0

So I'm currently working on a google assistant project, with dialogflow, firebase and google storage, and thus far I have a conversational agent that is working, but after searching the whole day a way to play .mp3 files stored in my google storage bucket, I'm still helpless.

Here's what the intent is supposed to do : 

conv.ask(
`<speak>
  <audio src="https://storage.cloud.google.com/path_to_my_bucket/mp3_file_name">
    Couldn't read the mp3 file !
  </audio>
 </speak>`);

Unfortunately, the sound is not played, and I got the 'Couldn't read the mp3 file !' message instead. The mp3 file is conform to the requirement in the DialogFlow documentation

Here is the response :

{
  "payload": {
    "google": {
      "expectUserResponse": true,
      "richResponse": {
        "items": [
          {
            "simpleResponse": {
              "textToSpeech": "<speak><audio src=\"https://storage.cloud.google.com/path_to_my_bucket/mp3_file_name\">Couldn't read the mp3 file !</audio></speak>"
            }
          }
        ]
      }
    }
  },
...

I tried with the https://console.actions.google.com/ test platform, on all the devices available.

This is not an authorization problem : i set all my files as public in my google storage bucket, (and that's why I obviously didn't type the real audio file link...)

Prisoner
  • 49,922
  • 7
  • 53
  • 105
Lucas Rioust
  • 1
  • 1
    Does your audio URL contain any special characters like `?` Like https://developers.google.com/assistant/actions/reference/ssml#urls_in_ssml? – Sairaj Sawant May 27 '20 at 14:17
  • Thanks for your answer, but there is not a single ampersand in the URL so there shouldn't be any XML formatting issue. – Lucas Rioust May 28 '20 at 07:55

1 Answers1

1

Unfortunately, I'm unable to identify the exact issue but it looks like you're addressing a few of the requirements already for playing audio files, such as HTTPS and publicly accessible.

Please check if your mp3 file meets all of the requirements in this checklist

https://developers.google.com/assistant/conversational/df-asdk/ssml#audio

mandychannyc
  • 115
  • 6