Chat

const reply = await RunAnywhere.llm.generate([
  { role: 'system', content: 'You are a concise assistant.' },
  { role: 'user', content: 'What is on-device inference?' },
  { role: 'assistant', content: 'Running the model on your own hardware.' },
  { role: 'user', content: 'Why does that matter?' },
])

for await (const event of RunAnywhere.llm.generateStream(messages)) {
  if (event.type === 'textDelta') reply += event.text
}

import type { ChatMessage } from '@runanywhere/core'
import { useCallback, useRef, useState } from 'react'

export function useConversation() {
  const messages = useRef<ChatMessage[]>([
    { role: 'system', content: 'You are a helpful assistant.' },
  ])
  const [transcript, setTranscript] = useState<ChatMessage[]>([])

  const send = useCallback(async (text: string) => {
    messages.current.push({ role: 'user', content: text })
    setTranscript([...messages.current])

    let reply = ''
    for await (const event of RunAnywhere.llm.generateStream(messages.current)) {
      if (event.type === 'textDelta') reply += event.text
    }

    messages.current.push({ role: 'assistant', content: reply })
    setTranscript([...messages.current])
  }, [])

  return { transcript, send }
}

if (messages.current.length > 21) {
  messages.current.splice(1, messages.current.length - 21)
}