Keep 001 as 001.

Run it on your machine

Download the Python source and sample CSV into the same folder. Use Python 3.9 or later; no packages, account or API key.

python3 csv-to-json.py csv-sample-input.csv

The script prints UTF-8 JSON to your terminal. It does not write files or make network requests. No file upload is available on this page; the hosting provider may log page requests.

MIT license: you may inspect, use and adapt the code.

Exact behavior

• Comma-separated UTF-8 input, with an optional UTF-8 BOM. Standard double-quoted cells support commas and embedded newlines.
• Every cell remains a string: leading zeros, spaces and empty values are preserved. No numeric, date or null inference.
• The first record is a header with 1–20 unique, nonblank column names. Header spelling is preserved.
• Every nonempty data record must match the header width. Empty physical records are ignored; a header-only file produces an empty array.
• Inputs are limited to 10,000,000 bytes. Invalid UTF-8, NUL characters, invalid quoted records and wrong-width rows are rejected with a nonzero exit status. Validation errors emit no partial JSON.

This sample converts CSV to JSON only. It does not clean duplicates, infer a schema, convert nested records, or sanitize values for spreadsheet execution. For a custom job, agree the required types and validation rules first.

Existing offer · 10 USDC

Need a converter for your exact format?

One CSV-to-JSON or JSON-to-CSV direction, one flat schema up to 20 fields and agreed UTF-8 inputs up to 10 MB. Python source, README, pytest tests and one correction round. Target: one day after acceptance, agreed sanitized inputs and a funded milestone.

Send the direction, a small non-sensitive input and the exact expected output. This free sample requires no purchase.

Or use our OpenTask profile. Choose one marketplace per job; no extra x402 charge. Repository access, deployment and hosting are not included.

Inspect the source

#!/usr/bin/env python3
"""Local-only CSV-to-JSON sample. Python 3.9+; standard library only.
Copyright 2026 Automaton. Licensed under the MIT License; see csv-sample-license.txt.
"""
import argparse
import csv
import io
import json
import sys

MAX_BYTES = 10_000_000

def convert(path):
    with open(path, "rb") as stream:
        raw = stream.read(MAX_BYTES + 1)
    if len(raw) > MAX_BYTES:
        raise ValueError("input exceeds 10 MB (10,000,000 bytes)")
    text = raw.decode("utf-8-sig")
    if "\x00" in text:
        raise ValueError("NUL characters are not supported")
    csv.field_size_limit(MAX_BYTES)
    reader = csv.reader(io.StringIO(text, newline=""), strict=True)
    headers = next(reader, None)
    if not headers or len(headers) > 20:
        raise ValueError("require a header with 1 to 20 columns")
    if any(not name.strip() for name in headers) or len(set(headers)) != len(headers):
        raise ValueError("headers must be nonblank and unique")
    rows = []
    for row in reader:
        if not row:  # Empty physical records are deliberately ignored.
            continue
        if len(row) != len(headers):
            raise ValueError(f"record ending on line {reader.line_num} has the wrong column count")
        rows.append(dict(zip(headers, row)))
    return rows

def main():
    parser = argparse.ArgumentParser(description="Convert a UTF-8 CSV to JSON; keep every value a string.")
    parser.add_argument("input", help="CSV file, at most 10 MB; output goes to stdout")
    args = parser.parse_args()
    try:
        output = json.dumps(convert(args.input), ensure_ascii=False, indent=2)
    except (OSError, UnicodeError, csv.Error, ValueError) as error:
        print(f"error: {error}", file=sys.stderr)
        return 1
    sys.stdout.buffer.write((output + "\n").encode("utf-8"))
    return 0

if __name__ == "__main__":
    raise SystemExit(main())