@tanstack/ai-svelte

Svelte bindings for TanStack AI.

Installation

npm install @tanstack/ai-svelte
# or
pnpm add @tanstack/ai-svelte
# or
yarn add @tanstack/ai-svelte

Usage

<script>
  import { createChat, fetchServerSentEvents } from '@tanstack/ai-svelte'

  const chat = createChat({
    connection: fetchServerSentEvents('/api/chat'),
  })
</script>

<div>
  {#each chat.messages as message}
    <div>{message.role}: {message.parts[0].content}</div>
  {/each}

  {#if chat.isLoading}
    <button onclick={chat.stop}>Stop</button>
  {/if}

  <button onclick={() => chat.sendMessage('Hello!')}> Send </button>
</div>

API

createChat(options)

Creates a reactive chat instance. Returns an object with reactive getters and methods:

Reactive Properties (no $ prefix needed):

• messages - Array of messages in the conversation
• isLoading - Boolean indicating if a response is being generated
• error - Current error, if any

Methods:

• sendMessage(content) - Send a message
• append(message) - Append a message
• reload() - Reload the last assistant message
• stop() - Stop the current response generation
• clear() - Clear all messages
• setMessages(messages) - Set messages manually
• addToolResult(result) - Add a tool result
• addToolApprovalResponse(response) - Respond to a tool approval request

Svelte 5 Runes

This library uses Svelte 5 runes ($state) internally, providing a clean API where you don't need to use the $ prefix to access reactive state:

<script>
  const chat = createChat({ ... })
  
  // No $ needed - these are reactive getters!
  console.log(chat.messages)
  console.log(chat.isLoading)
</script>

<!-- Reactivity works automatically in templates -->
{#each chat.messages as message}
  ...
{/each}

License

MIT