Skip to content


Highlight for this release is Locomotion 2.0


As is tradition, let's start off with some of the cool Ragdoll material surfacing across the interwebs since the last release.

Mechanical Spider

"Mechanical Spider" by vertexmonster

60s Marionettes

Some excellent use of Ragdoll here by Ernst-Jan on the forums

Human Ball

Zombies + Radial field = This, by Jason

Oh Deer

Another one by Jason from last Christmas.

Fast Food

Jason has been on fire lately, here's another knockout!

A Whale of a Time

Jason jason jason.

Locomotion 2.0

First released in June 2022, this automated method of generating complete character locomotion has seen a major update!

With a new and much more intuitive workflow, ability to locomote along a path, with feet snapping to terrain and interactive updates as the solver works, you'll be able to achieve much greater results with far less headaches.

New Workflow

Locomotion ❤ Markers.

Borrowing from Markers, Locomotion no longer occupies your animation channels directly. Instead, a preview is shown for you to evaluate and later record.

Iterative Updates

More insight into while you wait.

Locomotion will now show you each result it comes up with, as it searches for one that both looks great and fulfils each of your requirement; targets, limits, mass and gravity.

New Manipulators

Editing targets and limits was a huge chore with Locomotion 1.0. This new release makes authoring plans much less error prone and more fun.


Drag the gizmo to translate a target. Notice how it is no longer possible to author invalid targets that hover above the ground, a significant burden to anyone trying to achieve nice results has been lifted!


Hold Ctrl (on Windows and Linux, Cmd on Mac) to rotate a target.


Hold Shift to adjust the height of the body. Feet must always touch the ground.

Terrain Snapping

Aligning feet to the ground in the case of having a terrain was even more daunting; but no longer! Now feet automatically snap to any terrain.


Go there, then here. Do a twirl, and then go over there!

In Locomotion 1.0, you were able to specify a start and end position for your character. But now, you can specify any number of points inbetween, effectively a path!


Start fast, then walk slowly. Then finish fast!

With more than 1 target, you also need to decide at when you intend on a target being reached. Click and drag on the targets in the UI to adjust.


You'll notice that each target has two shades, a top and bottom.

Shade Meaning
Top Distance, dark is less
Bottom Timing, dark is less


Targets are guidelines to Locomotion. It'll try to follow them, but if some are physically impossible or there are conflicting goals then it will try and find a compromise. Sometimes, you need the target to be more respected, and that's when you can "pin" a target.


Create complex locomotion by stringing together two or more plans.

Complete the illusion by blending them all together with traditional Marker physics.

Here's how it works.

  1. Select a plan
  2. Select a second plan
  3. Run Align Plans

The command will start the second plan where the first plan ends, creating one continuous motion.

Animation to Plan

A "make good" button?

We've talked about generating an animation from a plan, but what about the reverse? What if we could use your animation as input to the plan, and output a refined version of it? 🤔

The Animation to Plan menu item evaluates your animation and converts it into (1) targets and (2) stepsequence; the ingredients to a fully defined plan.

Terrain Accuracy

Plans now more accurately follows your terrain and enables you to visualise what Locomotion sees.

Reset Starting Position

If you change the starting position of your plan, or move the character after having assigned a plan, you can re-align the two with the Reset Starting Position command.

Terrain Debugging

If your terrain has two sides, Locomotion can struggle to snap its feet to the one you intend.

Here's a problematic result.

If we look at the feet snapping, we can see how they tend to snap towards the lower side of this terrain.

We can resolve this by removing the lower side altogether.

And presto, all good!

New Tutorials

Mohamed Elbebawy has produced a new series of tutorials for Ragdoll, available here.


Maya 2024


Released just a few days ago, Ragdoll is fully onboard. 🥳

On the other hand, Maya 2018 has left the chat and is no longer supported by Ragdoll (or Autodesk, for that matter!)

Account Management

You can now monitor and manage your account with us, starting with performing remote deactivation of your node-locked licences.


  • Deactivate machines for nodelocked licences
  • Download receipts and invoices from past purchases
  • Update payment details for subscription licences
  • Activate licence offline
  • Add seats to a licence
  • Add products to your account
  • Cancel a subscription

Checked items are ready for you, unchecked are coming soon!

For the time being, please contact us for login details.

Blend Simulation

There are now 2 ways in which to blend between simulation and animation; before or after.


There is now an option to blend between animation and simulation before recording!

Here are some more examples.

This will also work alongside Translate Motion = Soft.


Ragdoll typically records your simulation onto a layer, and this layer has a Weight attribute that you can animate to blend between simulation and animation.

But, most of the time, it isn't doing what you'd expect. 🤔

A new option has been added to the Record Simulation option dialog to solve this.


Here you can see the effect in action.

Stepped Simulation

Japanese anime, and 2D animation in general, is normally animated on 2's, 3's and sometimes 4's and sometimes at arbitrary times altogether.

Previously, Ragdoll was only ever able to produce 1 frame per Maya frame, meaning the animator would retroactively need to reduce the keys to only happen when animation happens. No longer!

Smooth input, Stepped output

In the simplest case, Ragdoll can be instructed to output simulation ever other or every third frame, or every fourth. Creating a 2D or stopmotion kind of look.

Stepped input, Stepped output

In the more complex case, consider this animation.

Notice how there are pauses between frames, and that keyframes are not distributed evenly. Some have a 2 frame pause, some a 5 frame pause. Here's what would happen if you were to try and simulate this.

Notice how the torso and especially the tail gets jerked around. Because during each pause, Ragdoll goes ahead and simulates it as though the character came to a complete halt. And then kicks off again when it moves, at a very high velocity.

You can now match simulation to when animation occurs, via the new Warp Time menu item.

