Limits

The open-source component does not enforce a fixed max file size or max row count. Parsing and preview run in the user's browser, so memory and CPU on that device are the real limits.

Preview renders the first 100 rows while counts reflect the full file. Prefer reasonable customer uploads (thousands of rows, not multi-hundred-MB workbooks) and fail clearly in your own product copy if you need a business-level cap. Optional useWorker moves parse off the UI thread when the browser provides a Worker, then falls back to the main thread. This is not a million-row canvas editor.

Text that is not valid UTF-8 is decoded as Windows-1252. UTF-8, including a UTF-8 BOM, is unchanged. UTF-16 still requires a BOM. There is no encoding picker. On the map step, Separator can be Auto, comma, semicolon, tab, or pipe for that file only. The choice is not saved. .tsv / .tab Auto stays tab. Neither behavior adds a file-size or row cap.

Related: File formats · Errors