vLLM Playground
vLLM Playground is a web interface for running, managing and chatting with vLLM inference servers, supporting GPU and CPU backends including Apple Silicon and OpenShift or Kubernetes deployment.
vLLM Playground provides a web UI for managing and interacting with vLLM inference servers. It supports GPU and CPU modes with optimizations for Apple Silicon Macs and for enterprise deployment on OpenShift and Kubernetes, and it can run subprocess, container, and remote vLLM servers side by side, switching between saved configurations from a management screen.
Features include image and speech generation through vLLM-Omni integration, running Claude Code against locally served open models, MCP tool support with human-in-the-loop approval, chat with vision-language models such as Qwen2.5-VL and LLaVA, and per-model routing for gateways that expose a dedicated base URL per model. It installs from PyPI and includes an observability dashboard, a PagedAttention visualizer and speculative-decoding support.