A local JSON workbench for Chrome
OmniJSON is a Chrome extension for exploring large JSON from URLs and local files. Parse locally, search keys and values, filter fields, and copy exact paths—even in 100MB-class documents.
Practical guides
Steps, working queries and limits for the questions that come up when exploring large JSON.
URLs, local-file permissions, large-document mode and recovery when Open fails.
Try keys, array indices, wildcards and recursive lookup—with expected matches.
Understand source requests, recent URLs and the information that stays on your device.
Understand loading stages, document shape, query budgets and historical validation.
Refresh old tabs, recover startup and collect a useful support report.
The viewer
Open a raw JSON URL or local file in a dedicated viewer. Your original tab stays in place. Expand a branch, search keys and values, or keep only a few named fields—even while the document is still loading.
Once parsing finishes, use JSONPath to reach a specific object or the same field across an array. Large documents use JSONPath Lite for structural queries. Select any row to inspect its value and copy a JSONPath, JavaScript, jq, or key path.
18-second walkthrough
Expand, search, choose fields, query, and inspect. An edited walkthrough of real screenshots using synthetic data.
How it works
Start with a URL or a local file. Browse, narrow the view, then take the value or path you need.
After installing, click the OmniJSON extension icon in Chrome. In version 0.1.13 and later, choose Open viewer in the launcher; earlier versions open the viewer directly. Paste a source into the viewer’s address field and choose Open.
The viewer opens separately and leaves your original tab in place. To reopen a recent source, type oj then Space or Tab in Chrome’s address bar. Suggestions come from OmniJSON’s local recent-URL list.
Expand a branch or use the depth control to collapse nested content. Then choose a filter for the question you are asking.
| Mode & example | What it finds |
|---|---|
Searchpaid |
Case-insensitive substring matches in keys and scalar values.Available while loading |
Fieldsid, status, total |
Exact key names, separated by commas. Keeps parent paths and each matched field’s subtree.Available while loading |
JSONPath$.items[*].id |
Structured matches, such as every item’s ID. In Search mode, start with $ to run a JSONPath query.Available after parsing finishes |
Use Prev/Next to jump between matches. Large documents use JSONPath Lite for structural paths such as $.items[0] and $..id.
Select a row to see its path and inspect its value. Long text is paged; escaped JSON can be viewed in a formatted form.
Copy actions use the complete selected value within the limits below, even when the inspector shows one page or a shortened preview.
Open one raw JSON object or array per document, from an HTTP/HTTPS URL, a local file, or a supported extension-resource URL. API data, exported datasets, and configuration files all fit this workflow.
JSONL / NDJSON, multiple documents in one stream, editing, and saving are outside the current scope. OmniJSON copies jq-style paths; it does not execute jq queries.
For a URL, check that it opens as raw JSON in Chrome. If automatic detection misses it, enter the URL directly in the viewer.
For a local file, check Allow access to file URLs in OmniJSON’s extension details, then reopen the viewer. Use an absolute path or a file:/// URL; relative paths and ~/ shortcuts are not supported. You can also open the file in a Chrome tab and use Open in OmniJSON when detected.
If parsing fails, the viewer shows the line, column, and source context to help locate malformed JSON.
After an extension update, refresh the original JSON tab. Version 0.1.13 introduced a toolbar launcher with connection checks, recovery steps and Copy diagnostics. See the opening and recovery guide.
Wait for the current document to finish parsing. Keyword search and Fields mode can work on the data already loaded.
Regular documents use full JSONPath, subject to query limits. Large documents use JSONPath Lite: keys, array indices, wildcards, and recursive key lookup. Predicates, slices, unions, functions, and comparisons are not supported in Lite mode.
Regular mode lets you page through the complete selected value. Large-document mode may shorten the inspector preview; it labels this and shows the full value size. The visible page does not determine what is copied.
Requests over these limits show an error instead of silently copying a partial value.
From roughly 50 MB or 500,000 nodes, the viewer switches to a lighter mode that loads visible rows on demand. The first view may be shallow; expand the branches you need. Search, Fields, match navigation, path copying, and JSONPath Lite remain available.
Select a scalar to load its details in the inspector. Hover or click the status area to see where time is spent reading and parsing the source.
The header shows the total elapsed time for the current load. Hover or click the status area for the available phase metrics.
Streaming phases can overlap, so their durations should not be added together. Document structure, node count, hardware, source, and cache all affect loading time. Timings shown in demo screenshots describe those individual runs.
Privacy
Last updated:
Parsing, filtering, and JSONPath run in your browser. The extension does not upload documents, query results, recent sources, clipboard output, or usage data to OmniJSON-operated servers. It includes no analytics, advertising SDKs, telemetry, or remote formatting.
The content script checks page content locally to detect raw JSON and offer the Open action. JSON you choose to open from a URL, file, or accessible extension resource is used for tree browsing, Search, Fields, JSONPath, previews, and copy tools. Parsed documents and indexes stay in memory; they are not saved to extension storage.
Opening a web source requests the URL you choose. Chrome may include applicable cookies or authentication credentials and follow server redirects, including to another host. Source servers receive those requests. OmniJSON does not inspect Chrome’s cookie store or upload the returned JSON to a separate processing service. If fetching fails, the original JSON tab can supply its visible content as a fallback.
Reading file:// sources or absolute local paths requires Chrome’s
Allow access to file URLs setting. Paths are converted to file URLs
for direct reading, with an already-open JSON tab available as a fallback. File
contents are parsed locally.
For oj and viewer suggestions, OmniJSON saves up to 50 complete
source URLs, derived display titles, and last-opened timestamps in local
extension storage, even if loading fails. URLs retain query parameters, fragments,
and file paths, including any tokens or sensitive text in them. This list does not
separately store parsed JSON, copied content, browser cookies, or filter/JSONPath
expressions. OmniJSON does not use Chrome sync or read Chrome browsing history.
Records have no time-based expiry; new sources beyond the 50-entry limit replace the oldest. Type oj then Space or Tab in Chrome’s address bar and use a suggestion’s delete control to remove it from both suggestion lists. Reopening that source saves it again. Uninstalling OmniJSON clears its local storage; clearing Chrome’s browsing history or cache does not.
Choosing Copy writes the selected path or content to your system clipboard. OmniJSON does not read the clipboard programmatically. Copied content is then handled by your system clipboard and any destination where you paste it.
Version 0.1.13 and later retain up to 20 local startup failure records: timestamp, extension version, failure stage, error code and elapsed time. Copied reports also include the browser version. These reports exclude JSON, source URLs, credentials and raw error text, and are never uploaded automatically. New records replace older ones after the limit; uninstalling the extension clears its local storage.
Chrome’s local developer console can contain source URLs, filter/JSONPath expressions, status, and errors. The extension does not upload these logs or send automatic error reports.
This site’s own code has no analytics scripts and sets no cookies or browser storage. Visiting it still sends web requests to the hosting service; hosting-level logs or analytics depend on deployment settings. Install links carry UTM campaign labels to the Chrome Web Store, including source, medium, or campaign values from the page URL. They do not include extension documents or recent-source records. Following a JSON value link or the help link also opens that destination, whose own privacy policy applies.
For privacy questions, use the publisher’s support channel on the Chrome Web Store listing. Recent-source records are stored on your device and can be removed as described above.
Version notes
These notes describe versioned changes. Check the Chrome Web Store listing for the currently available version and the publisher’s support channel.
For help, start with opening and recovery steps. When contacting support, include your installed extension version, Chrome version and reproduction steps. Do not include private JSON or source URLs containing credentials.
