Octane Chat UI | TanStack AI 简体中文

Install @tanstack/ai-octane. Import the UI factory from @tanstack/ai-octane/ui. Call createChatHook({ options, ...components }) once at module scope. Your app calls useAppChat() to create the instance. Render <chat.AppChat />.

The factory API matches React. Widgets are Octane components. Import hooks from octane.

The factory needs a toolsComponents entry for every tool name in chatOptions. It also needs an interruptsComponents.generic entry for every interrupt id. generic.fallback is optional.

The server route matches the React page. The chat UI recipes show the same option groups. The code there is React, and the shape carries over.

Client

import { fetchServerSentEvents } from '@tanstack/ai-octane'
import { createChatHook } from '@tanstack/ai-octane/ui'
import { toolDefinition } from '@tanstack/ai'
import { z } from 'zod'

const getWeather = toolDefinition({
  name: 'getWeather',
  description: 'Look up weather',
  inputSchema: z.object({ city: z.string() }),
  outputSchema: z.object({ temperature: z.number() }),
}).client()

const chatOptions = {
  connection: fetchServerSentEvents('/api/chat'),
  tools: [getWeather],
}

const { useAppChat } = createChatHook({
  options: chatOptions,
  components: {
    layout: (props) => (
      <main>
        <props.Messages />
        <props.Interrupts />
        <props.Queue />
      </main>
    ),
    message: (props) => (
      <article>
        <props.Parts />
      </article>
    ),
  },
  partsComponents: { fallback: (props) => <span>{props.part.type}</span> },
  toolsComponents: {
    getWeather: (props) => <strong>{props.part.input?.city}</strong>,
  },
})

export function ChatScreen() {
  const chat = useAppChat()
  return <chat.AppChat />
}