$
npx @buildinternet/releases get apollo-graphqlRecently Shipped24 releases · updated Oct 12, 2026
Summit week delivered a cluster of production features, agent-infrastructure launches, and a stable Rover CLI.
Rover 1.0 replaces static API keys with OAuth. The 1.0 release rebuilds auth, configuration, extensibility, and checks/publish around how teams run Rover today — in local development, CI, and agentic tooling. The rover subgraph check JSON output also gains per-contract-variant detail, with json_version moving to "3".
GraphOS Router 3.0 starts up without the management plane. Released in Preview at Summit, 3.0 packages persisted queries into Graph Artifacts so routers can start without reaching Apollo's management plane. A new query planner with bounded memory usage replaces the existing algorithm. Native circuit breaking and a rewritten request pipeline ship alongside it.
Apollo Connectors v0.4 is generally available. Teams can now express abstract types (interfaces and unions), inline -> transforms, and connectors that return data without an HTTP request — all declared directly in schema with no resolver code or subgraph server.
GraphOS Agent Services governs API access for agents. Announced at Summit, it addresses MCP credential sprawl: a central layer manages tool definitions, credentials, and what data returns to agent context — limiting field exposure without modifying upstream APIs.
Apollo Client 4.3 ships custom scalar support. The most-requested feature lands in 4.3 alongside type-safe cache access and improved TypeScript ergonomics for incremental data. Patch releases through 4.3.3 follow, including a deferred-fragment memory leak fix.
Apollo Kotlin 5.2.0 adds the HTTP QUERY method and empty selection sets, rounding out the window's mobile-SDK updates.
Patch Changes
There’s a clear chain of logic that most companies adopting agents discover through trial and error:
Creating your first agent was easy. Someone pointed an MCP client at a few internal services, wrote some skills by hand, and in a couple of days it was summarizing campaign performance emails or triaging tickets. Leadership really liked your demo.
Apollo GraphQL Blog · v3.0
GraphOS Router runs your graph, bringing data from its services together into an API your apps can use. With GraphOS Router 3.0, launching today in Preview, we’re putting…
GraphOS Factory is an agent skill that turns an existing REST API into an Apollo Connectors subgraph. Point it at an OpenAPI or Swagger spec, or at the API’s public docs, and it inventories the API, then works with you through the design questions: which operations and fields…
Apollo GraphQL Blog · v0.4
Apollo Connectors let subgraph schemas turn REST or HTTP APIs into typed GraphQL with no resolver code and no subgraph server to run.
Apollo GraphQL Blog · v1.0
Rover has been the command-line entry point to the Apollo ecosystem since the earliest days of Apollo Federation. It has been the tool to publish subgraphs, run schema checks, compose supergraphs, and generally act as the bridge between your terminal and GraphOS. Today we’re…
In part one we introduced the Runtime Testing Framework (RTF), which we use to deploy, test and measure services in production-like environments.
Today we’re introducing the GraphOS Router Support Tool, a new way to package up a diagnostic snapshot of your GraphOS Router and its Kubernetes environment, including router version and config, logs, pod health, and…
“Testing distributed systems is easy”, said no software engineer ever.
Today we are launching a new query planning algorithm which is built to scale to the most complex graphs, as part of GraphOS Router 3.0. The existing planning algorithm has brought us a long way, and two years…
Important: 3 potentially breaking changes below, indicated by ❗ BREAKING ❗
Patch Changes
This release was automatically created by GitHub Actions.
🚀 Features
Building or testing a GraphQL client usually means dealing with data you don’t control yet: a backend that isn’t finished, a staging environment that’s flaky or rate-limited, or a dataset that simply doesn’t have the edge cases you need for a given test.