A weekly newsletter of the best news, articles and projects about GraphQL

Reverb, the world's largest online marketplace for musical instruments, migrated a decade-old GraphQL stack to a federated supergraph — query by query, with zero client disruption — using Hive Gateway and The Guild's open source tooling.

A reusable framework for sound and precise GraphQL query analysis, with an IBM cost model, optimality proofs, and an open-source Rust implementation.

Exposing GraphQL APIs as tools AI agents can call directly earned Apollo MCP Server a 2026 API Award for Best Innovation in AI API. See why.

Why a GraphQL mock that returns the same fixed body for every request to /graphql tells you almost nothing — with GraphQL, it's the client's query, not the URL, that determines what a real server sends back. Covers what mocking GraphQL properly actually requires.

Data is the foundation of AI. It acts as the expe­riential element, providing AI systems with the nec­essary information to understand patterns, make decisions, and predict outcomes. Apollo GraphQL was also named on that list!

We questioned why conventional GraphQL execution incurs hidden costs, and rewrote it in a faster breadth-first manner to avoid them.

You want to call a GraphQL API from your Falcon Foundry app, and you’ve just discovered that API integrations only accept OpenAPI specs. GraphQL endpoints don’t have OpenAPI specs. So what do you do?

American Airlines built its API orchestration layer before AI agents existed, then used it as the agent layer. See why using GraphQL is the real head start.

I have spent a lot of years being ambivalent about GraphQL. I have watched it get oversold as a REST-killer and I have watched people pile onto it for reason...

Escape's unique, in-house dynamic security scanner is native to GraphQL and fully embraces its recursive nature, which is exactly what Sigma was missing.

Using GraphQL for service-to-service communication is not necessarily an anti-pattern. In many cases, it's actually turns out to be a great choice!

If you’ve been around GraphQL long enough, you remember the early federation era: a little bit of schema stitching, a little bit of magic, and a lot of “why is this field resolving from three different places?” In 2026, federation looks very different. It’s less of a clever integration trick and more of a platform discipline.

Buffer used to have a public REST API, but closed it back in 2019, and has since been deep in migrating everything internally to GraphQL!

This is a classic retrieval problem, and the usual answer is RAG: embed every coordinate or type, as well as the user’s question, and pull the top matches. It works well for documentation. It works less well for GraphQL, and the reason is specific to how schemas are designed.

Igor Kusakov shared the Apollo Tooling -> GraphQL Codegen migration journey at Yelp

We decided to build Buffer's API with GraphQL, and today it's live and ready for you to start building with. Everything is driven by a strongly-typed GraphQL schema. Every field has a defined type and shape, so you always know what you're sending and what you'll get back. Your editor can autocomplete queries and catch mistakes before you run them. And you can ask for exactly the fields you need rather than parsing through payloads you didn't want.

GAPs is the new home for community-written specifications and extensions to GraphQL.

A slightly different resolver paradigm can have an outsized impact

Hot Chocolate and Fusion now ship an MCP adapter. Add two lines, author tools and prompts on disk, publish them with Nitro, and connect any MCP host to your GraphQL API.

GraphQL's schema and flexible query style make it an interesting candidate for agentic AI systems — but is it actually the right choice? Let's dig in.

The agentic age of software brings new challenges for our APIs. Semantic Introspection makes GraphQL discoverable, scalable, and precise for LLMs.

Explore how Shopify rebuilt GraphQL execution from scratch. Cardinal's breadth-first engine delivers 4s latency savings and what it means for Ruby developers.

We fact-checked 18 claims comparing GraphQL and REST against peer-reviewed research, official docs, and security reports. Only 3 held up. Here's what the evidence actually says.

GOV.UK are exploring how the APIs that deliver content for GOV.UK could evolve to better support an experience spanning multiple digital channels. Find out what we’re doing and how GraphQL being explored!

Apollo Skills cut AI agent token usage and completion time down significantly — same result, far less cost. Here's what the experiment revealed.

From the AWS blog: This post shows you how to implement AWS Lambda authorizers with Direct Lambda Resolvers in .NET.

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 to query exactly the data they need without multiple round trips. Trivago's use case and perspective were different.

Event Sourcing is emerging as a preferred approach to building complex software systems. It enables rapid development, smooth evolution over time, and opens the door to AI-based acceleration.

GraphQL adoption within enterprise environments follows a predictable pattern. A single team implements it successfully, word spreads and suddenly dozens of services want to expose data through your graph / Ramesh Karsala

Hello there, this is a nice content that could be interesting for folks running GraphQL in JVM applications.

April Fools' Special: A look at the most controversial (and entirely fictional) bug fix of the year.

Hemnet is Sweden’s largest property platform, serving millions of users who browse, save, and search through real estate listings every day. This case study shares their migration story from Apollo Router (Rust) to Hive Gateway (JS).

Boilerplate became cheap, but GraphQL still wins when a durable schema contract matters across teams and clients.

Your GraphQL API already serves your frontend developers well. But increasingly, it needs to serve another kind of client: AI agents that can discover, understand, and interact with your data automatically, retrieving information, performing actions on behalf of users, and chaining multi-step workflows without human intervention.