Apicurio Registry

A runtime server that stores, versions, and validates API definitions and event schemas so producers and consumers share one source of truth.

• Category: Messaging & Streaming
• CNCF maturity: Sandbox
• Language: Java (source 17, runtime 21), built with Quarkus and Maven
• License: Apache License 2.0
• Repository: Apicurio/apicurio-registry
• Documented at commit: 3443acd9 (main, 2026-06-25, near tag v3.3.0)

What it is

Apicurio Registry (ARG) is a server that stores artifacts and the schemas that describe them. An artifact is a versioned document such as an Avro schema, a Protobuf definition, a JSON Schema, an OpenAPI or AsyncAPI contract, a GraphQL schema, or a WSDL or XSD file. Clients push and pull these artifacts over a REST (Representational State Transfer) API and reference them at runtime.

The most common use is schema management for event streaming. A Kafka producer registers the Avro schema it writes with, the registry assigns it a global identifier, and consumers fetch that schema by identifier to deserialize messages. The registry enforces rules on every change, so an incompatible schema can be rejected before it reaches a topic. Validity, compatibility, and integrity rules can be set per artifact, per group, or globally.

It runs as a single deployable artifact whose storage backend is chosen at startup. PostgreSQL is the canonical store; Kafka, a Git repository, or Kubernetes ConfigMaps are alternatives. Apicurio also exposes a Confluent Schema Registry compatible API so existing Kafka clients can point at it without code changes.

When to use it

• You run Kafka, NATS, or Pulsar and want a self-hosted, vendor-neutral schema registry under Apache 2.0.
• You need to store more than one kind of artifact: Avro and Protobuf next to OpenAPI and AsyncAPI definitions.
• You want compatibility checks enforced server-side before a schema change ships.
• You want a Confluent-compatible API without the Confluent Community License.
• It is a weaker fit if you only need a hosted, single-cloud schema registry and do not want to operate a server; a managed offering like AWS Glue Schema Registry may be simpler.
• It is overkill if you have one Protobuf-only service and your build already pins schemas in source control.

In this deep-dive

• History: origin, milestones, and why it exists.
• Architecture: components and how requests flow.
• Adoption & Ecosystem: who runs it and what surrounds it.
• Internals: the code paths that matter, read from source.
• Getting Started: install and a first working setup.

Sources

1. Apicurio Registry README (versioning and support policy, build configuration, storage variants): https://github.com/Apicurio/apicurio-registry
2. README getting-started and Docker run instructions: https://github.com/Apicurio/apicurio-registry/blob/main/README.md
3. GitHub API repository metadata (stars, forks, contributors, created date, SPDX license): https://api.github.com/repos/Apicurio/apicurio-registry
4. CNCF Sandbox application issue #72 (Apicurio Registry, 2023-11): https://github.com/cncf/sandbox/issues/72
5. CNCF Sandbox issue #461 (ecosystem integration: Strimzi, CloudEvents, xRegistry, 2026-02): https://github.com/cncf/sandbox/issues/461
6. Apicurio Blog (Studio origin 2016, Registry 2019, Studio integration, 3.3.0 GitOps): https://www.apicur.io/blog/
7. Apicurio Registry releases (1.0.4.Final through 3.3.0): https://github.com/Apicurio/apicurio-registry/releases
8. ADOPTERS.md (Axual, Castor, IBM, Libon, Red Hat, ZenWave 360): https://github.com/Apicurio/apicurio-registry/blob/main/ADOPTERS.md