How to use this tool
- Paste the delimited text or choose a file and encoding.
- Select the delimiter and whether the first record is a header.
- Convert; enable typed values only when their meaning is clear.
How the operation works
A record parser handles quoted separators, escaped quotes and multiline fields. With headers enabled, each record becomes an object; without headers it becomes an array. Optional typing recognizes exact JSON number tokens, true, false and null. Values such as 00123 stay strings because they are not JSON number tokens.
CSV records are not always text lines
A newline inside a quoted field belongs to that field. Splitting source text at every line break would incorrectly create extra records. The parser reads quoted fields, embedded delimiters and doubled quotes before checking record widths. A malformed quote or inconsistent number of fields stops the conversion.
Choose comma, semicolon, tab or pipe deliberately. A semicolon-delimited export can otherwise appear to be one column. The tool does not guess the delimiter or repair damaged quoting. Select the source file’s encoding before loading it; a failed decode asks you to choose a supported encoding rather than silently substituting damaged characters.
Headers determine the JSON shape
With First record contains headers enabled, each data record becomes an object. Header names must be nonempty and unique, preventing accidental field overwrites. Without a header, each record becomes an array of fields and the first record remains data.
If an export has a report title or explanatory rows before its real header, remove those deliberately before conversion. Use CSV Viewer to check the structure, or CSV Column Editor to keep, rename and reorder fields. Conversion does not infer which row was intended as the header.
Strings by default; typed values by choice
Default output keeps each field as a JSON string. This preserves identifiers, leading zeros and text such as true or null without guessing their meaning. Enabling type conversion recognizes exact JSON number spellings, true, false and null. The field 00123 remains a string because it is not a valid JSON number token.
CSV: id,active
00123,true
Default: {"id":"00123","active":"true"}
Typed: {"id":"00123","active":true}Numeric tokens are emitted directly rather than rounded through a JavaScript number. A downstream application can still round them if it uses limited-precision parsing. Inspect that application’s requirements before treating a long account identifier as a number. Whitespace within a field is significant and is not automatically trimmed.
Inspect and export the converted JSON
Copy and Download include all accepted records, not only a screen preview. Use JSON Formatter for alternate indentation, JSON Viewer for branches, or JSON Compare to compare converted exports by structure.
This operation checks table structure, not whether an email address, date or business value is correct. Limits are 5 MiB per input, 50,000 records and 200 columns. It does not expand a cell containing JSON into a nested object automatically. Keep the original export so you can revisit delimiter, header and typing decisions.
A worked example
id,name 00123,Ada
[{"id":"00123","name":"Ada"}]
Duplicate or blank header names are rejected rather than overwriting fields. A CSV value spelled null is text by default; enabling typing changes its meaning.
Processing and privacy
CSV to JSON parses the table in page memory; the file or pasted rows are not uploaded. Input stays in browser memory until you reset or leave the page. Copy and download run only when you choose them. No tool input is sent to a processing service. Website pages and scripts still make ordinary hosting requests; the privacy policy separates those from tool processing.
Frequently asked questions
Will numeric identifiers be rounded?
Default string mode preserves them. Typed numeric tokens are serialized directly, but the application reading that JSON must also support their precision.
What if headers are duplicated?
They are rejected rather than overwriting fields. Fix the source or rename in Column Editor first.
Can a cell contain JSON?
It stays a string unless you deliberately parse it elsewhere. Conversion does not expand cell JSON into objects.
TSV as well as CSV?
Yes—select tab as the delimiter and the correct header mode.
How this tool was checked
CSV to JSON was checked with quoted delimiters, embedded newlines, duplicate and blank headers, typed versus string values and numeric identifiers that must not be rounded. See the testing method or report a failure.