Triton Control, an open-source Kubernetes operations UI for Triton

[aiexpert-cyber]

We built an Apache-2.0 management and workflow layer around Triton Inference Server. It supports existing and self-deployed Triton instances, S3-backed model repositories, model configuration inspection, inference tests, Perf Analyzer, browser-based code-server workspaces, MLflow, and Argo Workflows.

Triton remains the serving layer; the project is intended to reduce the repeated operational glue around it. We currently include examples for Python, LibTorch, ONNX, TensorRT, TensorRT-LLM, and vLLM.

We would appreciate feedback on API compatibility, repository workflows, and the controls teams need for real Kubernetes deployments:

https://github.com/ai-lab-tech/triton-control