material-agent-cli

VLM-based material assignment uses vision-language models to predict and apply materials to USD assets. This pipeline automates material binding by analyzing rendered asset views and matching them against a materials manifest with USD bindings.

You can run the material-agent CLI to launch an end-to-end pipeline that assigns materials to USD files. It requires a Python environment, provider credentials for VLM backends, a render endpoint, and a materials manifest with USD bindings.

Yes, the CLI supports resuming failed runs during VLM-based material assignment. You can configure, resume, benchmark, or simulate material prediction pipelines on USD assets with end-to-end control over each step.

You need a Python environment, provider credentials for your selected VLM or LLM backends, a remote render endpoint configuration, and a materials manifest containing USD bindings to execute the material assignment pipeline.

Yes, the CLI enables benchmarking predictions and building datasets from USD renders. You can simulate material assignment workflows and evaluate VLM prediction accuracy across configured pipeline runs.

A CLI enables fast, repeatable material pipelines and end-to-end orchestration of VLM-based material prediction. It allows direct configuration, benchmarking, and resumption of automated material assignment on USD assets with minimal local setup.