LARGE FILE MODE · 0 BYTES UPLOADED
Convert the CSV files
other tools refuse.
Most converters cap out at a few megabytes, or make you upload your data to find out. Ours reads, converts and exports multi‑megabyte tables entirely in your browser, in a background worker, with nothing sent anywhere.
Free · no account · works offline once loaded
Why large files break other converters
It's never your file's fault. It's a business model that can't afford to process it for free, or won't risk touching it at all.
“Drop your file to begin”. Your data leaves your machine before anything happens.
A quiet 5 MB ceiling, discovered only after you've already handed the file over.
A spinner on someone else's hardware, then a 504. Your file was never the problem.
“Upgrade to Pro to convert files this size.” The paywall is the size limit.
Built to take the big one
Read off-thread
The file is read in a background Web Worker, so a 40 MB table never freezes the page.
Convert in one pass
The RFC 4180 parser walks the table once, on your machine. No round-trip, no queue.
Preview + full download
A capped preview appears in the editor; the complete result ships as a one-click download.
Nothing transmitted
There is no endpoint to send it to. None was ever built. Zero bytes uploaded, provably.
Where the ceiling really is
We'd rather tell you the real limit than invent a comfortable one. Because everything runs on your machine, there's no quota we impose and no server to time out. But there's also no server-grade RAM. The boundary is your device, and we say so up front.
See the live proof it never uploads →- Measured baseline on 8GB Win10, Node 24, V8 2240MB heap (median of 3, seed 42, 10 cols): 1K 37.7ms, 10K 41.3ms, 100K (9.4MB) 1,531ms, 500K (46.8MB) 5,553ms parse and 12,686ms pipeline
- 1M rows (93.7MB) OOM at 2GB heap — preview capped at 100K chars, download holds full file; close tab to reclaim memory
- CSV → JSON output expands about 2.9× input size, so it hits the ceiling first; give it the most headroom
- CSV ⇄ TSV stays about 1× and handles the largest tables of the three
- Editor shows capped preview so rendering never locks the tab; full file is always the download. No accounts, no upload, no servers
Large files, asked plainly
How large a file can I actually convert?
No fixed cap, but the ceiling is measured: 100K rows (9.4MB) parses in 1,531ms, 500K rows (46.8MB) in 5,553ms and 12,686ms full pipeline on 8GB Win10 Node 24 (median of 3, seed 42). 1M rows (93.7MB) OOMs at 2GB heap — preview is capped at 100K chars. CSV ⇄ TSV handles the largest files at about 1× size; CSV → JSON is tightest at about 2.9×.
Does the file get uploaded to check its size?
No. The file is read locally by your browser into a Web Worker. Nothing is transmitted anywhere. There is no backend for it to be transmitted to.
Why do I get a preview instead of the full result in the editor?
Rendering tens of megabytes into a text editor would lock your tab. So we render a capped preview you can inspect, and hand you the complete result as a download. You lose nothing.
Which tool should I use for a large file?
CSV → TSV or TSV → CSV for the biggest tables (1× output). CSV → JSON when you need structure. Its output is ~2.9× the input, so give it the most headroom.