[Inception member] Natural-language control for circuit/current simulation and 3D-printing setup — what's available in the NVIDIA stack?

Hi everyone,

We’re MyMeaning, an NVIDIA Inception member based in Korea. We’ve built a natural-language “product factory”: a user describes a product in plain language, and we generate URDF/USD and run physics simulation (Newton / Isaac / Warp).

We’d like to extend natural-language control to two more areas and are looking for the right NVIDIA building blocks:

  1. Electronics — driving circuit / current (SPICE-like) simulation from natural language. Is there anything in the NVIDIA stack (CUDA-X, NIM, Nemotron agents) for electrical/circuit simulation we could wrap behind an LLM agent? Or is EDA partner software (e.g. Synopsys/Cadence on GPU) the only path?

  2. 3D printing / manufacturing — setting geometry, slicing, and material parameters from natural language, ideally through Omniverse / OpenUSD (e.g. USD Code NIM?).

Are there recommended blueprints, NIM microservices, or Omniverse/Kit extensions for either of these? Any pointers would be much appreciated.

Thanks!

Hi @choi5, thank you for posting your question. Quick, honest answer on the two extensions.

1. Electronics / circuit (SPICE-like) simulation — out of scope for Isaac Sim

This one is outside Isaac Sim’s domain: Isaac Sim (and Newton/Warp) is a robotics and physics simulator, not an electrical/circuit simulator, so there’s no Isaac building block that will do SPICE-type analog/current simulation. NVIDIA also doesn’t ship a first-party circuit simulator to wrap directly.

Where I’d point you instead:

  • Simulation engine: GPU-accelerated circuit/analog simulation today comes from the EDA partners (e.g. Cadence, Synopsys), or an open-source engine like ngspice for a lighter path.
  • The NL/agent layer is where NVIDIA fits: an LLM agent (hostable via NIM, with Nemotron models) can translate natural language into a netlist and drive whichever simulator you pick. That part is provider-agnostic on the engine side.

2. 3D printing / manufacturing from natural language — partly in scope

This is closer to home on the geometry side:

  • Geometry authoring: OpenUSD is the right substrate, and USD Code NIM fits the “NL → OpenUSD/Python” step. Omniverse/Kit is then your visualization / digital-twin layer.
  • Slicing / toolpaths / material params is not an Omniverse feature — you’d drive a dedicated slicer (e.g. open-source CuraEngine / PrusaSlicer) from your agent with the parameters it derives from the prompt.
  • For USD Code NIM and Omniverse specifics, the OpenUSD / Omniverse developer forums are the right venue (again, more targeted than this Isaac Sim category). OpenUSD - NVIDIA Developer Forums

Hope that helps!

Hello!

We noticed that this topic hasn’t received any recent updates from anyone reporting this issue, so we are closing it for now to help keep the forum organized.

If you are still experiencing this issue or have related questions, please create a GitHub Discussion or Issue in the Isaac Sim repository and include a link to this topic along with updated details. Mentioning or linking to this original topic provides helpful context and makes it easier for others to assist you.

Thank you for being part of the NVIDIA Isaac Sim community.

Best regards,
The NVIDIA Isaac Sim Forum Team