Refactor: separate camera state from renderer-owned rasterization (cairo) + consolidate renderer related modules by behackl · Pull Request #4989 · ManimCommunity/manim

behackl · 2026-09-04T13:51:23Z

Overview: What does this pull request change?

This PR makes the camera/renderer ownership boundary explicit: Cairo cameras describe semantic views, while CairoRenderer owns raster targets, PyCairo drawing, composition, static-image reuse, readback, and cleanup.

It also packages concrete implementations under manim.renderer.cairo and manim.renderer.opengl, gives the default Cairo Camera an animatable frame and auto_zoom(), and preserves ThreeD, ZoomedScene, nested-camera, image, point-cloud, background, and static-cache behavior.

This change assumes the renderer-independent writer boundary introduced by #4976. Renderer-owned play loops, clocks, cache decisions, and writer scheduling remain unchanged for the later Manager execution refactor.

Motivation and Explanation: Why and how do your changes improve the library?

Cairo cameras previously combined semantic view state with pixel buffers, PyCairo contexts, drawing, composition, and readback. That made camera movement a special case and made render-resource ownership difficult to reason about.

After this change:

Camera, MovingCamera, MultiCamera, and ThreeDCamera own semantic frame, background, ordering, projection, shading, tracker, and nested-view state;
CairoRenderer and private helpers own all Cairo resources and raster work;
only production modules in the Cairo backend import PyCairo;
nested views use renderer-owned secondary targets, detect cycles, exclude their own display family, and clean up with the parent renderer;
background-colored vector composition reuses a renderer-owned scratch target; and
both backend implementations and OpenGL shader assets live under canonical backend packages.

For ordinary Cairo scenes, camera movement no longer requires a specialized scene class:

class CameraExample(Scene):
    def construct(self):
        square = Square()
        self.add(square)
        self.play(self.camera.frame.animate.move_to(square).scale(0.5))
        self.play(self.camera.auto_zoom([square], margin=1))

MovingCamera and MovingCameraScene are deprecated: use Camera and Scene. ZoomedScene is now a plain Scene subclass.

Breaking changes and migration

Supported top-level imports remain unchanged:

from manim import Camera, CairoRenderer, MovingCamera, MultiCamera, ThreeDCamera

Implementation-module imports are intentionally a clean break; no duplicate compatibility modules are retained.

Module paths

Removed path	Replacement
`manim.camera.camera`	`manim.renderer.cairo.camera`
`manim.camera.moving_camera`	`manim.renderer.cairo.camera`
`manim.camera.multi_camera`	`manim.renderer.cairo.camera`
`manim.camera.three_d_camera`	`manim.renderer.cairo.camera`
`manim.renderer.cairo_renderer`	`manim.renderer.cairo.renderer` or `manim.renderer.cairo`
`manim.renderer.opengl_renderer`	`manim.renderer.opengl.renderer`, `manim.renderer.opengl.camera`, or `manim.renderer.opengl`
`manim.renderer.opengl_renderer_window`	`manim.renderer.opengl.window`
`manim.renderer.shader`	`manim.renderer.opengl.shader`
`manim.renderer.shader_wrapper`	`manim.renderer.opengl.shader_wrapper`
`manim.renderer.vectorized_mobject_rendering`	`manim.renderer.opengl.vectorized_mobject_rendering`

OpenGL shader resources move from manim/renderer/shaders to manim/renderer/opengl/shaders. Plugins loading those resources directly must update their package-relative paths.

Camera construction and raster access

Cairo camera constructors now accept semantic view settings only. The following raster-oriented arguments are removed:

pixel_width, pixel_height, and frame_rate;
image_mode, n_channels, and pixel_array_dtype;
background pixel arrays and cairo_line_width_multiple; and
MovingCamera.fixed_dimension.

