Skip to main content
Every endpoint page in the API reference ships its samples in the same shape: a cURL command you can run from any shell beside the same request in the SDK languages, grouped in one block so you can switch between them without losing your place. This page explains that shape — the two languages every sample guarantees, the six languages the headline operations carry, and the rules the samples follow. For the page-by-page list of which treatment each endpoint gets — full six-language groups versus the trimmed pair — see Which languages each endpoint page ships.

The two-language floor

Every request sample contains cURL and Node.js (TypeScript) at minimum. Those are the two most-used forms measured by usage, and they are promised on every sample: if a snippet exists at all, those two are in it.

How the samples are written

Each language form is deliberately dependency-free, so it runs even when the SDK for that language has not published yet:
  • cURL — a complete command: method, URL, headers, and a real request body with example values.
  • Node.js (TypeScript) — the typed SDK helper where one exists.
  • Python — plain requests, so the snippet works whether or not you have installed the SDK.
  • Go, Ruby, PHP — plain standard-library HTTP (net/http, Net::HTTP, the curl extension).
When a route has no typed helper in the published SDKs, samples use the untyped request method on the client, and each language page walks that form end to end: Node.js, Python, Go (client.Request), Ruby (client.request), PHP ($client->request), plus Java and C# (client.request / client.RequestAsync).

How long pages keep samples readable

Endpoint pages with more than 15 operations trim the tail: the first 15 operations show the full six-language group — cURL, Node.js (TypeScript), Python, Go, Ruby, and PHP, always in that order — and everything after the fifteenth shows cURL and TypeScript. Hero flows that sit past the cut still get the full-group treatment inside a Worked sequences section lower on the page. A page with 15 or fewer operations never trims: every operation shows all six languages.

What’s in a response sample

Response samples carry the complete envelope, so what you see on the page is what your code receives:
  • the data payload for the success path, and
  • the meta block with request_id and timestamp.
Where an operation documents an error shape, the sample includes the full error envelope under a labelled status tab (for example 422 Error). A 204 No Content operation says so inline instead of fencing an empty body.