A big JSON document is a hostile environment. Five nested levels deep, a key named something almost what you were expecting, and a dozen objects that all look the same at a glance. Ctrl+F helps, but it's blunt — it highlights every occurrence of the literal string, regardless of structure. Jsonize's fuzzy search goes a step further, and that step makes huge JSON files feel navigable.

What "fuzzy" means here

"Fuzzy" gets thrown around loosely. In Jsonize it means three things:

• Structure-aware. Search matches are filtered against the parsed JSON tree, so a match in a key label looks different from a match in a value. The surrounding object stays visible so you keep context.
• Forgiving. Partial matches win. Searching for user finds userId, user_name, and currentUser. You don't need to know the exact casing or the full key name.
• Live. Results update on every keystroke. No hitting enter, no waiting. The document filters itself as you type.

How to use it

There's a single search input at the top of the viewer. Start typing and the document filters in real time — irrelevant branches collapse or hide so you're left with just the paths that contain a match. Clear the search with the x button or by emptying the input, and the tree returns to its prior state.

In the formatted card view the matching nodes stay highlighted so you can spot them quickly. In the code view the match is shown inline alongside its parent path.

Patterns that work especially well

1. "Find every API error"

In a large response, search for error. Every key or value containing the substring — error, errors, errorCode, _error_ — shows up together. You can immediately see how the API reports trouble.

2. "Where is this user?"

Got a user ID or email? Paste it into the search. Fuzzy matching means you don't have to worry whether it lives in user_id, userId, ownerId, or inside a free-text note field.

3. "Schema drift"

Looking at a response from a new version of an API? Search for a known key and watch how it appears in each object. If some objects have userName and others have user_name, the search surfaces both in the same filtered view.

4. "Find all timestamps"

Search for a year like 2025 or a format fragment like T00:00:00. Every timestamp-ish value stands out at once — handy when you're verifying that a batch job used the expected date range.

Keyboard shortcuts that compound

Search is more powerful when combined with navigation shortcuts:

• Ctrl/Cmd + E — expand all nodes. Run this after a search to fully reveal every match.
• Ctrl/Cmd + W — collapse all. Good for getting back to a bird's-eye view after an expansion.
• Ctrl/Cmd + Alt + C — copy the (pretty-printed) JSON. Handy for extracting a filtered section and pasting it somewhere else.
• Escape — close any modal so you can get straight back to the search input.

Performance in practice

Search runs locally in your browser. On modern hardware, searching files up to a few megabytes is effectively instant. For larger files (tens of megabytes) the tree render takes longer to initially paint, but once it's up, filtering remains fast. There's no network latency to fight with because there's no network round trip.

Tip: If a JSON file is really large, collapse everything first (Ctrl/Cmd + W), then search. The viewer only has to highlight the matching subtrees instead of the whole document.

Why it feels fast

Traditional Ctrl+F searches text. Fuzzy search in Jsonize searches structure — keys, values, and paths, with forgiving matching. That combination is what lets you go from "I think the value is somewhere near the shipping block" to "there it is" in one keystroke instead of five.

If you've been reaching for an IDE every time a JSON file gets too big, give it a try. The search alone is worth the switch.