How does it decide what value to generate for a field?

It reads the key name. Keys containing uuid produce a real UUID, keys ending in _id produce incrementing integers or short ids, and keys matching email, name, date, or price produce values of that kind. Matching is word-aware, so userId is treated as an id while valid is not.

Does this use an AI model or an external service?

Neither. It is a few hundred bytes of deterministic key-matching rules running entirely in your browser, with no network request, no API key, and no faker-style dependency. Nothing you paste is transmitted anywhere, which matters when the schema you are mocking came from a real system.

How many array items does it generate?

It templates from the first item in the array and repeats that structure two to four times, varying per generation. That keeps list rendering realistic — enough rows to expose layout problems — without producing so much data that the output becomes hard to read.

~/tools / schema-to-mock

Schema-to-Mock Data Simulator

Feed it the shape — a raw JSON object or a cURL command with a JSON body — and it walks every key, infers intent from the key names, and emits realistic randomized data in the identical structure. No faker, no API, no LLM: a few hundred bytes of deterministic key-context rules running entirely in your browser.

input — json or curl

— waiting for input —

randomized mock output

// mock data will appear here

Persistent live endpoints backed by Supabase SSR edge routing are under development. Everything else on this page is fully functional and runs entirely in your browser.

cat ./tldr.txt

A schema-to-mock generator takes a raw JSON object or a cURL command and returns randomized data in the identical structure, inferring realistic values from key names so ids, emails, dates, names, and prices look plausible rather than being generic placeholders.

./when-to-use-this

You have one example response from an API and need fifty plausible variations for a seeded list view.
A colleague sent a cURL command and you want mock data matching its body without hand-writing it.
You need demo data that looks credible in a screenshot without exposing real customer records.

./frequently-asked

How does it decide what value to generate for a field?: It reads the key name. Keys containing uuid produce a real UUID, keys ending in _id produce incrementing integers or short ids, and keys matching email, name, date, or price produce values of that kind. Matching is word-aware, so userId is treated as an id while valid is not.
Does this use an AI model or an external service?: Neither. It is a few hundred bytes of deterministic key-matching rules running entirely in your browser, with no network request, no API key, and no faker-style dependency. Nothing you paste is transmitted anywhere, which matters when the schema you are mocking came from a real system.
How many array items does it generate?: It templates from the first item in the array and repeats that structure two to four times, varying per generation. That keeps list rendering realistic — enough rows to expose layout problems — without producing so much data that the output becomes hard to read.

cat ./tldr.txt

./when-to-use-this

./frequently-asked

ln -s ./related-tools