TwiML™ Voice: <Play>

The <Play> verb plays an audio file back to the caller. Twilio retrieves the file from a URL that you provide.

Verb Attributes

The <Play> verb supports the following attributes that modify its behavior:

Attribute Name Allowed Values Default Value
loop integer >= 0 1
digits integer >= 0, w no default digits for Play


The 'loop' attribute specifies how many times the audio file is played. The default behavior is to play the audio once. Specifying '0' will cause the <Play> verb to loop until the call is hung up.

Use it like so:

<?xml version="1.0" encoding="UTF-8"?>
    <Play loop="10"></Play>


The ‘digits’ attribute lets you play DTMF tones during a call.

For example, if you need to test an IVR system, you can use this feature to simulate digits being pressed to navigate through the menu options.

If you are dialling a phone number and need to play DTMF tones to enter the extension, you should use the sendDigits attribute for <Number>.


The "noun" of a TwiML verb is the stuff nested within the verb that's not a verb itself; it's the stuff the verb acts upon. These are the nouns for <Play>:

Noun Description
plain text The URL of an audio file that Twilio will retrieve and play to the caller.

Twilio supports the following audio MIME types for audio files retrieved by the <Play> verb:

MIME type Description
audio/mpeg mpeg layer 3 audio
audio/wav wav format audio
audio/wave wav format audio
audio/x-wav wav format audio
audio/aiff audio interchange file format
audio/x-aifc audio interchange file format
audio/x-aiff audio interchange file format
audio/x-gsm GSM audio format
audio/gsm GSM audio format
audio/ulaw μ-law audio format

Nesting Rules

You can't nest any verbs within <Play>. You can nest <Play> within a <Gather> verb, with one major exception - you can't play "digits" within a <Gather>.


Example 1: Simple Play

This TwiML document tells Twilio to download the cowbell.mp3 file and play the audio to the caller.

<?xml version="1.0" encoding="UTF-8" ?>

Example 2: Using digits

We are going to test our IVR menu to make sure users can navigate properly. We know that the length of the initial greeting and the menu number we need to enter. We can add a few leading 'w' characters to add a pause. Each 'w' character tells Twilio to wait 0.5 seconds instead of playing a digit. This lets you adjust the timing of when the digits begin playing to suit the phone system you are dialing.

<?xml version="1.0" encoding="UTF-8" ?>
    <Play digits="wwww3"></Play>

Hints and Advanced Uses

  • Twilio will attempt to cache the audio file the first time it is played. This means the first attempt may be slow to play due to the time spent downloading the file from your remote server. Twilio may play a processing sound while the file is being downloaded.
  • Twilio obeys standard HTTP caching headers. If you change a file already cached by Twilio, make sure your web server is sending the proper headers to inform us that the contents of the file have changed.
  • Audio played over the telephone network is transcoded to a format the telephone network understands. Regardless of the quality of the file you provide us, we will transcode so it plays correctly. This may result in lower quality because the telephone number does not support high bitrate audio.
  • High bitrate, lossy encoded files, such as 128kbps mp3 files, will take longer to transcode and potentially sound worse than files that are in lossless 8kbps formats. This is due to the inevitable degradation that occurs when converting from lossy compressed formats and the processing involved in converting from higher bit rates to low bit rates.
  • <Play>ing a file that is longer than 40 minutes can result in a dropped call. If you need to <Play> a file longer longer than 40 minutes, consider splitting it up into smaller chunks.