Use config.pixel_width, config.pixel_height, and config.frame_rate for render execution settings. Cairo output is standardized on renderer-owned uint8 RGBA targets. Use background_image, background_color, and background_opacity for camera backgrounds. Configure logical view geometry with frame=, frame_width=, frame_height=, or the public camera.frame mobject. Unknown camera keywords now raise TypeError instead of being retained silently.

Camera-owned raster APIs are also removed, including camera.pixel_array, capture_mobjects(), get_image(), Cairo-context caches, and display_* / apply_* drawing helpers. Direct raster users should render through CairoRenderer:

with CairoRenderer(camera=camera) as renderer:
    renderer.render_mobjects(mobjects)
    pixels = renderer.get_frame()
    image = renderer.get_image()

For one-off mobject images, mobject.get_image(camera=camera) manages that temporary renderer automatically.

Custom camera subclasses should implement semantic hooks such as object ordering, point projection, shading, and movement controls. Backend drawing extensions belong in a renderer integration rather than a camera subclass.

Removed camera implementations

MappingCamera, OldMultiCamera, and SplitScreenCamera are removed.

Nonlinear mapping should be applied explicitly to sufficiently subdivided geometry with apply_function() rather than through MappingCamera.
Nested or split views should use MultiCamera with positioned ImageMobjectFromCamera display mobjects.

BackgroundColoredVMobjectDisplayer, previously exported from manim.camera.camera and the top-level manim namespace, is removed without replacement. CairoRenderer composites VMobjects with background images itself; color them with VMobject.color_using_background_image().

ImageMobjectFromCamera remains available, but it is now a geometry-bearing Mobject rather than an AbstractImageMobject. It stores no pixel_array and has no get_pixel_array() API; CairoRenderer renders the source camera into a secondary target and composites it into the display geometry. Existing positioning, display-frame, and resampling usage remains supported.

For ThreeD camera positioning, use camera.frame or camera.frame_center; the private _frame_center mobject is removed.

ThreeDCamera.rotation_matrix is read-only and computed from the current angles when read, so projection queries between frames no longer need a draw first. Change the angle trackers, or override generate_rotation_matrix().

Camera.auto_zoom() raises ValueError for mobjects with neither width nor height unless margin is positive. Before, the frame collapsed to zero size.

CairoRenderer.update_frame() takes excluded_mobjects as a keyword-only argument.

Links to added or changed documentation pages

Further Information and Comments

This PR does not move scheduling or resource lifecycle into Manager, redesign OpenGL context lifetime, or unify Cairo and OpenGL cameras. Tests etc. pass locally.

AI assistance was used during implementation and review. The resulting design, code, documentation, and tests were reviewed by the author.

nikolajmunk

Did a superficial deslopping pass of some of the documentation and comments.

More generally, I think it would be useful to consciously choose some language for the terms used in a camera/render context. I don't like "logical/semantic" "frame/view". I would suggest that we adopt camera frame (or just frame when the context is clear) to describe the rectangle which controls the camera view (the position, angle, and size of the view), and I would suggest that the rectangle which defines the output image is called the viewport - in graphics engine circles, this is the usual nomenclature for the "screen rectangle". Another choice might be canvas but that doesn't work quite as well for 3D.

So for example, you might say "When both the width and height of the camera frame are provided, you must ensure that the pixel dimensions of the viewport have the same aspect ratio of the frame, otherwise (bad things happen)."

Happy to discuss more! I would also be happy to give the new camera guide a more thorough rewrite in a later PR once things settle down a bit.

nikolajmunk

Everything reads much more nicely now! I nevertheless went through and remarked on a few LLM-isms.

As a general note, I'm not personally convinced that "semantic (camera) view" and "owned" are useful terms to the reader.

