Scrydon
Vendor Catalog

vLLM

Self-hosted high-throughput inference

Vendor ID: vllm · Categories: AI

vLLM — high-throughput, memory-efficient inference engine. Exposes an OpenAI-compatible API surface.

Auth

CredentialNotes
apiKey (optional)If your vLLM deployment is gated. Many self-hosted setups don't enable it.
noneIf the vLLM server is reachable directly.

Capabilities

CapabilityWire protocol
LLMopenai-chat-v1
Embeddingopenai-chat-v1-compatible

Configured per integration with a baseUrl pointing at the vLLM endpoint.

On this page

On this page