Interface: SseResponse
Defined in: ai/src/provider/sse.ts:10
Minimal Server-Sent-Events support for OpenAI-compatible streaming.
Some gateways (e.g. go4ai) return a spurious 500 when they buffer a long non-streaming completion, even though the model generated it (and charged for it). Streaming sidesteps that: the response is delivered in chunks, so there's no large buffered body to choke on.
Properties
chunks
> chunks: AsyncIterable<string>
Defined in: ai/src/provider/sse.ts:16
Decoded text pieces of the streaming body.
ok
> ok: boolean
Defined in: ai/src/provider/sse.ts:11
status
> status: number
Defined in: ai/src/provider/sse.ts:12
Methods
text()
> text(): Promise<string>
Defined in: ai/src/provider/sse.ts:14
Full body — used for the error message on a non-2xx.
Returns
Promise<string>