llm4s-docs/org.llm4s/org.llm4s.llmconnect/org.llm4s.llmconnect.spi/EmbeddingProviderDescriptor

EmbeddingProviderDescriptor

org.llm4s.llmconnect.spi.EmbeddingProviderDescriptor

Everything llm4s needs to know about one embedding provider, supplied by the provider itself.

This is ProviderDescriptor's counterpart for the embedding half of the SPI, and it is deliberately a separate trait rather than a method on the chat descriptor: the two sets overlap but neither contains the other. OpenAI and Ollama supply both a chat client and an embedding provider, Voyage supplies only embeddings, and Anthropic only chat. Folding embeddings into ProviderDescriptor would force an embedding-only provider to implement buildClient and buildConfig only to fail them.

Register it by listing it in an Llm4sProviderModule's Llm4sProviderModule.embeddingProviders, exactly as a chat provider is registered, and EmbeddingClient.from can reach it with nothing in llm4s-core edited.

Attributes

Example

object JinaEmbeddings extends EmbeddingProviderDescriptor:
 val id = ProviderId("jina")
 def build(config: EmbeddingProviderConfig): Result[EmbeddingProvider] =
   Right(JinaEmbeddingProvider.fromConfig(config))

Graph

Supertypes

class Object

trait Matchable

class Any

Known subtypes

object OllamaEmbeddingProvider

object OpenAIEmbeddingProvider

object VoyageAIEmbeddingProvider

Members list

Value members

Abstract methods

Constructs the embedding provider for an already-resolved config.

The config has been read and validated by the caller, so a Left here means something only this provider can detect — a model its API does not offer, say.

Attributes

Canonical id, e.g. ProviderId("voyage"). Must be unique among the embedding providers in a registry.

Chat and embedding ids live in separate namespaces, so a provider that supplies both — OpenAI, Ollama — uses the same id for each without a clash. That is what lets EMBEDDING_MODEL=ollama/nomic-embed-text and LLM_MODEL=ollama/llama3 name the same provider.

Attributes

Concrete methods

Alternative spellings accepted in EMBEDDING_MODEL=<name>/<model>, folded onto id.

Attributes

In this article

Generated with