hashbrown

experimental_webllm

Creates a local transport from an application-owned WebLLM engine.

Supports text, structured output, and trusted UI without importing WebLLM. Tools are rejected; images require an explicit media opt-in and resolver. Overlapping runs sharing an engine fail with MODEL_BUSY; applications must also avoid using that engine directly while a Hashbrown run owns it.

API

export declare function experimental_webllm(
  options:  ExperimentalWebLlmTransportOptions ,
):  TransportFactory;
experimental_webllm
@paramoptions:
@returns

API

export declare function experimental_webllm(
  options:  ExperimentalWebLlmImageTransportOptions ,
):  TransportFactory;
experimental_webllm
@paramoptions:
@returns