minson 0.1.1

Minify JSON to reduce LLM API token costs

minson

minify + json — strip

Install

pip install someson

Usage

In Python

from minson import minify

# From a dict/list
minify({"name": "Alice", "age": 30})
# → '{"name":"Alice","age":30}'

# From a JSON string
minify('{ "name": "Alice",  "age": 30 }')
# → '{"name":"Alice","age":30}'

# From a file
minify("/path/to/data.json")
# → '{"name":"Alice","age":30}'

Check your savings

from minson.core import token_savings

stats = token_savings({"name": "Alice", "age": 30, "city": "London"})
# {
#   'original_chars': 48,
#   'minified_chars': 38,
#   'chars_saved': 10,
#   'estimated_pct_saved': 20.8,
#   'estimated_tokens_saved': 2
# }

Minify a file (and write output)

from minson.core import minify_file

minify_file("data.json", "data.min.json")

From the command line

# Print minified JSON
someson data.json

# Write to file
someson data.json --output data.min.json

# Show token savings stats
someson data.json --stats

# Pipe from stdin
echo '{"a": 1, "b": 2}' | someson -

Why?

LLM APIs charge per token. Pretty-printed or verbose JSON wastes tokens on whitespace that carries no meaning. Minifying before sending data to the API is one of the simplest, zero-effort ways to cut costs.

Format	Relative token cost
Pretty JSON	100% (baseline)
Minified JSON	~69–80%
YAML	~82%
XML	~114%

License

MIT