Reading stored document
JSON:API document viewer
Follow the pointer.
JSON:API keeps related resources in a sibling included array, so
following a relationship means hunting a UUID through the whole payload. Paste a
document and every pointer becomes a link you can click — with real browser
history, deep links and find-in-page.
.json file
A pointer that resolves is a link. Clicking it scrolls to that resource and adds a history entry, so Back returns you to where you were.
A pointer that resolves to nothing says so. That distinction is usually
the thing you are looking for — a missing include parameter, or a
server that dropped something.
Nothing leaves your browser. Parsing, indexing and rendering all happen on this page, and the document is stored in local IndexedDB so a reload keeps your deep links working. There is no server to send it to.
Questions
What people ask before pasting a payload into somebody else's tool.
Does anything I paste leave my browser?
No. Parsing, indexing and rendering all happen on this page, and the document is kept in your browser's IndexedDB. There is no upload step and no account. The one exception is opt-in: a share link is encrypted in your browser first, and the key travels in the link itself, so the server only ever stores ciphertext it cannot read.
What is a JSON:API document?
A response shaped by the JSON:API specification: a top-level data
key holding resource objects, whose relationships are
{type, id} pointers, and a sibling included array
holding the resources those pointers name. Following one relationship by hand
means searching the whole payload for a UUID.
How do I follow a relationship?
Click it. Every pointer that resolves is a real anchor to that resource on the same page, so Back and Forward, deep links, find-in-page and “copy link address” all behave the way they do everywhere else.
What does “not in document” next to a pointer mean?
The relationship names a resource that is in neither data nor
included. Usually that is a missing include parameter
in the request, or a server that dropped something — and telling that
apart from a resource that is present is normally why you opened the payload.
How large a document can it read?
25.7 MB and 56,821 resources render in about 1.6 seconds in Chrome on Apple Silicon, and scroll smoothly afterwards. The binding constraint is DOM node count rather than payload size, so the practical ceiling is around 100,000 resources.
Is there an API, a server or a sign-up?
None of the three. It is a static page — markup, one JavaScript bundle and self-hosted fonts — with no analytics, no cookies and no third-party requests. The only server-side code in the project stores encrypted share blobs it cannot decrypt.