Class OrchestrationResponse

constructor

• new OrchestrationResponse(rawResponse: HttpResponse): OrchestrationResponse

  Parameters

  • rawResponse: HttpResponse

  Returns OrchestrationResponse

Readonly_data

ReadonlyrawResponse

findChoiceByIndex

• findChoiceByIndex(index: number): LlmChoice | undefined

  Parses the response and returns the choice by index.

  Parameters

  • index: number

    The index of the choice to find.

  Returns LlmChoice | undefined

  An LLMChoice object associated with the index.

getAllMessages

• getAllMessages(choiceIndex?: number): ChatMessages

  Messages that can be used for subsequent prompts as message history. When reasoning content is present, it is automatically included in the returned messages. Pass the result directly as messagesHistory in the next request to maintain reasoning context.

  Parameters

  • OptionalchoiceIndex: number

    The index of the choice to parse.

  Returns ChatMessages

  A list of all messages including any reasoning content.

getAssistantMessage

• getAssistantMessage(choiceIndex?: number): AssistantChatMessage | undefined

  Gets the assistant message from the response.

  Parameters

  • OptionalchoiceIndex: number

    The index of the choice to use (default is 0).

  Returns AssistantChatMessage | undefined

  The assistant message.

getCitations

• getCitations(): Citation[] | undefined

  Gets the citations from the orchestration response. Citations are returned by models like Perplexity Sonar that provide source references.

  Returns Citation[] | undefined

  The citations, or undefined if there are none.

getContent

• getContent(choiceIndex?: number): string | undefined

  Parses the orchestration response and returns the content. If the response was filtered, an error is thrown.

  Parameters

  • OptionalchoiceIndex: number

    The index of the choice to parse.

  Returns string | undefined

  The message content.

getFinishReason

• getFinishReason(choiceIndex?: number): string | undefined

  Reason for stopping the completion.

  Parameters

  • OptionalchoiceIndex: number

    The index of the choice to parse.

  Returns string | undefined

  The finish reason.

getIntermediateFailures

• getIntermediateFailures(): OrchestrationError[] | undefined

  Gets the intermediate failures from the orchestration response. When using module fallback, this contains errors from module configurations that failed before a successful one was found.

  Each failure includes:

  • message: Description of what went wrong
  • code: Error code identifying the type of failure
  • location: Position in the fallback chain (e.g., "config[0]", "config[1]").

  Returns OrchestrationError[] | undefined

  The intermediate failures, or undefined if there were none.

  Example

  const response = await client.chatCompletion();
  const failures = response.getIntermediateFailures();
  if (failures) {
    failures.forEach(f => console.log(`${f.location}: ${f.message}`));
  }
  Copy

getIntermediateResults

• getIntermediateResults(): ModuleResults

  Gets the intermediate results from the orchestration response.

  Returns ModuleResults

  The intermediate results.

getReasoningContent

• getReasoningContent(choiceIndex?: number): string[] | undefined

  Parses the orchestration response and returns the reasoning content generated by the model. Only available for models that support reasoning (enabled via the harmonized reasoning_effort model parameter). Returns only the user-visible reasoning text; encrypted or redacted reasoning blocks are excluded. To preserve reasoning context in multi-turn conversations, pass the result of getAllMessages as messagesHistory in the next request instead of using this method directly.

  Parameters

  • OptionalchoiceIndex: number

    The index of the choice to parse.

  Returns string[] | undefined

  The content of each reasoning block as an array, or undefined if no reasoning content is present.

getRefusal

• getRefusal(choiceIndex?: number): string | undefined

  Parses the orchestration response and returns the refusal message generated by the model.

  Parameters

  • OptionalchoiceIndex: number

    The index of the choice to parse.

  Returns string | undefined

  The refusal string.

getRequestId

• getRequestId(): string

  Gets the request ID from the orchestration response.

  Returns string

  The request ID.

getTokenUsage

• getTokenUsage(): TokenUsage

  Usage of tokens in the response.

  Returns TokenUsage

  Token usage.

getToolCalls

• getToolCalls(choiceIndex?: number): MessageToolCalls | undefined

  Parses the orchestration response and returns the tool calls generated by the model.

  Parameters

  • OptionalchoiceIndex: number

    The index of the choice to parse.

  Returns MessageToolCalls | undefined

  The message tool calls.