Replicated the laundry task in mujoco web, and also self-trained the RL WBC. By Opus 5.5 in 3 hrs. Agentic real2sim is nearly solved.

Wow nice work!

Thanks Lindon! Thrilled to get acknowledged by you. I am always fascinated by what Dyna has done.

@ZeYanjie Cool! How did you trained a urr based rl policy?

forwarded to our eng 🫡

awesome! does mujoco web support cloth folding in sim?

Mujoco supports cloth simulation, but the performance is not good and can't really do real-time in web. So in the video the cloth is only presented as rigid body.

Did the three-hour build match how the towels behave, or mainly the robot and laundry room?

Mujoco supports cloth simulation, but the performance is not good and can't really do real-time in web. So in the video the cloth is only presented as rigid body.

@ttktjmt let's add this to MjSwan ?

Wow very impressive, I will look into that 🔥

laundry task in 3 hrs with opus is kinda nuts

Very cool, mindblowing what coding agents are capable of.

This is very cool

Was the whole simulation also replicated by Opus 5.5?

The robot trajectory is retrieved from official website (only kinematic) and Opus did the scene and robot dynamics to match this trajectory.

This is so cool. Are the towels modeled as well in sim?

Currently all rigid bodies so it's easier to run on web.

super cool!

3 hours? I could not help myself. This wants to live on a factory cart. Print brackets, bolt on wheels, and let Opus do the boring RL tuning. OH YES.

That looks great! Are you going to try the same with XLeRobot? It is cheap enough to actually test sim-to-real on the real hardware

Hi will you release the codebase

is this mujoco? the rendering quality look really high?

Mujoco web, so I can have additional renders customized (while still runs in real time in the browser)

what additional renders? how are you doing that

Basically I pre-baked the lighting using blender first.

The rigid-cloth clarification is useful. Which physical properties were estimated from the source assets, and which used defaults? A sensitivity sweep on friction and inertia would be interesting before treating this as a faithful dynamics model.

Wow this is insane! Did the real2sim modeled the robot as well? Would be great if you can share the harness - I have been trying lately as well

Robot file is extracted from Dyna's website and added physical properties, and also generated collision mesh. Didn't use special designed harness, just prompted my requirements in claude code.

That's sooo brilliant! Love the idea - and still amazed by the real2sim abilities, agentic robotics is here

will it be open-sourced?

After I confirm with Dyna, it’s their IP after all

again it needs to be proved effective when deployed in real world!

Nice work！How do you generate URR data？

~98% synthetic (random keyframes + min-jerk + step targets → FK), seeded by ~97 s of Dyna's public ego-replay clips

This looks really good. Nice work.