API calls with a single function call that uses the Apollo Router to fetch data from the graph. This simplifies frontend data fetching without giving up existing REST services and patterns, making it easier to maintain and scale the codebase.

Building MCP tools with GraphQL can simplify the development process and provide better tools for Large Language Model (LLM) interaction. A commonly recommended pattern is to create a simple tool for every Entity defined in an API endpoint, but this approach has limitations due to the need for additional code and deployment. GraphQL operations as MCP tools offer a better alternative by providing no writing tool code, allowing the graph to do the API orchestration, and enabling the addition of tools without code deployments. This approach enables rapid iteration, hot-reloading, and improved developer experience, making it an attractive solution for building MCP tools with GraphQL.

The Apollo MCP Server is a standardized interface for AI systems to interact with graph-based API orchestration, making it easier to integrate AI into existing infrastructure. The server is written in Rust and can be built from source or used as a Docker image. It provides a secure and scalable way to bring data and services together. The MCP Inspector tool allows developers to view and test out tools for the server. Pre-defined operation tools are generated based on the provided GraphQL API schema, but there is also an option to use introspection to create dynamic tools. The server also supports Server-Side-Events (SSE) transport. To get started with the Apollo MCP Server, users can clone the repository, build it, and run it using STDIO or SSE transport. The tool can be configured in the inspector to provide a standardized interface for AI systems like Claude and GPT to interact with the graph-based API orchestration.

The Apollo MCP Server enables AI systems like Claude and GPT to interact with organization's APIs through GraphQL, a standardized protocol that standardizes how Large Language Models (LLMs) interface with external systems. The server provides an abstraction layer between the AI and APIs, exposing a set of Model Context Protocol (MCP) tools that provide specific capabilities backed by orchestrated API calls. GraphQL is an ideal foundation for this orchestration layer due to its declarative, performant, self-documenting, standards-based, and self-service nature. The server aligns with the needs of complex AI systems, which require deterministic execution, policy enforcement, efficiency, and agility. By using GraphQL, teams can build modern API infrastructure that accommodates the pace of AI development and facilitates continuous improvement. The Apollo MCP Server represents a continued investment in making GraphQL the best way to build modern API infrastructure, providing an entry point for connecting AI to existing graph APIs.

The application architecture of online services often requires the orchestration of multiple APIs to retrieve data from various sources, such as inventory, reviews, and prices. This can lead to engineering challenges, increased costs, and slowing innovation due to cumbersome authentication and authorization credentials management. Apollo's GraphOS platform addresses these challenges by leveraging GraphQL and graph-based API orchestration, abstracting underlying data sources from clients, and providing a declarative approach to requesting data, which reduces the necessity for procedural code and optimizes API calls. By adopting this approach, developers can concentrate on solving business issues rather than generating boilerplate orchestration code, leading to significant enhancements in engineering efficiency.

Booking.com, a leading online travel company, has successfully integrated Apollo Federation into its service architecture to improve user experience and streamline API orchestration. The company's monolithic approach became increasingly complex with over 1,000 simultaneous experiments running across multiple domains, resulting in performance bottlenecks, development bottlenecks, difficulty maintaining consistent experiences, and limited ability to optimize orchestration patterns. Booking.com migrated to a distributed, federated model using Apollo Federation, decoupling subgraphs, improving performance, and enhancing developer experience through independent delivery cadence. The company implemented comprehensive monitoring and governance measures, including query-level latency monitoring, timeout budgets, schema caching, and alerting dashboards. As a result, Booking.com supports 140 service connections handling over 8 billion requests per day, with product delivery accelerated by 40% and developers enjoying improved autonomy and productivity. The company plans to further refine its use of API orchestration approach by migrating to Apollo Router, promising significant performance improvements.

PlayStation transformed its digital store experience by integrating GraphQL into its architecture, powering experiences across multiple platforms and services ecosystem. The company overhauled its digital store experience with federated GraphQL using Apollo, effectively orchestrating multiple backend APIs into a unified high-performance experience. This centralized approach to API orchestration allowed for business logic to be centralized, reducing duplicated code and data inconsistencies across devices. PlayStation adopted Apollo Client 3.7’s @defer directive to create a progressive loading experience for users, improving speed and revenue. The company emphasized the importance of cross-functional collaboration and maintaining platform-specific flexibility to balance consistency with customization. With GraphQL federation, PlayStation has built a robust foundation that accelerates performance, boosts developer efficiency, and streamlines updates.