GniLudio

When I try to run the tests, I get an error:

> pytest .
Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "D:\Documents\manim\.venv\Scripts\pytest.exe\__main__.py", line 10, in <module>
    sys.exit(console_main())
             ~~~~~~~~~~~~^^
  File "D:\Documents\manim\.venv\Lib\site-packages\_pytest\config\__init__.py", line 223, in console_main
    code = main()
  File "D:\Documents\manim\.venv\Lib\site-packages\_pytest\config\__init__.py", line 199, in main
    ret: ExitCode | int = config.hook.pytest_cmdline_main(config=config)
                          ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_hooks.py", line 512, in __call__
    return self._hookexec(self.name, self._hookimpls.copy(), kwargs, firstresult)
           ~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_manager.py", line 120, in _hookexec
    return self._inner_hookexec(hook_name, methods, kwargs, firstresult)
           ~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_callers.py", line 167, in _multicall
    raise exception
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_callers.py", line 121, in _multicall
    res = hook_impl.function(*args)
  File "D:\Documents\manim\.venv\Lib\site-packages\xdist\plugin.py", line 223, in pytest_cmdline_main
    auto_num_cpus = config.hook.pytest_xdist_auto_num_workers(config=config)
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_hooks.py", line 512, in __call__
    return self._hookexec(self.name, self._hookimpls.copy(), kwargs, firstresult)
           ~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_manager.py", line 120, in _hookexec
    return self._inner_hookexec(hook_name, methods, kwargs, firstresult)
           ~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_callers.py", line 167, in _multicall
    raise exception
  File "D:\Documents\manim\.venv\Lib\site-packages\pluggy\_callers.py", line 121, in _multicall
    res = hook_impl.function(*args)
  File "D:\Documents\manim\.venv\Lib\site-packages\xdist\plugin.py", line 23, in pytest_xdist_auto_num_workers
    count = psutil.cpu_count(logical=use_logical) or psutil.cpu_count()
            ^^^^^^^^^^^^^^^^
AttributeError: module 'psutil' has no attribute 'cpu_count'

behackl · 2026-10-29T11:22:52Z

When I try to run the tests, I get an error:
[...]
AttributeError: module 'psutil' has no attribute 'cpu_count'

Hmm, nothing here changes the way our tests are setup; this is unchanged from main. Maybe try uv sync to fix a potentially broken psutils?

Retire unused nested targets, release retained buffers, and guard closed operations. Reject unused renderer keywords and build dispatch tables once per draw.

Expose Scene.get_image() and Scene.show() through Manager without stepping animations or emitting movie frames. Preserve active raster targets and capture OpenGL meshes on the context-owning thread.

Compute default frame bounds directly without a semantic cache, preserve custom frame hooks, and share perspective scaling between axes. Add mutation, boundary, and projection parity tests.

Co-authored-by: nikolajmunk <28557236+nikolajmunk@users.noreply.github.com>

Address review feedback on camera terminology, backend controls, custom camera selection, snapshot behavior, and inset views. Preserve the earlier camera test cleanup.

CairoRenderer and OpenGLRenderer are public and referenced from the guides and docstrings, but their modules were missing from the reference, so those links could not resolve.

Remove tests that repeat others or only check implementation details: a nine-way closed-target matrix on a private class, weakref checks of private buffers, scratch-target reuse, nested target sizing and subtree bookkeeping read from private attributes, a mocked check of draw arguments, the re-derived projection formula, trivial frame and constructor echoes, show() delegating to get_image(), and a slow CLI copy of the in-process square-geometry test. Trim the square-geometry and closed-renderer matrices to cases that differ.

The context was scaled to the pixel size and then given a full matrix, which replaces the scale. Pointed out in review.

update_frame took **kwargs only to read excluded_mobjects from them (as main's camera did); a keyword-only parameter says so, and other keywords are rejected instead of ignored.

The two loaders of Cairo background images (the camera's background and a vectorized mobject's background image) repeated the same open, convert, resize code, and each target kept its own file cache: every camera view's target read the files again. One function now loads an image at a size, and the renderer's targets share one cache of files, by path and size.

