What is the CSV Reader?
Raw CSV text is hard on the eyes — endless commas, wrapped lines, no column alignment. The CSV Reader flips that around: paste a file and you get a readable table with aligned columns, so the data actually tells you something. Use it when you just need to see what's inside a CSV someone sent you.
This isn't about editing or transforming — it's about reading. When a vendor drops a 3,000-row customer export in your inbox, or when you're debugging why the ETL job complained about row 2,147, the Reader gives you a human-friendly view in one paste. It handles quoted fields and common edge cases quietly, so you can focus on the data itself. For heavier workflows, Python's csv module and Excel work too — this tool is for the 30-second "let me see the file" moment.
Reading happens in your browser. Files never touch a server, which matters when the CSV contains names, emails, or other sensitive columns. CSV itself is a plain-text format described well in the Wikipedia article on CSV and formalized in RFC 4180.
Why This Reads Better Than a Text Editor
Aligned Column View
Columns line up. Your eye can scan down a column without counting commas. Huge difference when the CSV has 15+ columns and mixed-length values.
Per-Column Search
Type in any column header to narrow the view. Works like a spreadsheet filter — useful when you need to see only rows where `department` is `Engineering` or `annual_salary` starts with a 1.
Value Inspection
Click into a cell to read long values without truncation. Helpful when one column holds a long note or serialized payload.
Hand Off as JSON
Need to pass the data to a teammate in a different format? Export as JSON in one click. The reader already figured out the types, so the JSON is usable.
Drag-and-Drop
Drop a `.csv` file directly onto the page. No upload progress bar because nothing is uploaded — the file reads locally.
Hide Noise Columns
Hide columns you don't need so you can focus on what matters. The underlying data stays intact.
Reading a CSV in Three Steps
Paste or Drop the File
Paste CSV into the left pane, or drop a `.csv` file onto the upload button. The sample button loads an employee roster with real-looking names (Marguerite Alderwood, Bartholomew Kendrick) so the reading experience is obvious.
Scan the Table
Your CSV shows up on the right as an aligned table. Sort by clicking through headers mentally — and use the per-column filters to narrow what you're reading. Spot outliers like a missing value or an unexpected string in a numeric column.
Export If You Need To
Done reading? Hit `JSON` to save a clean structured version, or `Download` to re-export the CSV. Most of the time you'll just close the tab — the reader did its job.
When You Just Need to Read a CSV
Reading a File Someone Sent You
A client emails a CSV export. Opening it in Excel auto-formats dates and corrupts leading zeros. Use this reader instead — the gene-name-autoconvert problem is a real thing and Excel is famous for it.
Debugging a Pipeline Failure
Your import job says row 2,147 is malformed. Paste the CSV, jump to that row, and the offending value is right there. Much faster than opening the file in a text editor and counting lines.
Quick Data Profiling
Got a new CSV and want to know its shape before writing code? Read it here first. Column names, value distributions, obvious nulls — all visible in seconds. Good companion to a formal tool like <a href="https://pandas.pydata.org/docs/reference/api/pandas.read_csv.html" target="_blank" rel="noopener">pandas.read_csv</a>.
Code Review of Test Fixtures
Reviewing a PR that updated a CSV test fixture? Paste the before and after side by side in two tabs — much easier to spot what changed than reading the diff raw.
CSV Reading Questions
How is this different from opening the CSV in Excel?
Excel tries to "help" by auto-converting types — SKU codes become numbers, leading zeros vanish, `1/2` becomes a date. The reader here shows the CSV as-is. For readonly inspection, that's usually what you want. See the Nature article on gene-name corruption for why this matters.
Does the file get uploaded?
No. The CSV is read by your browser's File API. Nothing is sent over the network. If you open devtools and watch the Network tab while loading a file, you'll see zero requests triggered by the upload.
How big a file can I read?
A few hundred MB of CSV typically works in a modern browser. The reader holds the whole file in memory, so the practical limit is your device's RAM. For gigabyte-scale files, stream them with a tool like csvkit or DuckDB.
Why are some cells showing as numbers and others as strings?
Type inference. If a column value looks like a number (`12.00`) we render it as a number; otherwise it stays a string. That's why a ZIP code column like `07001` might show as `7001` — the leading zero is lost. If you need strict string handling, wrap those values in quotes in the source CSV.
Can I read a CSV with a different delimiter?
The reader is tuned for comma-separated files. For semicolon-delimited European CSV or tab-separated files (TSV), swap the delimiter before pasting, or use a tool that lets you pick — most CSV libraries do.
What about CSV files with a BOM?
Byte-order marks (`\uFEFF`) at the start of a file are stripped silently. You'll sometimes see a BOM on CSVs exported from Excel. See the W3C note on byte-order marks for context on why they exist.
Other CSV Tools
The Reader is for reading. If you need to parse, validate, or convert, these are the right tools: