Guide
Open a large CSV file
Excel's row limit is 1,048,576, and it does not break the news gently. The file opens, the numbers look plausible, and the rows past the line are simply not there.
Open a CSV PDF, Word, PowerPoint, PNG, JPG, WebP, Text · XLSX, CSV, Parquet, JSONDrop the file in
CSV, TSV, Parquet, JSON or Excel. It loads into a database that runs inside your browser, so the rows are not uploaded anywhere. You see the columns, their types and the row count before you ask anything.
Ask in a sentence
“How many rows”, “total by region”, “which days are missing”. A query is written for the question and run locally against the whole file, so the answer comes from every row rather than the first few.
Keep going
Each answer stays in the conversation, so the next question builds on it rather than starting over. Suggestions for what to ask next come from what the file actually contains.
The rows do not leave the machine
What goes out is the shape of the table: column names, their types, and a handful of sample values so the query can be written correctly. The data stays in your browser and the query runs there. A million-row file costs a couple of kilobytes of traffic — that is the profile, not the data.
Why won't Excel open my file?
Most likely the row limit: 1,048,576. Excel opens what fits and leaves the rest, and it is not loud about it — which is worse than refusing, because the totals still look like totals.
How many rows can this open?
Up to a million on a desktop and fifty thousand on a phone, where a browser tab gets killed rather than warned. Past that the first rows are kept and the page says so, so a truncated answer is never presented as a complete one.
Can I run SQL on a CSV?
That is what happens underneath: your question becomes a SQL query and runs against the file. The query is shown to you, which is the point — the figure is checkable rather than asserted.
Do I need to install anything?
No. The database is WebAssembly running in the page you are already on. Nothing to install, nothing to configure, and no server to send the file to.
Is it faster than Excel?
For reading and aggregating, usually by a lot — a sum over a million rows is milliseconds. The first file on a fresh page pays a short pause while the engine loads; every file after it is instant.
Can I open Parquet too?
Yes, and TSV, JSON and NDJSON. All of them are free — a paid plan buys more files and more questions, not more file types.
Can it edit or split the file?
No, and deliberately. The table is read, never written, so nothing is sorted, edited or saved over by accident. What comes back is answers, not a modified file.
Is my data uploaded?
The rows are not. Only a short description of the table is sent so the query can be written; the query itself runs in your browser against the full file.
Do I need an account?
No. Three files and twenty questions a week without one.