vLLM Manager

Command preview

vllm serve …

This is exactly what will run when you click "Start vLLM Server". The HF token (if any) is passed as an environment variable, not on the command line.

Environment variables applied to the process:

(none)

Name	Model	Updated

Loading…

Model	Size on disk	Files	Last used

Checking server status…

Send a test message

System prompt (optional)

Your message

Max tokens

Temperature

Response

(no response yet)

Checking server status…

Request

Pick a preset for a ready-to-run example against a real vLLM route, or set Method / Path / Body yourself to hit anything vLLM exposes — chat, legacy completions, tokenize/detokenize, models, health, metrics, embeddings, or any newer endpoint this app doesn't have a dedicated button for yet. Requests go straight to the running server exactly as shown.

Preset

Method

Path

Request body (JSON — only sent for POST)

Response

(no response yet)

Loading status…

Auto-refresh

(no logs yet)