Embabel Agent Documentation/com.embabel.agent.api.common/Ai

Ai

interface Ai

Gateway to AI functionality in the context of an operation. This includes both LLM and embedding models.

Functions

rag

abstract fun rag(): RagService

Return the RagService

withAutoLlm

open fun withAutoLlm(): PromptRunner

Get a configurable PromptRunner for this context using automatic model selection criteria. This may consider prompt and tools, so is not the same as default.

withDefaultEmbeddingModel

open fun withDefaultEmbeddingModel(): <Error class: unknown class>

withDefaultLlm

open fun withDefaultLlm(): PromptRunner

Get a configurable PromptRunner for this context using the default model selection criteria.

withEmbeddingModel

abstract fun withEmbeddingModel(criteria: <Error class: unknown class>): <Error class: unknown class>

Return an embedding model matching the given criteria.

open fun withEmbeddingModel(model: String): <Error class: unknown class>

Return an embedding model with the given name

withFirstAvailableLlmOf

open fun withFirstAvailableLlmOf(vararg llms: String): PromptRunner

withLlm

abstract fun withLlm(llm: <Error class: unknown class>): PromptRunner

Get a configurable PromptRunner for this context using the given LLM. Allows full control over LLM options.

open fun withLlm(model: String): PromptRunner

Get a configurable PromptRunner for this context choosing the given model by name and the default LLM options. Does not allow for any other LLM options to be set.

withLlmByRole

open fun withLlmByRole(role: String): PromptRunner

Get a configurable PromptRunner for this context choosing the given model by role and the default LLM options. Does not allow for any other LLM options to be set. Users must configure roles, for example in application.properties.