NARIKJSON opens 10+ GB JSON, XML & CSV in a blink.
A dual-engine, Rust-powered data workbench. It opens, navigates, queries and converts enormous JSON, NDJSON, XML, YAML, CSV and TSV files without ever loading them into memory — 10+ GB on a laptop, and 50+ GB on a 32–64 GB machine.
1{ 2 "events": [ 3 { "id": 8842013, "level": "error", 4 "service": "checkout", "ms": 1284 }, 5 { "id": 8842014, "level": "info" } 6 ] 7}
One viewer for every large data format
Engineers are already opening the impossible.
Every one of these files was opened with NARIKJSON on someone's own machine — no uploads, no cloud, no waiting.
Counts include log files, API dumps and event streams — updated as the community grows.
// a developer, to a developer
Built by a developer who lives in big data — for the ones who do too.
No bloat, no telemetry, no progress bars — just the tool we wished existed every time an editor choked on our files. That tool is NARIKJSON.
Open it. Navigate it. Find anything.
Every file is memory-mapped and parsed once into a compact index — only the rows on screen are ever built.
Open the file
Drag in a JSON, NDJSON, XML, CSV or TSV file up to 10+ GB (50+ GB with 32–64 GB of RAM) — or paste from the clipboard or fetch a URL with auth. It opens instantly, no progress bars.
Navigate the tree
Everything renders as one uniform, virtualized tree. Jump to any path, bookmark nodes, and read collapsed counts like “products : […], 31 items”.
Search & export
Regex find, live filter, or JSONPath / XPath queries backed by parallel native search. Then export or convert to JSON, XML or CSV.
Built for scale, byte by byte.
NARIKJSON's dual engine treats your data as structure, not text — in-memory for speed or database-scale for the biggest files, auto-selected. Memory tracks your working set, not the file size.
The result speaks for itself.
Approximate load time for a large JSON file. Where general-purpose editors stall or run out of memory, the structural index just opens.
Illustrative comparison on a large JSON file. Actual results vary by hardware, file shape and configuration.
A data workbench, not a text editor.
Browse, query, transform and convert your files like a database — with the tools you'd expect from an enterprise DBMS.
Virtualized tree view
Objects, arrays, XML elements and CSV records render in one uniform, collapsible tree. Cost per frame tracks visible rows (~40), never file size.
Lightning search
Trigram-indexed literal or regex search scoped to keys, values or attributes — finds anything in milliseconds, even on massive files.
JSONPath, XPath & jq
Query JSON/CSV with $.store.items[*].name and XML with //item/@id, or run a full jq program in the built-in bar — results open in a new tab.
Dual-engine architecture
Two engines in one — in-memory for speed, database-scale for the biggest files — auto-selected for your data. Files too large to index eagerly still open on demand.
Spreadsheet-grade tables
Pin, filter, sort and profile millions of rows like a spreadsheet, with a SQL console over CSV, TSV and delimited files. Leading-zero cells like ZIP codes stay strings.
Follow tail (tail -f)
Follow a growing JSON / NDJSON / log file. Only newly appended lines are parsed each tick, so it stays cheap on huge, live files.
Every format, one click
JSON, NDJSON, XML, YAML, CSV, TSV and delimited files in one app — convert freely between them, batch-export, or save your search matches as a dataset.
Integrated API workbench
Postman-style requests with None / Basic / Bearer / API-Key auth — explore live responses as structured data and copy any request as cURL.
Schema, diff & insight
Generate and validate schemas, diff two files, check column coverage, and read child counts, type histograms and numeric min / max / avg for any node.
Other tools weren't built for gigabytes.
General-purpose editors treat JSON as plain text. NARIKJSON treats it as structured data.
| NARIKJSON | Text editors | Browser / online | |
|---|---|---|---|
| Opens 10+ GB files | ✓ | ✕ | ✕ |
| Memory tracks working set, not file size | ✓ | ✕ | ✕ |
| Structural tree navigation | ✓ | ✕ | Limited |
| Parallel regex search | ✓ | Slow | ✕ |
| JSONPath / XPath queries | ✓ | ✕ | ✕ |
| Follow a live / growing file | ✓ | ✕ | ✕ |
| Cross-format convert (JSON/XML/CSV) | ✓ | ✕ | Partial |
| Data never leaves your machine | ✓ | ✓ | ✕ |
What people say.
“Opened an 8 GB NDJSON log that crashed everything else. It was on screen before I finished my coffee.”
“The JSONPath query bar plus regex find replaced three scripts I used to run to inspect API dumps.”
“Follow-tail on a live JSON log is the feature I didn't know I needed. Memory just doesn't move.”
Book a live demo.
Thirty minutes, screen-shared, on a file the size of the ones you actually work with. No slides.
- Bring your own file — or use one of ours
- Engine modes, SQL console and jq, walked through live
- Licensing and rollout questions answered
Pick a date and a time that suits you — we'll confirm by email.
or email openxmljson@gmail.com
What to know before you download.
How large a file can NARIKJSON really open?
On a typical machine it's comfortable around 10 GB. With 32–64 GB of RAM it opens 50+ GB files and hundreds of millions of rows — performance scales with the memory you give it. Files too large to index eagerly fall back to the database engine and on-demand indexing, so documents near or beyond your available RAM still open.
Which formats are supported?
JSON, JSON-Lines / NDJSON, XML, YAML, CSV, TSV and other delimited files — all in one app, with one-click conversion between them. You can also paste from the clipboard or open from a URL.
What are the two engines?
NARIKJSON runs an in-memory engine for speed and a database engine for scale, and picks the right one for your file automatically. The database engine uses SQLite for JSON and XML and DuckDB for CSV and delimited data, which is what makes multi-GB files and 100M+ rows practical.
Does my data leave my machine?
No. Files are processed locally by the native Rust core. Nothing is uploaded — the only network calls are the ones you make yourself from the API workbench.
What platforms does it run on?
macOS (Apple Silicon and Intel) and Linux today. A Windows build is coming soon — get notified.
What's the difference between the free and paid tiers?
They are the same complete application. NARIKJSON Free is billed monthly at no cost; the paid NARIKJSON license is billed yearly and adds volume and site licensing, managed deployment, centralized updates and a dedicated support SLA for teams.
Stop waiting on progress bars.
Download NARIKJSON and open your biggest file in seconds.