CSV to JSON Converter — Convert a Table to JSON Online

Paste a CSV (or drop a .csv file) into the input panel and the tool detects the delimiter automatically, parses each row into a JSON object using the header line as keys, and shows the output instantly. Switch the output format to JSON array, JSONL, or SQL INSERT to get exactly the shape you need.

Yes. The parser follows RFC 4180 — fields wrapped in double quotes can contain commas, newlines and even literal quotes (escaped as "") without breaking parsing. So a value like "Smith, John" remains a single cell, not two.

The tool counts tabs, commas, semicolons and pipes on the first non-empty line and picks the most-frequent one. If you have an unusual or ambiguous file, you can override auto-detect and select tab (TSV), comma, semicolon, pipe or Markdown table manually.

Switch the output format to SQL INSERT and enter a table name. The tool sanitises column names to valid identifiers (replacing spaces, dashes, etc. with underscores) and produces one INSERT per row with properly-escaped string literals, numeric values, and NULL for empty cells. Useful for seeding databases or migrating data.

JSONL (JSON Lines) puts one JSON object per line, with no surrounding array. It is the standard format for streaming data into BigQuery, Snowflake, Logstash, Elasticsearch and most machine-learning pipelines. Pick JSONL when each row will be ingested independently.

Yes. Values matching integer or decimal patterns are emitted as JSON numbers, the literals "true"/"false" become booleans, and empty cells become null. There is no option to keep numeric-looking values as text, so a value such as 02134 becomes the number 2134; add a non-numeric character to values that must stay strings before converting.

Yes. Select "Markdown table" as the delimiter and paste a standard pipe-delimited Markdown table (header row, separator row, body rows). The separator line is automatically skipped and cell whitespace is trimmed.

No. Parsing and conversion run in your browser rather than through the Pixoate processing API. Follow your organization's data-handling rules before opening confidential spreadsheets in any web application.

When checked, the tool treats your table's first row as column names and uses them as the JSON object keys (or generates col_1, col_2... for SQL if you leave it blank). Uncheck it if your data has no header row — a plain list of rows — and the tool will instead output raw JSON arrays or auto-name columns col_1, col_2, etc. for SQL INSERT statements.

Switch Output to SQL INSERT and a Table name field appears where you can type your target table's name, such as customers or order_items. The tool automatically sanitizes whatever you enter — replacing spaces and symbols with underscores and prefixing a leading digit — so the generated SQL always has a valid identifier even if your input is messy.

Uncheck 'First row is header' if your export has no column names, set Output to SQL INSERT, and type your real table name in the Table name field. The tool will auto-generate col_1, col_2... column names, coerce numbers and booleans automatically, and produce one properly-escaped INSERT statement per row that you can paste straight into your database client.

JSON (arrays) drops the field names and outputs each row as a plain list of values in column order — for example ["Alice", 34, "Berlin"] instead of {"name":"Alice","age":34,"city":"Berlin"}. It still respects the 'First row is header' checkbox, using that row only to know how many columns to expect rather than as keys. Choose it when the receiving code (a charting library, a positional database driver, or a spreadsheet import) reads values by position rather than by field name; otherwise stick with JSON (objects).

Try sample fills the input panel with a short three-row name/age/city table (tab-separated) so you can immediately see how the delimiter detection, header row and each output format behave before you paste real data. Once there's content in the box, a Clear button appears — next to the Input label in the editor view, and again as a full-width button in the Settings panel — that wipes the textarea back to empty in one click instead of selecting and deleting the text manually.

No — Table to JSON / SQL runs entirely in your browser, so what you paste never leaves the machine you are on. That matters if you are checking something that contains a token, a key or customer data. Closing the tab is enough to clear it.

The error points at the first position that could not be parsed, which is usually just after the real mistake rather than on it — a missing comma, an unclosed quote or bracket, or a smart quote pasted in from a document instead of a straight one. Check the line above the reported position first.

Table to JSON / SQL supports free standard use within the published processing allowances. Free downloads are clean — no watermark. No signup is required for standard tool use. Photo effects and AI edits have a web-size limit on Free, with the original-resolution result available through Premium or one credit. Crop, resize, rotate, compress and convert stay full size. Published processing allowances and file limits still apply.

No — Table to JSON / SQL runs in the browser on desktop and phone. There is nothing to install, no trial to start, and no license to buy before you can see whether it does what you need. Try it on the file you are actually stuck with rather than a sample.

Cells made only of digits are converted to numbers, and a number has no leading zeros, so 02134 becomes 2134. The tool has no option to keep a column as text. Convert the table, then restore those fields as strings in your code, or add a letter prefix to the values before converting.

In the JSON object for each row, a key can appear only once, so when two headers share a name the later column replaces the earlier one without a warning. Rename one header before converting, for example to phone_home and phone_work, so every column keeps its own key.

A double quote opens a quoted section wherever it appears, even in the middle of an unquoted value such as 5" long. Everything after it, including delimiters and line breaks, is read as part of that value until the next quote. Write the quote as two quotes inside a quoted value to keep rows apart.

With a comma delimiter, the comma in 1,000 is read as a separator, so the value becomes 1 and 000 in separate columns. Put the whole value in double quotes in the source, or choose a delimiter that does not appear inside your numbers, such as a semicolon or a tab.

Column names are written exactly as sanitized, without quotes or brackets, so a reserved word such as order, group or select is read as part of the statement. Rename those headers before converting, or add the quoting your database uses to the column list in the output.

An uploaded file must be 5 MB or smaller, and a larger one is refused with a message that keeps your current text. Pasted text is not subject to that upload limit, but a very large paste depends on your browser and device memory. For large exports, split the file into parts.