ThreeDCamera stored its rotation matrix and refreshed it only when a frame was prepared, so projections between frames (project_point after set_phi, or get_mobjects_to_display, which reads get_rotation_matrix) used the matrix of the previous frame. The matrix is now computed from the angle trackers when read; _drawing() still computes the view once per frame, so drawing costs the same. rotation_matrix is a read-only property, reset_rotation_matrix does nothing, and the _prepare_for_render hook that only refreshed the matrix is gone. Found in review.

Every camera has an animatable frame and auto_zoom() now, so these classes add nothing: MovingCamera is a Camera, and MovingCameraScene a Scene with the default camera. Both are deprecated with manim.utils.deprecation and log a warning when created. ZoomedScene, the init template and the examples use Scene and Camera, so they do not warn. Suggested in review.

Title "Working with cameras"; wording. The camera movement examples are rendered (manim directives) so readers see the result. Scenes that choose a camera take camera_class as a keyword with a default, so callers and subclasses can still choose another one. The OpenGL thread note and a redundant renderer hint are gone. In the deep dive, the camera-view step is split in two, and the frame array is described as row-major. ImageMobjectFromCamera says why it is not an AbstractImageMobject: it holds no pixels of its own.

GniLudio · 2026-11-07T08:19:30Z

+        new_height = abs(top - bottom)
+        target = self.frame.animate if animate else self.frame
+        if new_width / self.frame.width > new_height / self.frame.height:
+            return target.set_x(x).set_y(y).set(width=new_width + margin)


Shouldn't it be new_width + 2 * margin?
I would expect the margin to be on both sides.

Confusingly, the meaning of margin on main also is the left and right margin added. Should be fixed in a follow-up.

The frame was a ScreenRectangle subclass that computed its boundary and center with a shortcut, for the many projections of a 3D frame. Each drawn frame now computes its view once, so the shortcut no longer makes a measurable difference (3D and text benchmarks within noise), and the frame is an ordinary rectangle again. Raised in review.

auto_zoom() set the frame to the mobjects' bounding box; for a point (or mobjects without width and height) and no margin, the frame got no size, and every later auto_zoom divided by its width and height (also on main). It now raises instead, and compares the aspect ratios without dividing. Bounds are read with get_left()/get_right()/get_top()/get_bottom(), the width and height without abs(), and the docstring says that margin is added to the total width or height, as it always was. Found in review.

Mobject.get_image created a renderer and closed it in a finally clause. CairoRenderer is now a context manager (it closes on exit), and get_image uses it. The renderer also reads config attributes instead of keys and conversions, chooses its camera in one expression, and keeps the file writer's type. Suggested in review.

The package's __getattr__ imported every camera class to pick one from a dict; it imports the module that has the name and takes it from there. The camera reads config.frame_width, config.background_color and config.background_opacity as attributes, defaults last. Suggested in review.

GniLudio · 2026-11-07T12:17:57Z

 from manim.typing import MatrixMN, Point3D

-from .. import config
-from ..utils import opengl
+from ... import config
+from ...utils import opengl


Suggested change

from manim.typing import MatrixMN, Point3D

from .. import config

from ..utils import opengl

from ... import config

from ...utils import opengl

from manim import config

from manim.typing import MatrixMN, Point3D

from manim.utils import opengl

The rest of the codebase is predominantly relative imports, switching this should be done more consistently across the codebase (if we wanted to)

The examples of moving a scene's camera declare construct() -> None and the types of their helper, and begin_ambient_camera_rotation() and add_fixed_orientation_mobjects() declare their return and keyword types. Suggested in review.

behackl

Thanks for all the big and small comments; these reviews have improved the first part of this stack considerably again.

auto_zoom() refused a zoom only when the larger of the mobjects' width and height plus the margin was not positive, but it then sets whichever side needs the larger frame for the frame's aspect ratio. A 1.5 x 1 rectangle with margin=-1.2 passed the check on its width and gave the 16:9 frame a height of -0.2, flipping the picture. The check now uses the side that is set. Raised in review.

