json library

Official documentation

The json module reads and writes JSON ("JavaScript Object Notation") data — a plain-text format built on nested dicts and lists, which makes it the standard way structured data moves between programs, files, and web APIs. Every example below actually runs in your browser: Pyodide gives each page its own in-memory filesystem, so open() works exactly like it would on a real computer, just without anything being saved outside this page.

Install

json ships with Python's standard library — nothing to install.

Import

The whole module is used through the json. prefix, so a plain import is all you need.

import json

Tool	Converts	Use it for
json.dump()	A Python object → an open file	Saving data to disk
json.load()	An open file → a Python object	Loading data from disk
json.dumps()	A Python object → a string	Sending data somewhere, like an API request body
json.loads()	A string → a Python object	Parsing JSON text you already have in memory

Writing JSON files

json.dump() writes a Python object straight to an open file — a dict becomes a JSON object, a list becomes a JSON array, automatically.

import json

snake = {"species": "ball", "length_ft": 4.5, "venomous": False}

with open("snake.json", "w") as file:
    json.dump(snake, file)

print("wrote snake.json")

Pretty-printing with indent

json.dump() writes everything on one line by default. Pass indent= to spread it across multiple, readable lines instead — handy when you'll open the file yourself later.

with open("snake.json", "w") as file:
    json.dump(snake, file, indent=2)

Run a writing JSON example

All the examples above, combined into one script:

import json

snake = {"species": "ball", "length_ft": 4.5, "venomous": False}

with open("snake.json", "w") as file:
    json.dump(snake, file)

print("wrote snake.json")

with open("snake.json", "w") as file:
    json.dump(snake, file, indent=2)

Reading JSON files

json.load() reads a file and reconstructs the original Python object — a JSON object comes back as a dict, a JSON array comes back as a list, with numbers and booleans already converted to int/float/bool instead of strings.

import json

with open("snake.json") as file:
    data = json.load(file)

print(data["species"])
print(data["length_ft"])

Nested data

Unlike a CSV file, which is strictly flat rows and columns, JSON can nest a list or another object inside a value — so one record can hold something like a snake's full sighting history, not just single values per column.

snake = {
    "species": "ball",
    "length_ft": 4.5,
    "sightings": ["2024-03-15", "2024-06-02"]
}

with open("snake.json", "w") as file:
    json.dump(snake, file)

with open("snake.json") as file:
    data = json.load(file)

print(data["sightings"][0])    # "2024-03-15"

Run a reading JSON example

All the examples above, combined into one script:

import json

snake = {"species": "ball", "length_ft": 4.5, "venomous": False}

with open("snake.json", "w") as file:
    json.dump(snake, file)

with open("snake.json") as file:
    data = json.load(file)

print(data["species"])
print(data["length_ft"])

snake = {
    "species": "ball",
    "length_ft": 4.5,
    "sightings": ["2024-03-15", "2024-06-02"]
}

with open("snake.json", "w") as file:
    json.dump(snake, file)

with open("snake.json") as file:
    data = json.load(file)

print(data["sightings"][0])

Working with strings instead of files

json.dumps()/json.loads() do the same conversion as dump()/load(), but to and from a string in memory rather than a file — the pair to reach for when the JSON is coming from somewhere other than disk, like an API response. The requests library's own .json() method — covered on that page — is really just calling json.loads() on the response text for you.

import json

snake = {"species": "ball", "length_ft": 4.5}

text = json.dumps(snake)    # '{"species": "ball", "length_ft": 4.5}'
data = json.loads(text)     # back to {"species": "ball", "length_ft": 4.5}

Run a strings example

All the examples above, combined into one script:

import json

snake = {"species": "ball", "length_ft": 4.5}

text = json.dumps(snake)
print(text)

data = json.loads(text)
print(data)