Using image style when configuring automator with vision results in an error
>>> [!note] Migrated issue
<!-- Drupal.org comment -->
<!-- Migrated from issue #3555656. -->
Reported by: [ultimike](https://www.drupal.org/user/51132)
>>>
<h3 id="summary-problem-motivation">Problem/Motivation</h3>
<p>When creating an automator with an "AI Provider" with vision, in the "Provider configuration", when I select any image style other than "Use original", get the following warning message (and no text suggestion is provided):</p>
<pre>Warning message<br>Error invoking model response: messages.0.content.1.image.source.base64.media_type: Input should be 'image/jpeg', 'image/png', 'image/gif' or 'image/webp'</pre><h4 id="summary-steps-reproduce">Steps to reproduce (required for bugs, but not feature requests)</h4>
<p>Drupal core 11.2.5, AI module 1.2.2, standard Article content type with image field. </p>
<p>Add a new "Text (plain, long)" field called "Image description" to the Article content type.</p>
<p>Add an AI automator to this field of type "LLM: Text", advanced input mode, prompt: "Please describe the provided image fully using no more than 250 words." </p>
<p>Advanced settings include using Field Widget, the default Vision model (I tested with both Claude Sonnet 4.5 gpt-4-turbo), Image field of "Image", and various image styles (only "Use original" worked for me.)</p>
<p>The field widget configuration for the "Image description" field is nothing special - just calls the automator. </p>
<p>-mike</p>
issue