Random and Seed

This section describes how to use the random number generator and how to seed it.


All of the data providers in this library are subclasses of the BaseProvider class, which has a random attribute. This attribute is an instance of the Random class from the module random.

If you are creating your own data provider, you should use this random attribute to access the random.

from mimesis import BaseProvider

class MyProvider(BaseProvider):

    class Meta:
        name = 'my_provider'

    def my_method(self):
        return self.random.randint(0, 100)

my_provider = MyProvider()
# Output: 42

See Data Providers for more information about custom providers.

Seeding random


Keep in mind that some methods of some providers cannot be used with seeded providers since their nondeterministic nature.


We support pytest_randomly and its global seed. If you use it during pytest runs, mimesis will have the same seed as shown in your pytest output: Using --randomly-seed=XXX

For using seeded data just pass an argument seed (which can be int, str, bytes, bytearray) to data provider:

from mimesis import Person
from mimesis.locales import Locale

person = Person(locale=Locale.TR, seed=0xFF)
# Output: 'Gizem Tekand'

Reseeding random

To reseed the random number generator, you can use the reseed() method of the BaseProvider class

from mimesis import Person
from mimesis.locales import Locale

person = Person(Locale.EN, seed='Wow.')

# Output: 'Fausto'


# Output: 'Fausto'

If you want to use the same seed for all your data providers, then using Generic() is your option:

from mimesis import Generic
from mimesis.locales import Locale

generic = Generic(Locale.EN, seed='Wow. Much seed. Much random.')

# Output: 'Donn'
# Output: '2021-09-04'
# Output: 'platform'

Probability and Weighted Choice

You might wish to produce data with a specific likelihood of appearing.

To illustrate, suppose you aim to produce random complete names for males and females, but with a greater likelihood of female names being generated.

Here’s one approach to accomplish this:

from mimesis import Person, Locale, Gender

person = Person(Locale.EN)

for _ in range(10):
    full_name = person.full_name(
                Gender.MALE: 0.2,
                Gender.FEMALE: 0.8,


Chieko Flynn
Jannet William
Rozella Church
Dorotha Flowers
Annis Garcia
Trudie Mcclure
Alfonzo Cox
Elsy Bridges
Darby Bates
Serita Cleveland


We are accessing random attribute of the Person class to ensure same seed.