Releases

First stable release of csv-to-json, a lightweight CSV-to-JSON converter for AI agents and developer workflows.

What it does

Turns CSV data into JSON — either as a Python library or a CLI — with zero external dependencies (Python 3.10+ standard library only).

Highlights

• Type inference — numeric values become int / float, true/false/yes/no become booleans, empty cells become null, everything else stays a string. Disable with --no-type-inference when you need round-trip fidelity.
• Nested key expansion — headers like user.address.city collapse into nested JSON objects when you pass --expand-keys.
• Streaming — stream_csv_to_json() yields records one at a time via a generator, so multi-GB files don't blow up memory.
• Flexible input — file paths, raw CSV strings, or stdin. Custom delimiters via -d (comma, tab, pipe, whatever).
• Two ways to use it — python src/csv_converter.py data.csv -o out.json from the shell, or from src.csv_converter import csv_to_json in your own code.

Install

No pip step needed — clone the repo and run. See the README for full CLI flags and library examples.

Compatibility

Python 3.10+. No external packages.

License

MIT.