A live JSON endpoint in four seconds.
Paste a document, press publish, get an HTTPS URL that returns it with correct headers and open CORS. Perfect for a mock API, a prototype's fixtures, or handing a payload to a colleague without setting anything up.
Loading the editor…
Three steps, no setup
The editor validates as you type and tells you the line and column of any syntax error. The panel beside it renders a live, collapsible tree of what you actually have.
An hour for a quick test, three days without an account, six if you sign in. The clock starts when you publish and the endpoint is removed when it runs out.
You get a URL plus ready-made curl, JavaScript and Python snippets. CORS is open, so a browser app can call it directly with no proxy.
Good uses, and one bad one
Mock an API you do not have yet
Front-end work rarely waits for the backend. Publish the response shape you agreed on and build against a real URL instead of a hard-coded fixture.
Share a payload
Faster than pasting 400 lines into a chat window, and the recipient can curl it. The link expires on its own, so nobody has to remember to clean up.
Demo, workshop or test fixture
Feed a live endpoint to a tutorial, a CodePen, an integration test or a no-code tool that insists on a URL rather than a file.
Feature flags for a prototype
A tiny config document your prototype polls. Change it with a PUT, no redeploy.
Webhook payload samples
Keep a known-good example of a provider's payload at a stable URL while you build the handler.
Not: anything private
Everything published here is readable by anyone with the URL. No credentials, no personal data, no production secrets. It is a scratchpad, and it should be treated as one.
Questions people actually ask
No. Paste JSON, press publish, and you have a URL. An account is optional and buys you two things: endpoints that live six days instead of three, and a list of everything you have created that follows you between devices.
Because GetJSON is a scratchpad, not hosting. Short, enforced lifetimes keep the service cheap to run, keep abandoned data from piling up, and set the right expectation — this is for a prototype, a mock API, a demo or a quick handover, not for production. The ceiling is enforced by a database trigger, so no API call can exceed it.
A plain HTTPS URL that returns your JSON with a proper application/json content type and permissive CORS headers. No API key, no wrapper object, no envelope — a GET returns exactly the document you saved. It works from curl, from a browser, from a fetch in a sandboxed frontend, from anywhere.
Yes. Publishing anonymously hands you an edit token once — send it as an X-Edit-Token header on a PUT or DELETE. Publishing while signed in ties the endpoint to your account instead, and you can edit or delete it from the dashboard without keeping a token anywhere.
No — treat everything you publish as public. Anyone with the URL can read it, and IDs, while random, are not secrets. Do not put credentials, personal data or anything confidential in an endpoint. If you need private data, you need a real backend.
256 KB per endpoint. That is comfortably more than any realistic mock API response and keeps the service fast for everyone.
Reads are cached briefly at the edge and are not metered for normal use. Sustained automated hammering will be throttled. If you need production-grade throughput, this is the wrong tool — copy the JSON into your own infrastructure.