BetaWiTwin Studio
The RF Digital Twin
Describe a scene to the Agent, run differentiable Maxwell, radar and channel solvers on it, and keep every Project an ordinary Python codebase.
What you can build
Radar, channel, EM and home simulations open in Studio as ready-made Projects. The same solvers cover satellite links and robots.
Get started
Studio for the editor; the solvers are Python packages you can use on their own.
Studio
AI-native editor: build scenes by chat, run solvers, keep every Project a Python codebase
From a sentence to a solved Scene
One editor and one set of open solvers for electromagnetics, radar and wireless channels
End-to-end differentiable
Maxwell, Radar and Channel plug into PyTorch autograd, so a loss on the simulated field, radar return or channel flows back to geometry, materials and positions. Studio runs gradient, sweep, Monte Carlo, genetic, PSO and Bayesian Studies on the same Scene.
Learn morePhysics inside your network
Solver outputs are torch tensors on the GPU. Put a simulator between two learned modules and train the whole model end to end, with the physics kept exact.
Learn more
Citation
If you use RF Digital Twin in your research, please cite our paper
@inproceedings{chen2026rfdt,
title = {Physically Accurate Differentiable Inverse Rendering
for Radio Frequency Digital Twin},
author = {Chen, Xingyu and Zhang, Xinyu and Zheng, Kai and
Fang, Xinmin and Li, Tzu-Mao and Lu, Chris Xiaoxuan
and Li, Zhengxiong},
booktitle = {Proceedings of the 32nd Annual International Conference
on Mobile Computing and Networking (MobiCom)},
year = {2026},
doi = {10.1145/3795866.3796686},
publisher = {ACM},
address = {Austin, TX, USA},
}Licensing
Research
For academic and non-commercial research purposes
- Full access to simulation capabilities
- Academic publications allowed
- Community support
Commercial
For commercial and enterprise applications
- Production deployment rights
- Priority support
- Custom integration assistance