chatStream()

Creates an LLM streaming chat with auto-cancellation and backpressure.

Signature

function chatStream(factory: ChatStreamFactory, opts?: ChatStreamOptions): ChatStreamResult

Parameters

Parameter	Type	Description
factory	ChatStreamFactory	Async iterable factory: receives messages + AbortSignal, yields text chunks.
opts	ChatStreamOptions	Optional configuration.

Returns

ChatStreamResult — send, stop, retry, clear, messages, partial, streaming, error.

Basic Usage

import { chatStream } from 'callbag-recharge/ai/chatStream';

const chat = chatStream(async function* (signal, messages) {
    const res = await fetch('/api/chat', {
        method: 'POST',
        body: JSON.stringify({ messages }),
        signal,
      });
  const reader = res.body!.getReader();
  const decoder = new TextDecoder();
  while (true) {
    const { done, value } = await reader.read();
    if (done) break;
    yield decoder.decode(value);
  }
});

chat.send('Hello!');
// chat.messages.get() → [{ role: 'system', content: '...' }, { role: 'user', content: 'Hello!' }]
// chat.partial.get() → accumulating response text...
// chat.streaming.get() → true

Options / Behavior Details

• Auto-cancel: Sending a new message while streaming cancels the current generation.
• Rate limiting: Pass a rateLimiter to throttle API calls (e.g., for LLM RPM limits).
• Message history: messages store includes the system prompt (if set) and reactively updates.