unfckr: a free, private data workbench that runs entirely in your browser

Drop a CSV, JSON, NDJSON, XML, Excel, or SQLite file and work on it right here - nothing is uploaded, because there is nowhere to upload it to. Files stream through a transform pipeline, so size is limited by your disk, not your RAM: multi-gigabyte CSVs open fine, well past Excel's 1,048,576-row ceiling.

how to use it

1. Drop a file anywhere on the page, or paste a table straight from Excel, Sheets, or a SQL client. Nothing uploads - there is nowhere to upload it to.
2. Read the ingest panel: what actually came in, and a watchlist of what will fight you later - empty columns, case collisions, duplicate headers. Where there is a safe fix, one click adds the step.
3. Build up a pipeline. Every step is visible, reorderable, switch-off-able, and undoable, and clicking a column header offers the steps that make sense for it. Several steps can be open at once, and the ⤢ button at the top of the panel opens the whole chain full-screen.
4. Export - CSV, JSON, NDJSON, or a .sql script, with the pipeline running over every row on the way out. Or open the ⚡ engine for joins, GROUP BY, full-text search, and a real .sqlite file.

what it does

• Profile every column from the whole file, not a sample: distinct counts, top values, min/max/mean, date ranges, null share, proven uniqueness.
• Clean and transform: flatten nested JSON, scrub PII (mask, hash, or drop emails, phones, SSNs, cards), filter, dedupe, sort, rename and drop columns, split one column into several at a delimiter, extract new columns with regex, fix individual cells - every step recorded, reorderable, and undoable.
• A real data engine: SQLite compiled to WebAssembly. Join across several files, group and aggregate, sort the full file, run read-only SQL in a console, build a full-text index over a text column and search millions of rows instantly, scroll the whole result rather than a first page of it, see a relationship map with discovered foreign keys - each candidate join measured before you build it, so you know whether it drops rows, multiplies them, or matches cleanly, and export a genuine .sqlite database.
• Developer artifacts: CREATE TABLE, INSERT, SELECT, and UPDATE statements for PostgreSQL, MySQL, SQL Server, and SQLite - column types suggested from your actual data - plus TypeScript interfaces and JSON Schema.
• Convert dates: read a column in any of eleven date formats - ISO, epoch in seconds through nanoseconds, US or European slashes, Apache, syslog, Go - and write it back as ISO, a plain date, US or European slashes, or epoch. The input format is detected from your data, and a value that does not match becomes empty rather than passing through, so a column that is half one format and half another shows you which rows.
• Regex painter: highlight two or three examples of what you want to match and get back a regular expression that is verified against your own highlights.
• Streaming exports: CSV, JSON, NDJSON, or a .sql script of your entire file - the pipeline runs over every row on the way out. CSV writes on your terms: comma, semicolon, tab or pipe, LF or CRLF, and an optional UTF-8 BOM so accented characters open correctly in Excel instead of as José.
• JSON that isn't already a table: paste or drop pretty-printed JSON, NDJSON, or an API response - the records inside an envelope ({"data": [...]}, results.items) become the table, other arrays in the document stay one click away, and columns still holding a blob prompt you to flatten them.
• A grid you can actually read: drag a column edge to resize it, double-click that edge to fit the column to its content, or fit every column at once from the header menu - which matters most on logs, where one column holds the whole line. Switch it to paper for dark text on white if that is how you read a table; your theme stays as it is everywhere else.
• A grid you can drive from the keyboard: click a cell and the arrows, Tab, Enter, Home/End and the page keys move a cursor, with Shift extending a selection. Start typing to replace a cell, or press F2 to edit what is there - Enter commits and drops a row, Tab commits and moves right. Every correction becomes one undoable pipeline step, so it survives export and travels in the recipe.
• Selections you can act on: copy a range as tab-separated text and paste it straight into a spreadsheet or an email, fill down or fill right from the edge of the selection, or clear it to empty - from the keyboard or a right-click on any cell. Each becomes one undoable step rather than a hidden change to the data.
• XML, flattened like JSON: drop or paste an XML document and the repeating element becomes the rows. Attributes arrive as @name columns, mixed text as #text, namespace prefixes are stripped unless two would collide, and a nested value is a JSON cell you can flatten again. Repeated elements become arrays consistently across the whole document, so two records of the same shape produce the same columns.
• Sorting, two ways: click a column header to sort the preview instantly (display only), or add a sort step - multi-key, nulls last - so the exported file itself comes out in that order.
• See what one step did: the step studio (⤢, or the commands palette) opens the chain full-screen and shows the rows entering a single step beside the rows leaving it, with the cells that step rewrote marked. Cells are left unmarked where a step drops or reorders rows, because the two sides are then not the same rows. In the engine, the same view carries a sql tab: the one statement your visual steps compiled into, ready to paste into any SQLite client.
• A history you can walk back: every file opened on this device is in recent in the header, from any screen rather than only the home one - and pasted tables are in it too, under a generated name, because a paste is stored exactly like a file. Reopening one brings its pipeline back with it. Nothing there was uploaded; the list and the bytes live on your device.
• Your steps stay with the file: reopen a file and its cleaning pipeline comes back, even after switching tools. A different file with the exact same columns is offered those steps instead of having them applied silently, because steps address columns by position.
• Repairs broken text: José back to José. Non-UTF-8 files are detected when you drop them, and each mojibake repair is proven by a byte roundtrip first, so words that only look wrong are left alone.
• Recipes and history: save or share a pipeline as a link (the recipe lives in the URL fragment, never sent to a server), and reopen recent files from an on-device list that follows you between tools.
• Two sample databases, one click: a real 11-table database (Chinook, MIT) where every declared foreign key measures clean and the two nobody declared are the ones that would have hurt you, plus a small deliberately messy one where a join drops rows, another multiplies them, and a third matches nothing at all.