Since the background image cache moved from each target to the renderer, nothing cleared it: close() promises to release the image buffers, but a closed renderer kept every decoded background file (about 33 MB for a 4K image) for as long as the renderer object lived, for example in a scene kept after rendering or in Jupyter. Targets cleared their cache in close() before. close() now clears the renderer's cache. Raised in review.

CairoRenderer is a context manager, so the temporary renderer that draws Scene.get_image() closes at the end of a with block instead of in a finally clause. Same behavior; flagged by CodeQL.

The cameras guide's TwoCameraViews example was plain code; it is now a manim directive, rendered in the docs like the other camera examples (about a second at low quality). Mobject.show had no docstring, so the API reference left it out and the guide's :meth: reference to it was no link. Raised in review.

behackl

Fixed the bugs uncovered in the latest review round + implemented some changes that I've had postponed to follow-ups before.

GniLudio

LGTM

github-advanced-security Bot found potential problems Sep 4, 2026

View reviewed changes

Comment thread manim/renderer/cairo/camera.py Resolved

Comment thread manim/renderer/cairo/camera.py Resolved

behackl added refactor Refactor or redesign of existing code breaking changes This PR introduces breaking changes labels Sep 4, 2026

nikolajmunk requested changes Sep 6, 2026

View reviewed changes

nikolajmunk reviewed Sep 8, 2026

View reviewed changes

behackl mentioned this pull request Sep 8, 2026

Refactor: move more render lifecycle responsibilities to Manager #5002

Open

nikolajmunk mentioned this pull request Sep 9, 2026

Move mobject positioning methods into separate class #4994

Open

7 of 11 tasks

behackl force-pushed the refactor/camera-renderer-ownership branch from 5623b1f to d0e8b25 Compare October 9, 2026 21:56

behackl removed this pull request from stack #5003 October 26, 2026 18:00

behackl added this pull request to stack #5104 October 26, 2026 18:01

behackl force-pushed the refactor/camera-renderer-ownership branch from 77a3071 to 9b7a29a Compare October 26, 2026 21:42

GniLudio requested changes Oct 28, 2026

View reviewed changes

Comment thread manim/renderer/cairo/rendering.py Resolved

Comment thread manim/renderer/cairo/camera.py Resolved

Comment thread manim/renderer/cairo/camera.py Resolved

GniLudio requested changes Oct 28, 2026

View reviewed changes

Comment thread manim/renderer/cairo/renderer.py Resolved

GniLudio requested changes Oct 28, 2026

View reviewed changes

behackl and others added 15 commits October 29, 2026 12:43

Package the OpenGL renderer backend

d346408

Move Cairo rasterization behind semantic cameras

6b4a603

Remove legacy Cairo camera implementations

1ea30e9

fix: read frame rate from config instead from camera

22b445b

Consolidate backend camera implementations

c19fe05

Fix default Cairo camera aspect without changing explicit geometry

89b921f

Release Cairo raster resources and reject closed drawing

aaf6913

Retire unused nested targets, release retained buffers, and guard closed operations. Reject unused renderer keywords and build dispatch tables once per draw.

Add fresh scene image capture for Cairo and OpenGL

cbd00e7

Expose Scene.get_image() and Scene.show() through Manager without stepping animations or emitting movie frames. Preserve active raster targets and capture OpenGL meshes on the context-owning thread.

Document camera controls, snapshots, and multi-camera composition

00e3cba

Reduce default camera frame query and ThreeD projection overhead

c3ad385

Compute default frame bounds directly without a semantic cache, preserve custom frame hooks, and share perspective scaling between axes. Add mutation, boundary, and projection parity tests.

Apply batched suggestions from code review

bad9902

Co-authored-by: nikolajmunk <28557236+nikolajmunk@users.noreply.github.com>

Clarify camera guides and snapshot documentation

d2f5805

Address review feedback on camera terminology, backend controls, custom camera selection, snapshot behavior, and inset views. Preserve the earlier camera test cleanup.

