Data Format

Data Format is the convention a system uses to encode, structure, and serialize information so it can be stored, transmitted, and — most importantly — reliably parsed by another system on the other end. JSON, XML, YAML, CSV, and Protocol Buffers are all data formats, and the choice between them quietly shapes how easy an API is to consume, how it versions, and how well it survives being read by both humans and machines.

• Encoding and serialization - The rules for turning in-memory structures into bytes and back again without ambiguity.
• Human- versus machine-readability - JSON and YAML lean human-friendly; Protocol Buffers and Avro trade readability for compact, fast wire formats.
• Media types - Formats are negotiated over HTTP through registered media types like application/json, letting a client and server agree on representation.
• Schema pairing - A format only describes syntax; a schema layered on top (JSON Schema, XSD) constrains what valid content looks like.

In everyday API operations, the data format is the contract’s surface: it determines what an OpenAPI definition serializes to, how payloads are validated, and how content negotiation behaves across clients. Most modern APIs standardize on JSON while keeping YAML for the definitions themselves, and the same conventions carry into agent-facing tooling, where a predictable, well-documented format is what lets an automated client parse a response without guessing. It sits directly alongside data models and data modeling — the format is the encoding, the model is the meaning.