OmniJSON / Guide
How to open a large JSON file in Chrome
Publisher and support: OmniJSON on the Chrome Web Store. Checked against version 0.1.14. Version notes; check the store listing for availability.
Updated
To open a large JSON file in Chrome, install OmniJSON, allow local-file access if needed, and open the file’s absolute path or URL in the dedicated viewer. OmniJSON parses locally and lets you browse, search and copy from 100MB-class documents. Loading time depends on the document, hardware and source.
OmniJSON is an extension for desktop Chrome. Install it to try these steps with your own JSON.
Open a URL or local file
- Install OmniJSON from the Chrome Web Store. Check the installed version in
chrome://extensions. - For a local file, open OmniJSON’s extension details and enable Allow access to file URLs. This setting is needed for local paths as well as
file:///URLs. - Click the OmniJSON toolbar icon. In version 0.1.13 and later, choose Open viewer in the launcher. Earlier versions open the viewer directly.
- Paste an HTTP/HTTPS URL, absolute file path or file URL into the viewer’s source field, then choose Open. The source must contain one JSON object or array.
- Browse the first rows as parsing progresses. Use Search or Fields while loading; wait for parsing to finish before running JSONPath.
| Source | Example |
|---|---|
| Web URL | https://example.com/data.json |
| macOS or Linux path | /Users/you/data.json or /home/you/data.json |
| Windows path | C:\Users\you\data.json |
| File URL | file:///Users/you/data.json |
Relative paths and ~/ shortcuts are not supported. These examples describe the format; replace them with your own source. On a detected raw JSON page, you can also use the page’s Open button. The original tab remains available.
Try the workflow with a small, synthetic file first: download the sample JSON. After it finishes loading, Search for paid to find one matching value, or follow the query examples and expected results. This sample uses regular mode; it is not a large-file benchmark.
What changes for a large document?
From roughly 50 MB or 500,000 nodes, OmniJSON uses large-document mode and requests visible rows on demand. The initial view may be shallow. Expand the branches you need or adjust the depth control.
- Search: case-insensitive substring matching over keys and scalar values.
- Fields: exact key names such as
id, status, preserving parent paths and matched subtrees. - JSONPath Lite: structural queries after parsing completes, such as
$.items[*].id. - Prev/Next: move between matches without manually scanning the tree.
Node count matters as well as file size: a smaller document with many tiny values can enter large mode. The product demo uses a synthetic 100 MiB document; its screenshot timings are not a throughput guarantee. See the real viewer examples.
Read about performance, practical limits and validation evidence.
What if Open does not work?
- After updating the extension
- Refresh the original JSON tab so it gets a fresh extension connection. In version 0.1.13 and later, an invalidated connection changes the page button to Refresh. Refresh, then click Open again.
- The toolbar opens but the parser is unavailable
- Version 0.1.13 and later provide a launcher with Check again, recovery instructions and Copy diagnostics. Opening the viewer and connecting to the parser are separate steps. Reload the viewer once; if it still fails, turn OmniJSON off and on in
chrome://extensions, then refresh the source. This interrupts active OmniJSON sessions. - A local file cannot be read
- Check the file-access setting and the absolute path. You can also open the file in Chrome and use the detected Open button.
- A URL fails or JSON is malformed
- Confirm that the URL serves raw JSON in Chrome. If refetching fails, an already-open source tab can provide its visible document as a fallback. A parse error shows line, column and source context.
Follow the full step-by-step Open recovery guide.
Which inputs are supported?
OmniJSON accepts one raw JSON object or array per document. JSONL/NDJSON, multi-document streams, editing and saving are outside the current scope. Local processing does not mean every source is offline: opening a web URL makes a request to that source.