"CSV" is a loose term here, and paste means paste anything: TSV, semicolon and pipe delimited files, ranges copied from Excel or Google Sheets, SQL client result grids, even tables copied off a web page (they arrive tab-separated) - the delimiter is detected automatically. One delimiter is chosen for the whole file; if a column turns out to carry its own separator inside it, click that column's header and choose split at delimiter to break it into real columns.

every tool in the box

• data workbench - profile, clean, transform, and export CSV, JSON, and SQLite - the full suite (this page)
• JSON flattener - nested JSON or NDJSON to flat columns, arrays and objects unrolled
• CSV PII scrubber - find emails, phones, SSNs, and card numbers - mask, hash, or drop them
• JSON to CSV - convert JSON or NDJSON to CSV with full column control
• CSV to JSON - convert CSV to JSON or NDJSON, streamed at any size
• Excel to CSV - read .xlsx workbooks and export CSV, sheets and all
• CSV to Excel - write a real .xlsx where leading zeros and dates survive
• SQL on CSV - real SQL over your files: joins, GROUP BY, a read-only console
• CSV to SQLite - drop several files, get one .sqlite database with discovered foreign keys
• CSV to SQL - CREATE TABLE and INSERT statements for Postgres, MySQL, SQL Server, SQLite
• JSON to TypeScript - TypeScript interfaces and JSON Schema derived from your real data
• large CSV viewer - open files past Excel's 1,048,576-row limit without loading them into RAM
• regex painter - highlight example matches, get a verified regular expression back
• encoding fixer - detect windows-1252/UTF-16 and repair garbled text (José → José)
• log parser - read log files as events: layouts detected, stack traces kept whole, timestamps normalized, and a volume-over-time strip drawn in SQL across every row
• visualize - chart any table you have open - bar, line, area or point - drag across it to filter, export svg, png, or the vega-lite spec that drew it, or edit that spec in place and run it
• theme builder - design, test, and export custom color themes with live previews, palette derivation, and contrast diagnostics

Like every unfckr tool, this runs 100% in your browser: no upload, no servers, no signup, no analytics. A Content-Security-Policy header blocks every programmatic network channel, so not even malicious code could phone home - and you can watch it happen: the privacy page runs six live escape attempts and shows the browser refusing each one.

↑ back to the tool