Generating Multiple Text-to-Speech Fails in AI Explorer
>>> [!note] Migrated issue
<!-- Drupal.org comment -->
<!-- Migrated from issue #3490970. -->
Reported by: [jbuttler](https://www.drupal.org/user/3437983)
Related to !319
>>>
<p>When <strong>not </strong>using the 'Save as media' feature, I am unable to generate more than one tts file reliably.</p>
<p>I am able to successfully generate audio files when I use the feature, but there is some kind of caching which keeps giving me the same audio file. I am currently unable to generate a new tts file after the one I created a couple of days ago.</p>
<p>I confirmed that my requests are being fulfilled by openAI, the explorer waits for the timeout, thinks its successful, then plays the previously generated file.</p>
<p>This is easy to reproduce on my site:<br>
1. Enter a prompt<br>
2. Do Not use 'Save as media' feature<br>
3. Press 'Generate an audio response' and wait<br>
4. Successfully receive the requested tts file.<br>
5. Enter a new prompt<br>
6. 'Generate an audio response' <em>(without 'Save to media')</em><br>
7. Wait a few seconds<br>
8. Attempt appears successful.<br>
9. Playing the file reveals its the same as the first file.</p>
issue