smoothStream()

smoothStream is a utility function that creates a TransformStream for the streamText transform option to smooth out text and reasoning streaming by buffering and releasing complete chunks with configurable delays. This creates a more natural reading experience when streaming text and reasoning responses.

1
import { smoothStream, streamText } from 'ai';
2

3
const result = streamText({
4
  model,
5
  prompt,
6
  experimental_transform: smoothStream({
7
    delayInMs: 20, // optional: defaults to 10ms
8
    chunking: 'line', // optional: defaults to 'word'
9
  }),
10
});

Import

import { smoothStream } from "ai"

API Signature

Parameters

Word chunking caveats with non-latin languages

The word based chunking does not work well with the following languages that do not delimit words with spaces:

• Chinese
• Japanese
• Korean
• Vietnamese
• Thai

Using Intl.Segmenter (recommended)

For these languages, we recommend using Intl.Segmenter for proper locale-aware word segmentation. This is the preferred approach as it provides accurate word boundaries for CJK and other languages.

1
import { smoothStream, streamText } from 'ai';
2

3
const segmenter = new Intl.Segmenter('ja', { granularity: 'word' });
4

5
const result = streamText({
6
  model: "anthropic/claude-sonnet-4.5",
7
  prompt: 'Your prompt here',
8
  experimental_transform: smoothStream({
9
    chunking: segmenter,
10
  }),
11
});

1
import { smoothStream, streamText } from 'ai';
2

3
const segmenter = new Intl.Segmenter('zh', { granularity: 'word' });
4

5
const result = streamText({
6
  model: "anthropic/claude-sonnet-4.5",
7
  prompt: 'Your prompt here',
8
  experimental_transform: smoothStream({
9
    chunking: segmenter,
10
  }),
11
});

Regex based chunking

To use regex based chunking, pass a RegExp to the chunking option.

1
// To split on underscores:
2
smoothStream({
3
  chunking: /_+/,
4
});
5

6
// Also can do it like this, same behavior
7
smoothStream({
8
  chunking: /[^_]*_/,
9
});

Custom callback chunking

To use a custom callback for chunking, pass a function to the chunking option.

1
smoothStream({
2
  chunking: text => {
3
    const findString = 'some string';
4
    const index = text.indexOf(findString);
5

6
    if (index === -1) {
7
      return null;
8
    }
9

10
    return text.slice(0, index) + findString;
11
  },
12
});

Returns

Returns a TransformStream that:

• Buffers incoming text and reasoning chunks
• Releases content when the chunking pattern is encountered
• Adds configurable delays between chunks for smooth output
• Passes through non-text/reasoning chunks (like tool calls, step-finish events) immediately