Now simulation only happens when the character actually moves. You can control the frequency and time delta of each frame by tuning the resulting animation curve.

Restore Time

Use Restore Time to get back to a non-stepped simulation.


Freaky Time

Time is controlled via a normal animation curve, and there's no reason it cannot also go backwards. 😊

Let your creativity out!

Time Method

Normally, during playback Maya will transition from one frame to the next. For example, from frame 1 to 2. But sometimes, when a rig is too heavy, Maya may skip a frame or two. For example from 1 to 3, without first visiting frame 2.

To Maya, the different in time between frame 1 and 2 is about 1/24 seconds, with a playback rate of 24 fps; that is, about 41 milliseconds. The difference between frames 1 and 3 is 83 milliseconds.

To Ragdoll, the difference is always 41 milliseconds. No matter the difference in Maya. Why does that matter? Well, consider this.

At frame 17, the character hits the ground. But what if we go to frame 17 immediately, and skip a few more frames?

Now the character hits the ground at frame 78! To Maya, 78 frames has passed; about 3 seconds. But to Ragdoll, barely a second has passed.

With the new Time Method attribute, you can now choose to handle this in a realistic way instead.

Collision Group

You can now more easily manage which Markers collide, and which does not!


Create New Collision Group

Assign to two or more Markers to create a new collision group.

Add to Existing Collision Group

Select a collision group along with one or more Markers to add.

Remove from Existing Collision Group

Similarly remove Markers.

Multiple Collision Groups

Select multiple groups to visualise them together. They will each occupy a (customisable) color.

Merge Collision Groups

Assign to two or more groups to combine them.


You can now visualise the trajectories of your Markers via the Solver Display attribute.


Mass Independent Fields

Fields were introduced in Ragdoll 2022.02.14 to help apply environment effects such as wind and turbulence to your animations.

Forces are applied as they would in real life... which isn't always practical!

Consider this.

Here, as the character grows in size, so does its mass. As the mass increases, the effect of the turbulence force diminishes. Since the same force is applied to a now much-heavier object. Like if you pushed on a woodden door and then tried pushing a car; your muscles remain as strong, but the object is now much heavier.

With this release, forces are instead applied evenly regardless of mass, which should result is more control and predictability of your simulation.

You can revert to the old (realistic) behaviour by changing Force Mode from Velocity Change to Force in the rSolver node.


Multi Replace Mesh

Ragdoll supports 4 geometry types for collisions.

  • Box
  • Sphere
  • Capsule
  • Mesh

Assign to a control, and then use Replace Mesh alongside any geometry to use this geometry for collision detection.


Sometimes collisions are more complex and require multiple meshes to make up 1 physical collision shape. You can now achieve this with ease.

You can now select any number of meshes and use the combination of them as a collision shape. 🥳

Pin Constraint Pivot

You can now control where on a Marker to Pin Constrain it.

Hold D to edit the position of the rotate pivot

Attach Constraint Pivot

Like the Pin Constraint, Attach Constraints can have their pivots edited too. Especially useful if you are looking to attach two Markers at a specific location.

Quality of Life

Last but not least, a few minor things to make your life easier.

New Behaviour Names

The options for a Marker's Behaviour attribute has been updated for more clarity.

Old Name New Name
Inherit -> Use Group
Kinematic -> Animated
Dynamic -> Simulated

The name "kinematic" is the technical term for when something is animated, so we figured we may as well call it what it is! A Marker set to Animated is entirely animated, with no simulation applied. It may affect other Simulated Markers, but cannot be affected by them.

Edits on Start Frame

Some things in Ragdoll can be animated, others cannot. Those that cannot can also not be changed on any frame other than the start frame.

This used to be a chore, and is no longer! Ragdoll will now automatically return to the start frame when it needs to, leaving you with less to worry about.

Symmetry and Undo

You can now undo both sides at once, as you'd expect!

Z-up Assets

You can now export assets from a Y-up Maya scene, and load it into a Z-up Maya scene! This means our default assets, which were made in a Y-up environment, will now load perfectly fine in your Z-up scenes.



Linked Solver Crash

Many linked solvers could potentially crash Maya, no longer!

Deprecating Max Mass Ratio

Ragdoll likes when the masses of objects that interact are somewhat similar. A 60kg character sitting on a 100kg horse is 👌. But a million ton planet falling onto a 60g ant is less than OK.

To combat this, an attribute was introduced a while back called Max Mass Ratio whose job it was to even out the differences ("ratio") between masses. But it didn't work that great, the simulation was no happier because of it.

So in this release, we are deprecating this attribute. It will still exist in your scene, and can be found at the very bottom of the Solver node in the Attribute Editor. If you find yourself needing it, let us know. If not, you have one less thing to worry about.

Restore T Key

There's an option to restore the functionality of the T key, and to change it to something else.

One-off bug

Ragdoll was reading values from Maya that were 1 frame too old.


I'm stepping through the simulation, notice how the pelvis is set to Animated (Kinematic) but how it doesn't actually have an effect until the next frame?


This has now been fixed.

Solver Update

The underlying maths behind your physics has been modernised to resolve subtle issues but primarily lay the foundation for new things to come. You shouldn't notice much of any difference, apart from subtle things like this.

  • Before 600 boxes: 69-81 fps
  • After 600 boxes: 71-84 fps (lots of independent rigid bodies)
  • Before Wyvern: 82 fps
  • After Wyvern: 79 fps (lots of connected and constrained rigid bodies)

Python API

For the Python folk out there, these members were added to the Python API.

from ragdoll import api

# Create a new group
group = api.assign_collision_group(markers)

# Add some more markers
api.add_to_collision_group(other_markers, group)

# Remove some markers
api.remove_from_collision_group(markers, group)

Last update: 2023-03-31