JSON Serialization

JSON serialization is the process of converting Python data into JSON.

diagram what a view is allowed to return mermaid

  flowchart TD
  A["a view returns …"] --> B["a string"]
  A --> C["a dict"]
  A --> D["a list"]
  A --> E["a tuple"]
  A --> F["a Response"]
  B --> B2["200, text/html"]
  C --> C2["200, application/json"]
  D --> D2["200, application/json"]
  E --> E2["(body, status) or (body, status, headers)"]
  F --> F2["used as-is"]
  C2 --> G["jsonify() is only needed for older Flask, or to set options"]
  D2 --> G

Flask accepts several shapes and converts each into a Response. A dict or a list becomes JSON automatically, so jsonify is no longer required for the common case. A tuple lets you set the status and headers without building a Response by hand. Anything else has to be a Response already.

What works out of the box

JSON supports:

strings
numbers
booleans
null
arrays
objects

In Python terms:

str, int, float, bool, None
list/tuple
dict

What does NOT serialize automatically

Common non-serializable types:

datetime
database model objects
Decimal

You must convert them.

Serialization strategies

1) Manual dict conversion

python

user_dict = {"id": user.id, "username": user.username}
return user_dict

2) Helper method on model

python

class User(db.Model):
    # ...
    def to_dict(self):
        return {"id": self.id, "username": self.username}

3) Schema libraries (later)

For larger APIs, people use:

Marshmallow
Pydantic

This tutorial track keeps it simple: manual to_dict() is fine.

Error response shape (recommended)

Try to keep error responses consistent:

json

{ "error": "validation_error", "message": "Email is required" }

Consistency makes frontend/mobile clients much easier to build.

🧪 Try It Yourself

Exercise 1 – Create a Flask App

Exercise 2 – Dynamic Route

Exercise 3 – Return JSON

pch.coffeeTagline

pch.coffeeCta

pch.feedbackHeading

pch.feedbackSubheading