OpenAIResponsesCompactionArgs
このコンテンツはまだ日本語訳がありません。
type OpenAIResponsesCompactionArgs = object;Session subtype that can run compaction logic after a completed turn is persisted.
Properties
Section titled “Properties”compactionMode?
Section titled “compactionMode?”optional compactionMode: "previous_response_id" | "input" | "auto";How the compaction request should provide conversation history.
When omitted, implementations use their configured default.
force?
Section titled “force?”optional force: boolean;When true, compaction should run regardless of any internal thresholds or hooks.
responseId?
Section titled “responseId?”optional responseId: string;The response.id from a completed OpenAI Responses API turn, if available.
When omitted, implementations may fall back to a cached value or throw.
store?
Section titled “store?”optional store: boolean;Whether the last model response was stored on the server.
When set to false, compaction should avoid previous_response_id unless explicitly overridden.