vLLM (vllm)
Local inference (text and vision language models).
Generated from catalog/components/vllm.yaml: do not edit by hand.
| Purpose | Local inference (text and vision language models) |
| Kit capability | Inference locale |
| Plane | models |
| State | integrated |
| Pinned version | v0.33.0 |
| Licence | Apache-2.0 |
| Licence text | https://raw.githubusercontent.com/vllm-project/vllm/v0.33.0/LICENSE (checked on 2026-11-03) |
| Licence status | allowed |
| Licence exception | - |
| Upstream repository | https://github.com/vllm-project/vllm |
| Owner | ai-platform |
| Distribution | Kubernetes chart, Navin Manager host service, Compose development stack |
| Navin Manager service | vllm |
Images and version pins
| Role | Kind | File | Value |
|---|---|---|---|
| Kubernetes | text | deploy/helm/enterprise-ai/values.yaml | tag: v0.33.0 |
| Kubernetes | text | deploy/helm/enterprise-ai/values.yaml | tag: v0.33.0-x86_64 |
| Navin Manager host | navin-manager-catalog | navin-manager/catalog/services/vllm.yaml | 0.33.0 |
| Compose (development) | compose-image | deploy/compose/docker-compose.yml (selector vllm) | vllm/vllm-openai:v0.33.0 |
Container images
vllm/vllm-openai:v0.33.0
The container images of the Navin Manager host service are listed on vllm.
Dependencies
Operations
| Supported | install, configure, start, stop, restart, rolling-restart, drain, scale, upgrade, rollback, repair, decommission |
| Not supported | backup, restore, rotate |
Back to the component index.