Persistent Data link

Ren'Py supports persistent data, saved data that is not associated with a particular point in a game. Persistent data is accessed through fields of the persistent object, which is bound to the variable persistent.

All data reachable through fields on persistent is saved when Ren'Py terminates, or when renpy.save_persistent() is called. Persistent data is loaded when Ren'Py starts, and when Ren'Py detects that the persistent data has been updated on disk.

The persistent object is special in that an access to an undefined field will have a None value, rather than causing an exception. If something other than None is to be the default of a persistent value, the default statement should be used:

default persistent.main_background = "princess_not_saved"

An example use of persistent is the creation of an unlockable image gallery. This is done by storing a flag in persistent that determines if the gallery has been unlocked, as in

label gallery:

    if not persistent.gallery_unlocked:
        show background
        centered "You haven't unlocked this gallery yet."
        $ renpy.full_restart()

    # Actually show the gallery here.

When the user gets an ending that causes the gallery to be unlocked, the flag must be set to True.

$ persistent.gallery_unlocked = True

As persistent data is loaded before init python blocks are run, persistent data should only contain types that are native to Python or Ren'Py. Alternatively, classes that are defined in python early blocks can be used, provided those classes can be pickled and implement equality.

Fields starting with two underscores (__) are supported, but will receive the name-mangling effect described in this section, which means they will be specific to the file they're defined in. This implies that if the file is renamed between two releases, access to the value that field had in the previous release will be broken.

In addition to this, these fields are not affected by the persistent._clear() method.

As a reminder, fields starting with a single underscore _ are reserved and should not be used.

Merging Persistent Data link

There are cases where Ren'Py has to merge persistent data from two sources. For example, Ren'Py may need to merge persistent data stored on a USB drive with persistent data from the local machine.

Ren'Py does this merging on a field-by-field basis, taking the value of the field that was updated more recently. In some cases, this is not the desired behavior. In that case, the renpy.register_persistent() function can be used.

For example, if we have a set of seen endings, we'd like to take the union of that set when merging data.

init python:
    if persistent.endings is None:
        persistent.endings = set()

    def merge_endings(old, new, current):
        return current

    renpy.register_persistent('endings', merge_endings)

Persistent Functions link

persistent._hasattr(field_name) link

Tests whether the field_name persistent field has been set or not. This allows you to distinguish fields that have been explicitly set to None from fields that have never been set.

persistent._clear(progress=False) link

Resets the persistent data, except for fields starting with __.


If true, also resets progress data that Ren'Py keeps.

Note that this will delete all persistent data, and will not re-apply defaults until Ren'Py restarts.

renpy.register_persistent(field, func) link

Registers a function that is used to merge values of a persistent field loaded from disk with values of current persistent object.


The name of a field on the persistent object.


A function that is called with three parameters, old, new, and current:


The value of the field in the older object.


The value of the field in the newer object.


The value of the field in the current persistent object. This is provided for cases where the identity of the object referred to by the field can't change.

The function is expected to return the new value of the field in the persistent object.

renpy.save_persistent() link

Saves the persistent data to disk.

Multi-Game Persistence link

Multi-Game persistence is a feature that lets you share information between Ren'Py games. This may be useful if you plan to make a series of games, and want to have them share information.

To use multipersistent data, a MultiPersistent object must be created at init time (preferably using define). The user can then update this object, and save it to disk by calling its save method. Undefined fields default to None. To ensure the object can be loaded again in a different game, we strongly advise against storing instances of user-defined types in the object.

class MultiPersistent(key, save_on_quit=False) link

Creates a new MultiPersistent object. This should only be called at init time, and it returns a new MultiPersistent with the given key.


The key used to to access the multipersistent data. Games using the same key will access the same multipersistent data.


If it is True, this object will be automatically saved when Ren'Py terminates.

save() link

Saves the multipersistent data to disk. This must be called after the data is modified.

As an example, take the first part of a two-part game:

define mp = MultiPersistent("")

label start:

    # ...

    # Record the fact that the user beat part 1.

    $ mp.beat_part_1 = True

    e "You beat part 1. See you in part 2!"

And the second part:

define mp = MultiPersistent("")

label start:

    if mp.beat_part_1:
        e "I see you've beaten part 1, so welcome back!"
        e "Hmm, you haven't played part 1, why not try it first?"