About
JSONForge is a JSON workbench that runs entirely in your browser. It exists because the default way people handle JSON — paste it into a random website — is a privacy problem disguised as a convenience.
Why it exists
Every online JSON formatter can read what you paste, and several do. The alternative people reach for is a local editor, which cannot diff two payloads or repair a truncated LLM response. JSONForge is the middle path: the capability of the web tool, the privacy of the local one.
How it is built
Static files on Cloudflare Pages. A Web Worker pool does every parse, diff, repair and inference job, so the main thread never blocks for more than a few milliseconds. Documents live in plain handles outside React, which is why a 100 MB file stays usable. Large files are read in 4 MB slices and rendered through virtualised views.
The part we are proudest of
Array diffing matches rows by a detected key instead of by index. Insert one row into a 10,000-row array and JSONForge reports one addition plus a handful of moves; every other tool reports 10,000 modifications. Moves get their own colour and their own counter, because a reorder is not a rewrite.
What we deliberately do not do
No accounts. No cloud sync. No file size paywall disguised as a "pro" tier. No IP-based language redirect, because serving different content to crawlers than to users is cloaking. No interstitial ad wall: if an ad is blocked, the slot collapses into an equal-height card instead of an obstacle.
Funding
The site is free and funded by advertising, which is why the privacy promise is phrased precisely: your JSON never leaves the browser — not "zero third-party requests", which would be false while ads load.
The source is public:jsonforge/jsonforge