@shadcn/helpers package lets you write chat conversations in code

Small, focused helpers for developing apps, starting with AI SDK and TanStack AI.

We are open sourcing @shadcn/helpers, a new package for small, focused helpers that make developing apps easier.

Starting with a helper for AI SDK and TanStack AI. It lets you write a conversation in code, then run it through the real useChat lifecycle without a model, API route, network request, or API key.

Your UI receives native framework messages and streaming events, so reasoning, tool calls, loading states, and message components behave the same way they do in production.

Here is a conversation written with the AI SDK helper:

import { createChat } from "@shadcn/helpers/ai-sdk"

const chat = createChat()
  .user("What changed in this release?")
  .assistant("The release adds keyboard shortcuts and faster search.")
  .user("Can you check the full release notes?")
  .sleep(500)
  .assistant(({ writer }) => {
    writer.stepStart()

    // Reasoning.
    writer.reasoning("I should read the release notes first.")

    // Tool call.
    writer
      .tool("getReleaseNotes", {
        title: "Reading release notes",
        input: { version: "latest" },
      })
      .sleep(500)
      .output({
        highlights: ["Keyboard shortcuts", "Faster search"],
      })

    // Source.
    writer.sourceUrl({
      title: "Release notes",
      url: "https://example.com/releases",
    })

    // Response.
    writer.text("The release adds keyboard shortcuts and faster search.")
  })

Pass its messages and transport to useChat:

import { useChat } from "@ai-sdk/react"

const initialMessages = chat.get(0)
const transport = chat.transport()

export function Chat() {
  const { messages } = useChat({
    messages: initialMessages,
    transport,
  })
}

The package includes two adapters:

• @shadcn/helpers/ai-sdk plugs into AI SDK's useChat as a transport.
• @shadcn/helpers/tanstack-ai plugs into TanStack AI's useChat as a connection and streams real AG-UI events.

Use it to build components, create previews and demos, write documentation, or test streaming interfaces with deterministic output and no backend dependency.

[

Read the Docs

](https://ui.shadcn.com/docs/helpers/ai-sdk)