AI
chat

Send Custom Body from useChat

experimental_prepareRequestBody is an experimental feature and only available in React.

By default, useChat sends all messages as well as information from the request to the server. However, it is often desirable to control the body content that is sent to the server, e.g. to:

  • only send the last message
  • send additional data along with the message
  • change the structure of the request body

The experimental_prepareRequestBody option allows you to customize the body content that is sent to the server. The function receives the message list, the request data, and the request body from the append call. It should return the body content that will be sent to the server.

Example

This example shows how to only send the text of the last message to the server. This can be useful if you want to reduce the amount of data sent to the server.

Client 

'use client';


import { useChat } from 'ai/react';


export default function Chat() {
  const { messages, input, handleInputChange, handleSubmit } = useChat({
    experimental_prepareRequestBody: ({ messages }) => {
      // e.g. only the text of the last message:
      return messages[messages.length - 1].content;
    },
  });


  return (
    <div>
      {messages.map(m => (
        <div key={m.id}>
          {m.role === 'user' ? 'User: ' : 'AI: '}
          {m.content}
        </div>
      ))}


      <form onSubmit={handleSubmit}>
        <input value={input} onChange={handleInputChange} />
      </form>
    </div>
  );
}

Server

We need to adjust the server to only receive the text of the last message. The rest of the message history can be loaded from storage.

import { openai } from '@ai-sdk/openai'
import { streamText } from 'ai'


// Allow streaming responses up to 30 seconds
export const maxDuration = 30


export async function POST(req: Request) {
  // we receive only the text from the last message
  const text = await req.json()


  // e.g. load message history from storage
  const history = await loadHistory()


  // Call the language model
  const result = streamText({
    model: openai('gpt-4-turbo'),
    messages: [...history, { role: 'user', content: text }]
    onFinish({ text }) {
      // e.g. save the message and the response to storage
    }
  })


  // Respond with the stream
  return result.toDataStreamResponse()
}