Move documents between JSON and YAML without leaving the browser. Use JSON→YAML when humans need a readable config snapshot. Use YAML→JSON when a tool chain only accepts JSON.
Verified round trip
JSON {"name":"모듈","tags":["a","b"]} becomes:
name: 모듈
tags:
- a
- b
Converting back yields an equivalent JSON object. Key order may differ after serialization; sort with the JSON formatter before textual diffs.
Collaboration pattern
Keep commented YAML as the source of truth in git. Emit JSON only at the boundary that requires it. If application code emits JSON that operators must read, convert to YAML for review, then discard that YAML rather than editing it as canonical.
Watch unquoted values: 01, NO, true may change type depending on YAML rules. Quote strings that must stay strings, then convert again and inspect types in JSON mode.
Limits
Huge documents, cyclic structures and exotic YAML tags may fail or simplify. Treat the page as a scalpel for samples and debugging, not a bulk ETL engine. Pair with the workbench when hashing or Base64 follows the conversion.
Concrete scenarios
Scenario A — document an API. Capture a JSON response, convert to YAML, and add human comments in the editor afterward. Remember that converting back to JSON drops those comments, so store the commented YAML separately if comments matter.
Scenario B — debug CI. Paste the failing YAML fragment, convert to JSON, and read the position-aware error. Mixed tabs and spaces, or a missing indent under a list item, are usual culprits.
Scenario C — feed a JSON-only CLI. Convert the ops YAML to JSON, pipe into the tool, and keep YAML in git. Automate the conversion in CI once the sample looks right here.
Type gotchas worth a second look
Unquoted NO, Yes, 01, or version-looking tokens may become booleans or integers depending on YAML rules. After YAML→JSON, inspect types explicitly. Quote anything that must remain a string. When comparing two JSON emits, run them through the JSON formatter’s sort mode so key order does not create fake diffs.
This page is intentionally narrow. Image work, character counting and romanization live elsewhere; chaining several text transforms belongs in the workbench.
What success looks like
A successful conversion is boring: types match intent, comments are either preserved in the YAML source of truth or consciously discarded, and CI consumes JSON without hand edits. Use this page to prove the boring path on a sample, then automate. If you find yourself pasting megabytes repeatedly, move the step into a scripted pipeline where it belongs.
Minimal acceptance test
Take one real (non-secret) config, convert YAML→JSON→YAML, and read the second YAML. Structure should match; comments will be gone. If types shifted, quote the offenders in the source YAML and repeat until the JSON types look right. That loop is the whole point of a browser-local converter.