What Avro types does this support?

It handles all primitive types (null, boolean, int, long, float, double, bytes, string) and the complex types record, enum, array, map, union, and fixed. Logical types attached to primitives are shown as a detail.

How are unions displayed?

A union is written in Avro as a JSON array of types, for example ["null", "string"]. The viewer renders it as a union node with one child per branch, which is the standard way to express an optional or nullable field.

What structural rules are checked?

A record must have a name and a fields array, each field needs a name and a type, an enum needs a non-empty symbols list, and a fixed type needs a numeric size. Array and map nodes must declare items and values respectively.

Does it validate data against the schema?

No. This tool validates the structure of the schema document itself, not records encoded with it. It is meant for reviewing and auditing schema definitions before registering them in a schema registry.

Is my schema sent to a server?

No. Parsing and validation run entirely in your browser in JavaScript. The schema never leaves your device, which is useful when working with internal Kafka topic definitions.

Avro Schema Viewer — Gera Tools

Apache Avro is a compact, schema-driven serialization format widely used with Kafka and the Confluent Schema Registry. An Avro schema is itself a JSON document describing records, enums, arrays, maps, unions, and fixed-size byte blobs. When schemas grow nested it becomes hard to read the raw JSON, so this viewer turns it into a collapsible tree and checks its structure.

How it works

The parser walks the schema recursively according to the Avro specification:

A string that is a primitive name (int, string, etc.) is a leaf; any other string is treated as a reference to a previously named type.
A JSON array is a union — each element is a branch, commonly used as ["null", "string"] for nullable fields.
A record must carry a name and a fields array; each field needs a name and a type, and an optional default is shown alongside it.
enum needs symbols, array needs items, map needs values, and fixed needs a numeric size.

As it walks, it collects any structural violations and reports them with a JSON-path-style location so you can jump straight to the problem.

Example

A nullable email field is written as:

{ "name": "email", "type": ["null", "string"], "default": null }

The viewer renders this as a union node with two children — null and string — and notes the default value. A record’s role field pointing at an inline enum shows its symbols joined as ADMIN | USER | GUEST.

Notes

Validity here means the schema is well-formed, not that it is compatible with an existing registered version. For producer/consumer compatibility checks (backward, forward, full) you still need the schema registry’s compatibility rules — but a clean structural pass is the prerequisite for any of that.