Interface: OpenAIResponsesReasoningItem
Defined in: src/batteries/llm/openai_responses/types.ts:193
A model-emitted reasoning item — the native wire representation of a "thought." Only ever sent back on input when reasoningReplay !== 'off' and its stored signature/prefix-fingerprint is still valid; see the reasoning-pairing adjacency-sweep pass and Known Gotcha #1.
Properties
| Property | Type | Description | Defined in |
|---|---|---|---|
content? | OpenAIResponsesReasoningContentPart[] | Ordered full-reasoning-text parts, present only under some include values. | src/batteries/llm/openai_responses/types.ts:201 |
encrypted_content? | string | null | Opaque, provider-signed reasoning payload. Present only when include requested it. | src/batteries/llm/openai_responses/types.ts:203 |
id | string | Item id, used to correlate this reasoning item with adjacent output items on replay. | src/batteries/llm/openai_responses/types.ts:197 |
status? | "completed" | "incomplete" | "in_progress" | Lifecycle status of the reasoning item. | src/batteries/llm/openai_responses/types.ts:205 |
summary | OpenAIResponsesReasoningSummaryPart[] | Ordered summary-text parts for this reasoning item. | src/batteries/llm/openai_responses/types.ts:199 |
type | "reasoning" | Item discriminator; always 'reasoning'. | src/batteries/llm/openai_responses/types.ts:195 |