Comma, semicolon or tab? Find out before you import.

Paste delimited text or drop a CSV below — get the separator detected with a confidence verdict, every candidate scored, a parsed preview, and one-click re-export. Detected entirely on your device.

  • Your file data never leaves the device
  • Comma, semicolon, tab, pipe, colon, space + custom
  • Batch up to 20 files + re-export
  • Quoted fields handled correctly

Paste text or add your files

Include the header row and a few data rows — at least two rows are needed for a confident call. Detection is scored on the first 100 KB of each source.

Try a sample:

Drag & drop CSV, TSV or text files here

or browse your files

Detected in your browser — every candidate delimiter is scored on column-count consistency across your rows, with quoted fields handled. Nothing is uploaded.

One character, scored alongside the built-ins.
Up to 20 files, 10 MB each. Pasted text and files are analysed together.
    No text or files yet. Everything stays on this device. Re-exports download instantly.

    Privacy you can verify, not just trust

    Most online delimiter tools send your data to their server first — the best-known dedicated detector processes pasted text through a PHP endpoint, and others sell the same check as a paid API. This one can't: your text and files are parsed entirely in your browser. Your file data never leaves the device — the only third-party requests are the site's own analytics, which count page views — and it uploads nothing, ever. There is no file fetched after load for the detection itself. Don't take our word for it:

    New cross-origin requests during use: 0
    1. Press F12 to open developer tools, then the Network tab.
    2. Detect a delimiter above. Watch: apart from the site's analytics, nothing new is requested — no uploads, and no file data ever leaves your device.
    3. Extra credit: disconnect your internet after this page loads — detection still works.

    How to detect a CSV delimiter without uploading it

    A CSV that will not import is usually not broken data — it is the wrong separator. A team expects commas and receives semicolons. A European export uses semicolons because commas are decimal points there. A data pipeline hands over tabs, a command-line tool hands over pipes. The file looks fine in a text editor, and every import either collapses into one column or explodes into dozens.

    This tool finds the separator where the file already is: on your own device. When you paste text or drop a file here, your browser:

    1. Scores every candidate delimiter — comma, semicolon, tab, pipe, colon and space, plus any custom character you supply — by parsing a sample of your text with a quote-aware parser and counting how consistently rows split into the same number of columns. A delimiter inside a quoted field never counts.
    2. Reports the winner with a confidence verdict — High, Good, Uncertain or Low — plus the full score table, so you can see why it won and how close the runner-up was. The header row is called out, and any row whose column count disagrees is named by line number.
    3. Shows a parsed preview and a one-click re-export — the first rows split on the detected delimiter, so a wrong call is visible at a glance, and a download that rewrites the file with the delimiter of your choice, quoting fields that need it.

    What makes this detector different

    The pages that top the results for this search mostly process your text on their server, cap the input at 200 KB, or sell the same check as a paid API — and the linters that include detection treat it as a side feature of a bigger report. This page is the only free, fully client-side dedicated delimiter detector we found that shows its work: every candidate scored in the open, a parsed preview, header detection, batch detection across 20 files, and a re-export with a chosen delimiter — with no daily limits, no account and no upload — plus a live network-request counter so you can verify the privacy claim yourself. It runs entirely in your browser — there is no server to upload to.

    When is detecting the right choice — and when is it the wrong one?

    Detecting is right whenever a file has to import somewhere else: a spreadsheet, a database, a CRM, a mail merge, or a data pipeline that expects a specific separator. It is the wrong choice for a file you only ever open yourself in a text editor, where the separator never matters, and for a file whose rows genuinely have different column counts — ragged data cannot be fixed by choosing a delimiter, and the score table will show that plainly. If the source system can export again, export with quoting switched on instead of patching the file.

    Honest limitations

    Heuristic, not a validator

    This tool scores candidates on column-count consistency — it is not a CSV validator. It cannot certify that a file is well-formed or that its values are correct.

    Unquoted delimiters in data

    If a comma appears inside unquoted free text in a comma file, no detector can tell data commas from separator commas. Quote those fields in the source export.

    Space is the weakest candidate

    Ordinary sentences can split “consistently” on spaces by accident. A space win with a low score is reported as uncertain — check the preview before trusting it.

    Header detection is a guess

    The first row is called a header when it is all distinct, non-numeric text. A data row that happens to look like that will be miscalled — the preview shows the call.

    Sampled, not whole-file

    Detection is scored on the first 100 KB and 200 rows of each source. A file whose delimiter changes halfway through will be scored on its opening — the ragged-row warning flags the rest.

    It never changes your file

    This tool only reports and re-exports to a new download. Your original file is never modified — to fix broken quoting, repair the export in the application that made the file.

    Frequently asked questions

    Do my files upload anywhere?

    No. Your text and files are analysed entirely in your browser — your file data never leaves the device, and the only third-party requests are the site's own analytics, which count page views. You can verify it yourself with the network-request counter in the privacy section below the detector.

    How does the detector work?

    For each candidate delimiter — comma, semicolon, tab, pipe, colon and space — the tool parses a sample of your text with a quote-aware parser and scores the candidate on how consistently rows split into the same number of columns. The most consistent candidate wins, and the score table shows every candidate so you can see the reasoning. A delimiter inside a quoted field never counts, and a candidate that only ever gives one column scores zero.

    Which delimiters can it detect?

    Comma, semicolon, tab, pipe, colon and space, plus any single custom character you type in the custom-delimiter box. European exports often use semicolons because commas are decimal separators there, data pipelines often use tabs, and command-line tools frequently produce pipe-separated text — those are exactly the cases this tool is built for.

    What if my CSV has commas inside quoted fields?

    That is handled: the parser reads fields wrapped in double quotes as single values, so a comma inside "Smith, John" does not split the row, and a doubled quote ("") inside a quoted field is read as one literal quote. If your file has commas inside unquoted text with no quoting at all, no detector can tell data commas from separator commas — the score table and parsed preview will show the confusion, and you should quote those fields in the source export.

    Why does my European CSV use semicolons?

    Because in much of Europe the comma is the decimal separator (1,5 means one and a half), so spreadsheet exports there use a semicolon as the field separator to avoid ambiguity. The same file opened with a comma assumption collapses into one column. Detect the delimiter first, then import with the semicolon setting — or use the re-export here to convert it to commas.

    What does the confidence score mean?

    The score is the share of sampled rows that split into the same column count under that delimiter: 100 means every row agrees. 90 and above is high confidence, 70 to 89 is good, 50 to 69 is uncertain — check the parsed preview before importing — and below 50 means no clear winner. A single row can never score above 49, because one row cannot prove a pattern.

    Can I detect the delimiter in several files at once?

    Yes. Drop up to 20 files at a time, up to 10 MB each, and detect them all in one go. Every file gets its own verdict, score table and preview, and the summary strip totals the batch. There is no daily limit and no account — detection is scored on the first 100 KB of each file, which is enough for the pattern to show.

    How do I re-export with a different delimiter?

    After detection, pick the target delimiter in the re-export control under a file's preview and press Download re-export. The tool re-parses the text with the detected delimiter and writes it out with your chosen one, quoting any field that contains the new delimiter, a quote or a line break, so the result imports cleanly. Your original file is never changed.

    What does this tool NOT do?

    It is a heuristic detector, not a validator or a fixer. It cannot certify that a file is well-formed, repair broken quoting, detect character-encoding problems, or tell you whether the values themselves are correct — a file whose rows genuinely have different column counts, or whose delimiter appears inside unquoted data, can be scored wrong. The per-candidate scores and the parsed preview are shown so you can judge the call yourself. For a full structural check, use a CSV linter or validator.