TypeError: Object of type X is not JSON serializable — how to fix

TypeError: Object of type datetime is not JSON serializable
TypeError: Object of type Decimal is not JSON serializable
TypeError: Object of type ndarray is not JSON serializable

Quick fix: Convert the value before dumping, or pass a fallback: json.dumps(data, default=str). For more control write a default function that handles datetime (.isoformat()), Decimal (float/str), set (list) and numpy (.tolist() / .item()).

What it means

Python's json module maps a fixed set of types to JSON:

Anything else raises TypeError: Object of type <name> is not JSON serializable. The type name in the message tells you exactly which value to convert.

Fix 1: quick fallback with default=str

import json, datetime
data = {"created": datetime.datetime.now()}
json.dumps(data, default=str)
# '{"created": "2026-11-04 09:30:00.123456"}'

default is called for every object json can't handle. str is fine for logs and debugging, but it turns everything into strings.

Fix 2: a proper default function

import json, datetime, decimal, uuid, dataclasses

def to_json(o):
    if isinstance(o, (datetime.datetime, datetime.date, datetime.time)):
        return o.isoformat()
    if isinstance(o, decimal.Decimal):
        return float(o)          # or str(o) to keep exact precision
    if isinstance(o, (set, frozenset)):
        return list(o)
    if isinstance(o, uuid.UUID):
        return str(o)
    if dataclasses.is_dataclass(o):
        return dataclasses.asdict(o)
    raise TypeError(f"Object of type {type(o).__name__} is not JSON serializable")

json.dumps(data, default=to_json)

Or subclass json.JSONEncoder and pass cls=MyEncoder; both work the same way.

Cause: NumPy and pandas types

numpy.int64, numpy.float32 and ndarray are not Python ints/floats/lists. They often sneak in from df["col"].sum() or arr.max().

import numpy as np
json.dumps({"n": np.int64(3)})            # TypeError: Object of type int64 ...

json.dumps({"n": np.int64(3).item()})     # convert scalar
json.dumps({"a": arr.tolist()})           # convert array

For DataFrames use pandas' own serializer, which handles timestamps and NaN:

df.to_json(orient="records", date_format="iso")

Cause: your own class

class User:
    def __init__(self, name): self.name = name

json.dumps(User("ann"))   # TypeError: Object of type User ...

Return a dict explicitly (json.dumps(vars(user))), make it a @dataclass and use dataclasses.asdict, or with Pydantic v2 use user.model_dump_json() / user.model_dump(mode="json").

Cause: bytes

bytes aren't JSON. Decode text (b.decode("utf-8")) or base64-encode binary data:

import base64
base64.b64encode(raw).decode("ascii")

In web frameworks

• Flask: jsonify handles dates in recent versions; for other types set a custom provider (app.json = MyJSONProvider(app) with a default method).
• FastAPI: return Pydantic models or use fastapi.encoders.jsonable_encoder(obj).
• Django: use django.core.serializers.json.DjangoJSONEncoder (handles datetime, Decimal, UUID): json.dumps(data, cls=DjangoJSONEncoder).

Last updated 2026-11-04.