asStream
Provides access to the callbacks that send data back to Llama Workspace. The method accepts an async function which provides pushText
as a parameter. When pushText
is invoked, a chunk of text is sent back to Llama Workspace and displayed to the user.
Parameters
processor: ({ pushText }: ProcessorCallbacks) => void
- The processor function that you will use to send data back to the client.
Returns
Promise<ReadableStream<Uint8Array>>
Example
import { LlamaWorkspace } from "llamaworkspace";
// app is an instance of express, koa, hono or any other web framework
app.post("/", async (req) => {
// ...instantiation code...
const stream = llamaWorkspace.asStream(async ({ pushText }) => {
// Your internal logic that eventually generates streaming data.
const myInternalStream = await myAIAgentProcess();
for await (const chunk of myInternalStream) {
// Every chunk of text generated by your code is sent back to Llama Workspace
pushText(chunk);
}
});
return new Response(stream, {
headers: {
"Content-Type": "text/event-stream",
},
});
});