Explore JSON, NDJSON & CSV Datasets

Inspect Datasets of Any Size and Depth

Stop struggling with unwieldy and undocumented data. Slicelytics automatically infers schemas and maps data types—even across arrays of objects and arbitrary depths. It builds an interactive structural map of your dataset on the fly, allowing you to instantly expand, collapse, filter, and understand the shape of your JSON or CSV, regardless of its original complexity.

To keep your experience smooth during inspection, the UI intelligently loads data in batches, whatever the format, showing the first 100 items and loading more only when you ask for them. "Load more" respects every active filter and brings in up to 100 matching items at a time, wherever they are in the dataset. Need to see everything at once? The Expand toggle opens all items, for datasets of any size.

Filtering stays responsive on large datasets too: a ticked checkbox or a new filter chip shows the moment you click, and the data catches up a moment later, dimmed while it updates.

Advanced Filtering at Your Fingertips

Focus on what matters. Our advanced filtering options allow you to zero in on specific properties and items within your dataset. Use our simple query language to quickly identify exactly what you need:

• Filter Items: Toggle individual indexes (e.g., 1,2,3) or specify whole index ranges (e.g., 0-10). Easily combine both approaches: 0-10,20,25,30.
• Filter Properties: Drill into the shape of your objects by toggling property filters instantly by typing their names (e.g., foo, bar), complete with native support for traversing deeply nested properties (e.g., foo.bar).
• Filter by Value: Keep only the items where a property has a given value (e.g., status="error"). Value filters reach into nested objects and arrays: records.type="sleep" keeps the matching items and trims the non-matching elements from each records array.
• Autocomplete: The properties filter suggests property names as you type, including nested ones, and = to start a value filter. Pick a suggestion with the arrow keys and accept it with Tab or Enter.
• Filter Chips: Every active filter is shown as a chip: hidden items and properties, value filters, and runs of items collapsed into ranges like 3-7. Click the x on a chip to remove that filter.

Hide the noise and focus on the signal.

Keyboard-Friendly and Efficient

We know developers love their keyboards. That's why Slicelytics relies on a robust set of hotkeys so you never have to reach for your mouse:

• Cmd/Ctrl + I: Instantly focus the Item filter search bar.
• Cmd/Ctrl + P: Instantly focus the Property filter search bar.
• Shift + F: Toggle distraction-free fullscreen mode.
• Shift + E / C / V: Rapidly switch between Explore, Compare, and Visualize modes.

Your Data Stays Yours (and Local)

Privacy is paramount. All data processing happens directly in your browser. Slicelytics natively parses JSON, NDJSON / JSONL, CSV, and TOON files entirely client-side. Your data never leaves your machine, ensuring complete privacy and security. We also support native Gzip and Zstandard (zstd) decompression, so you can work with massive compressed log dumps without sending them to an external server. By keeping the processing local, Slicelytics eliminates upload wait times and external bandwidth bottlenecks entirely.

Engineered for Gigabytes: Streaming & IndexedDB

Most browser-based JSON tools crash when handling files larger than a few megabytes because they attempt to load the entire text payload into memory at once. Slicelytics leverages advanced Web APIs to stream and parse data chunk-by-chunk.

For NDJSON files, our native streaming parser processes your data in manageable batches and efficiently caches them directly into your browser's IndexedDB. This zero-memory-leak architecture means you can safely analyze files scaling into the gigabytes right on your laptop. Slicelytics then safely pages records from IndexedDB to the view without locking up the main thread.

The file is read in a background worker, so the page stays responsive while it loads. The first items show up right away, and a progress bar shows how many items have loaded and how much of the file has been read. If a file can't be read to the end, for example one cut short, the items read until then are kept and Slicelytics tells you why it stopped.

Share Your Findings (Without the Server)

Need to share a slice of your data? Click Share to copy a link to exactly what is on your screen: the data as filtered, plus the properties and value filters you set. Whoever opens the link sees the same view. The data is compressed into the part of the link after #, which browsers never send to a server, so it's never stored by us. Browsers limit how long a link can be (about 2 MB in Chrome), so sharing suits small and medium datasets.

Watch a Folder for New Files

Running a script, test suite, or pipeline that writes a JSON, NDJSON, CSV, or TOON file per run? Click "Watch a folder" in the Explore view and pick its output folder. Slicelytics lists the data files in it, newest first, and loads the newest one. New files show up in the list as they are written and wait for your click, so what you are looking at never changes on its own. The folder is remembered across visits. Watching a folder works in Chrome and Edge.

Open Data from AI Agents and Scripts

AI coding agents such as Claude Code can open a dataset in Slicelytics by building a link, as described in slicelytics.com/llms.txt. The link can set up the whole view: the page (Explore, Compare, or Visualize), the properties to show, and value filters. Ask your agent to "show only the failed requests in responses.ndjson" and it opens Slicelytics with that filter already applied, leaving your file unchanged. The data travels in the part of the link after #, so it stays in your browser.

Install Slicelytics as a Desktop App

Prefer a dedicated window? Install Slicelytics from Chrome or Edge and open it like any other app. Links you open go to the app window you already have open, which offers them with a "Show" button instead of replacing what you are looking at. Opening a link while signed out brings you back to that link, with its data and view, once you have signed in.

How to Explore a JSON File Online

1. Load your data: drag and drop a JSON, NDJSON/JSONL, CSV, or TOON file - plain, Gzip, or Zstandard compressed - or watch a folder.
2. See its structure: Slicelytics infers the schema and shows your data as an expandable tree.
3. Filter: narrow it down by item, property, or value until only what you care about is left.
4. Share or switch: copy a link to the view, or move to Compare or Visualize with the same dataset.