// csv-to-md
CSV to markdown table.
Paste any CSV — comma, tab, semicolon, or pipe separated. Get a clean markdown table LLMs can read, query, and reason over. Auto-detects the delimiter. Handles quoted fields with commas inside.
Table appears here
Paste CSV on the left to start
// why markdown tables
LLMs read tables better than raw CSV.
When you paste raw CSV into a prompt, the LLM sees a stream of comma-separated values with no explicit structure. It has to guess what's a header and what's data. The more columns or quoted fields involved, the more likely it is to misparse a row.
Markdown tables use | as an unambiguous cell separator, with a header row explicitly separated by ---. Every major LLM was trained on markdown and parses it accurately even for complex multi-column data.
For large tables, set the max rows limit to keep your context window focused — send the LLM the most relevant slice, not the entire dataset.
name,age,city Alice,32,New York "Bob, Jr.",28,SF Carol,45,Austin
| name | age | city | | --- | --- | --- | | Alice | 32 | New York | | Bob, Jr. | 28 | SF | | Carol | 45 | Austin |
// faq
Frequently asked questions
What delimiters are supported?
Comma, tab, semicolon, and pipe. Auto-detect picks the most frequent delimiter in the first 2,000 characters — it works for almost every CSV format. Switch to manual if your file uses an unusual delimiter.
Does it handle quoted fields?
Yes. Fields containing commas, newlines, or pipes can be wrapped in double quotes — the parser handles them correctly, including escaped quotes inside fields (e.g. "it's a ""quote""").
What's the max rows limit for?
To keep your context window manageable. A 10,000-row CSV would use tens of thousands of tokens. Set a limit to send only the most relevant rows to the LLM — useful for sampling large datasets.
Can I paste tab-separated data?
Yes — copy directly from Excel or Google Sheets and select 'Tab' delimiter. When you copy a spreadsheet selection, it copies as TSV (tab-separated values) which this tool handles correctly.