Deploy to HF Spaces

Free debugger on OpenCode Zen; separate code and scene model lists

Raise the shared result-storage limit to 9 GB

Brand the app, strict per-tester GPU shares, free prompt enhancement

Fix scene generation for robots with an unnamed free joint

Route every task to a model automatically and budget AI usage in rupees

Add a web app for testers

Default the ROS 2 coding model to DeepSeek V4 Flash for the test phase

Default the ROS 2 coding model to DeepSeek V4 Pro

Let ROS 2 code use its own model (CODE_LLM_MODEL)

Refuse new jobs while the R2 bucket is near its free tier

Keep the default Modal deploy free of stored data

Add Isaac Sim RTX renders of finished runs

Export runs as LeRobot datasets

Add a first-pass video review with NVIDIA Cosmos Reason

Add robustness sweeps with MuJoCo Warp

Add /code/check, more robots, and a real hold controller

Track every backend job with its cost and stop at a monthly budget

Use GLM-5.3 via OpenRouter and run simulations on Modal (RunPod optional)

Keep free objects where the scene puts them when resetting to the robot keyframe

Fix scene validation breaking when a GL backend is configured

Harden simulation runs: official Foxglove embed, WebM video, CI image build

Add GPU simulation runs: RunPod worker, R2 storage, report page

Add ROS 2 code generation via OpenCode, README

Add Stage 1 MuJoCo scene generation API