NamedProviderConfig

Value parameters

apiKey

provider-dependent credential: an API key for most providers, or a path to a service-account credentials file for VertexAI

auth

the section's workload-identity block, if any: the identity token to exchange and the keys its provider declares in ProviderConfigSpec.authExtras. Validation guarantees it is never set together with apiKey. Redacted in toString.

baseUrl

optional base URL override

extras

the provider-specific keys its descriptor declares in ProviderConfigSpec.extras, as validation leaves them: required ones present, defaults filled in, deprecated aliases resolved to the current name, and undeclared keys dropped. Values are redacted in toString, since a provider may declare a credential here. Vendor-specific settings live here rather than as fields - Azure's endpoint and apiVersion, OpenAI's organization, the generic openai-compatible provider's contextWindow and reserveCompletion - so that this type does not change as providers come and go.

headers

extra HTTP headers sent on every request; read by openai-compatible and the OpenAI-format model listers, and ignored by the rest. Values are redacted in toString.

model

the model name to use

provider

the resolved ProviderId

timeouts

the section's timeouts block: how long a request and a stream may take. Absent values leave each client on its own default (org.llm4s.llmconnect.config.ProviderTimeouts). Every provider that makes HTTP calls reads it, so it is a built-in field like headers, not an extra.

Attributes

Companion

object

Graph

Reset zoomHide graphShow graph

Supertypes

trait Serializable

trait Product

trait Equals

class Object

trait Matchable

class Any

Show all