Clarify camera image rendering and inset documentation

f6d7cdf

Add the Cairo and OpenGL renderer modules to the API reference

6596fe1

CairoRenderer and OpenGLRenderer are public and referenced from the guides and docstrings, but their modules were missing from the reference, so those links could not resolve.

GniLudio requested changes Nov 6, 2026

View reviewed changes

Comment thread docs/source/guides/cameras.rst Resolved

GniLudio requested changes Nov 6, 2026

View reviewed changes

Comment thread docs/source/guides/cameras.rst Resolved

Comment thread docs/source/guides/cameras.rst Resolved

Comment thread docs/source/guides/cameras.rst Resolved

Comment thread docs/source/guides/cameras.rst Resolved

Comment thread docs/source/guides/cameras.rst Resolved

GniLudio requested changes Nov 6, 2026

View reviewed changes

Comment thread docs/source/guides/deep_dive.rst Resolved

Comment thread docs/source/guides/deep_dive.rst Resolved

behackl added 6 commits November 7, 2026 01:51

Remove a Cairo scale call that set_matrix overrides

67e7f1f

The context was scaled to the pixel size and then given a full matrix, which replaces the scale. Pointed out in review.

Name the excluded_mobjects argument of CairoRenderer.update_frame

8744c0a

update_frame took **kwargs only to read excluded_mobjects from them (as main's camera did); a keyword-only parameter says so, and other keywords are rejected instead of ignored.

GniLudio requested changes Nov 7, 2026

View reviewed changes

Comment thread manim/mobject/mobject.py Resolved

Comment thread manim/mobject/mobject.py Resolved

GniLudio requested changes Nov 7, 2026

View reviewed changes

behackl added 4 commits November 7, 2026 12:53

GniLudio requested changes Nov 7, 2026

View reviewed changes

github-advanced-security Bot found potential problems Nov 9, 2026

View reviewed changes

Comment thread manim/renderer/cairo/renderer.py Fixed

behackl commented Nov 10, 2026

View reviewed changes

GniLudio requested changes Nov 10, 2026

View reviewed changes

Comment thread docs/source/guides/cameras.rst Resolved

Comment thread docs/source/guides/cameras.rst Resolved

behackl mentioned this pull request Nov 10, 2026

Refactor: evaluate scenes without rendering, and draw requested frames #5005

Open

GniLudio requested changes Nov 10, 2026

View reviewed changes

Comment thread manim/renderer/cairo/camera.py Resolved

Comment thread manim/renderer/cairo/target.py Resolved

behackl added 4 commits November 11, 2026 12:19

Close the scene image renderer with a with block

0082bd8

CairoRenderer is a context manager, so the temporary renderer that draws Scene.get_image() closes at the end of a with block instead of in a finally clause. Same behavior; flagged by CodeQL.

behackl commented Nov 11, 2026

View reviewed changes

Comment thread docs/source/guides/cameras.rst Resolved

Comment thread docs/source/guides/cameras.rst Resolved

Comment thread manim/renderer/cairo/camera.py Resolved

Comment thread manim/renderer/cairo/target.py Resolved

GniLudio approved these changes Nov 11, 2026

View reviewed changes

Conversation

behackl commented Sep 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Overview: What does this pull request change?

Motivation and Explanation: Why and how do your changes improve the library?

Breaking changes and migration

Module paths

Camera construction and raster access

Removed camera implementations

Links to added or changed documentation pages

Further Information and Comments

Uh oh!

Uh oh!

Uh oh!

nikolajmunk left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

nikolajmunk left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GniLudio left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

behackl commented Oct 29, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GniLudio Nov 7, 2026

Choose a reason for hiding this comment

Uh oh!

behackl Nov 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

behackl commented Sep 4, 2026 •

edited

Loading

nikolajmunk left a comment •

edited

Loading

GniLudio left a comment •

edited

Loading