◆ File conversionFREE

Engineering ships JSON.
Translators just want Excel.

JSON⇄XLSXNested key paths, placeholders, escapes — all preserved · refill never drifts, hierarchy never collapses.
Converter· Local processing
Drop your .json here
or pick from your device · your file never leaves your browser
01 · what this tool actually does for you

Flatten a language file into an Excel sheet a translator can edit, then load it straight back to JSON — nested key paths, placeholders, escapes and key order all preserved, refill never drifts, hierarchy never collapses. Not generic JSON-to-table — the one step a localization workflow actually needs.

Files never leave your deviceIn-browser · zero upload · NDA-friendly
Deterministic conversionThe same input always yields the same output
Lossless round-tripJSON→Excel→JSON keeps every key
02 · two formats, one round-trip

JSON and Excel are each good at different things

This conversion exists so engineers and translators each use the tool they're comfortable with — without losing anything in between.

i18n JSON.json

The language file front-end and client teams use: key-value pairs, often nested (menu.start.label), with arrays and plural branches (items[0], ICU plural), values carrying placeholders ({name}, %s) and escapes (\n, \uXXXX). Great for engineers — not something to hand-edit row by row.

Excel.xlsx

The interface translators, reviewers and vendors find most natural: one row per string, key and translation side by side, easy to filter, bulk-edit and annotate. The catch: raw Excel flattens the nesting, treats {name} as plain text and auto-corrects \n — and then it won't load back into the codebase.

What this tool guards is exactly that "won't load back": on export it carries nested key paths, array/plural indices, placeholders and escapes faithfully into the sheet; on import it refills by key, restores the hierarchy, and doesn't drift even if rows were reordered. These are the fidelity details only a localization pro thinks to check — and that generic JSON-to-Excel tools mostly ignore.

03 · common questions

You're probably wondering

Will my file be uploaded?+
No. Parsing and conversion run entirely in your browser (a Web Worker handles large files); the file never leaves your machine and never touches a server. That's a hard requirement for NDA-bound localization work, so the tool is built fully client-side.
Does converting JSON to Excel lose anything?+
The essentials don't: nested key paths, array and plural indices, placeholders, escapes and empty values are all carried by rule, and the round-trip is faithful. A few constructs that can't map (such as comments in JSON5) are reported as not carried at export — never silently dropped. Every conversion comes with a fidelity report.
Converting Excel back to JSON — will the hierarchy or order break?+
No. Refill anchors by key, restoring flattened paths to their original nesting; reordering rows in Excel doesn't matter. Before refill it also checks each translation's placeholders against the source and flags mismatches, and any rows you added with no matching key are listed for review.
Why not just use a generic JSON-to-table tool?+
Generic tools treat JSON as plain data and flatten it: placeholders become text, escapes break, nesting can't be restored — and it errors on the way back into the codebase. This tool understands i18n language-file semantics and only converts between localization-native formats.
Does it support nesting, arrays and plurals, and is it free?+
It supports nested keys, arrays and ICU plural branches. This tier is completely free, no account, no usage limits — a client-side tool has no server cost. The exact JSON dialects and edge cases are described in the tool's format notes.
04 · who made this / why

Built by someone who's done this work

This isn't a thin wrapper over a large model, and it isn't an open-source library reskinned. It comes from years of game-localization frontline work and the pile of "nobody notices until it breaks" details that come with it.

Engineering hands over a JSON; the translator only wants to edit in Excel — the format loss in between shouldn't be a human's job to absorb.

Hand a language file to a vendor and nine times out of ten it comes back with placeholders "auto-corrected" away, \n turned into real line breaks, nesting flattened, plural branches misaligned — and it lights up red back in the codebase. We hit every one of those, so the tool guards them by default: key-anchored refill, placeholder checks, hierarchy restored.

Free, fully local, instant. It's one step in a workflow, not a platform — but it comes from a place that genuinely understands localization.