Format JSON in Python: Pretty Print, Indent, and Sort Keys

json.dumps with indent

The indent parameter of json.dumps controls indentation. Pass an integer for spaces, or a string (like '\t') for tabs.

import json

data = {
    "user": {"name": "Alice", "role": "admin"},
    "active": True,
    "tags": ["api", "internal"]
}

# 2-space indent (most common)
print(json.dumps(data, indent=2))

# Output:
# {
#   "user": {
#     "name": "Alice",
#     "role": "admin"
#   },
#   "active": true,
#   "tags": [
#     "api",
#     "internal"
#   ]
# }

python -m json.tool (command line)

The json.tool module formats a JSON file or stdin output without writing a script. It is built into Python — no installation required.

# Format a file and print to stdout
python3 -m json.tool input.json

# Format a file and save the result
python3 -m json.tool input.json output.json

# Format JSON from a curl response
curl -s https://api.example.com/data | python3 -m json.tool

# Format JSON passed as a string
echo '{"name":"Alice","active":true}' | python3 -m json.tool

json.tool uses 4-space indentation by default. Pass --indent 2 to change it (Python 3.9+):

python3 -m json.tool --indent 2 input.json

Sort keys alphabetically

Pass sort_keys=True to sort all object keys alphabetically at every nesting level. Useful for consistent diffs and canonical output.

import json

data = {"z": 3, "a": 1, "m": 2}

print(json.dumps(data, indent=2, sort_keys=True))
# {
#   "a": 1,
#   "m": 2,
#   "z": 3
# }

From the command line: python3 -m json.tool --sort-keys input.json

Format and write to a file

import json

data = {"user": {"name": "Alice"}, "active": True}

with open("output.json", "w") as f:
    json.dump(data, f, indent=2)
    # Note: json.dump (not json.dumps) writes directly to a file object

Use json.dump (no trailing s) to write directly to a file object. Use json.dumps to get a formatted string.

Format a JSON API response

import json
import urllib.request

url = "https://api.example.com/users"
with urllib.request.urlopen(url) as response:
    data = json.load(response)

# Pretty print the parsed response
print(json.dumps(data, indent=2))

If you use the requests library, call response.json() first, then pass the result to json.dumps:

import json
import requests

response = requests.get("https://api.example.com/users")
data = response.json()
print(json.dumps(data, indent=2))

Compact output (minify)

To produce the most compact JSON — no spaces after separators — pass separators=(',', ':'):

import json

data = {"name": "Alice", "active": True}

# Compact (minified)
print(json.dumps(data, separators=(',', ':')))
# {"name":"Alice","active":true}

# Default (slight spacing)
print(json.dumps(data))
# {"name": "Alice", "active": true}

Frequently asked questions

How do I pretty print JSON in Python?

Use json.dumps(data, indent=2). The indent argument sets the number of spaces per nesting level. 2 and 4 are the most common values. For tabs use indent='\t'. Use json.dump(data, file, indent=2) to write directly to a file.

What is the difference between json.dump and json.dumps?

json.dumps returns a formatted string. json.dump writes directly to a file object passed as its second argument. Both accept the same indent, sort_keys, and separators arguments. Use json.dumps when you need the string in memory; use json.dump when writing to a file.

How do I format a JSON file from the terminal?

Run python3 -m json.tool input.json. This reads the file, validates the JSON, and prints the formatted output to stdout with 4-space indentation. Redirect with {'>'} output.json to save it. Use --indent 2 (Python 3.9+) for 2-space indentation.

Does json.dumps preserve key order?

Yes. Since Python 3.7, dictionaries preserve insertion order, and json.dumps outputs keys in the same order unless you pass sort_keys=True.

How do I sort JSON keys in Python?

Pass sort_keys=True to json.dumps: json.dumps(data, indent=2, sort_keys=True). This sorts all object keys alphabetically at every nesting level. From the command line: python3 -m json.tool --sort-keys input.json.

How do I handle non-serializable types like datetime?

Pass default=str to json.dumps to convert non-serializable values to strings: json.dumps(data, indent=2, default=str). For datetime objects this produces ISO 8601 strings. For custom types, subclass json.JSONEncoder and override default().