* Announces an outbound model-layer HTTP request the runtime wants the SDK client to service. Carries the request head only; the body always follows as one or more httpRequestChunk frames keyed by the same requestId, even when the body is empty (a single chunk with end=true). * * @param pa
(params: LlmInferenceHttpRequestStartRequest)
| 17103 | * @returns Acknowledgement. Returning successfully simply means the SDK accepted the start frame; it does not imply the request will succeed. |
| 17104 | */ |
| 17105 | httpRequestStart(params: LlmInferenceHttpRequestStartRequest): Promise<LlmInferenceHttpRequestStartResult>; |
| 17106 | /** |
| 17107 | * Delivers a body byte range (or a cancellation signal) for a request previously announced via httpRequestStart, correlated by requestId. The runtime fires at least one chunk per request — when there is no body, a single chunk with empty data and end=true. Mid-stream the runtime may send a chunk with cancel=true to abort the request; the SDK then stops issuing httpResponseChunk frames and may emit a terminal httpResponseChunk with error set. |
| 17108 | * |
no outgoing calls
no test coverage detected