Use local faster-whisper only for transcription.

Drop cloud and Ollama engine choices so audio stays on-machine via Docker Whisper, and tighten orphan detection plus UI around a single local flow.

app/Http/Requests/TranscribeRecordingRequest.php

@@ -3,7 +3,6 @@
 namespace App\Http\Requests;
 
 use Illuminate\Foundation\Http\FormRequest;
-use Illuminate\Validation\Rule;
 
 class TranscribeRecordingRequest extends FormRequest
 {
@@ -17,27 +16,6 @@ class TranscribeRecordingRequest extends FormRequest
      */
     public function rules(): array
     {
-        return [
-            'driver' => ['required', Rule::in(['cloud', 'local', 'ollama'])],
-            'ollama_url' => [
-                Rule::requiredIf(fn () => $this->input('driver') === 'ollama'),
-                'nullable',
-                'url',
-                'regex:/^https?:\/\//i',
-            ],
-        ];
-    }
-
-    /**
-     * @return array<string, string>
-     */
-    public function messages(): array
-    {
-        return [
-            'driver.required' => 'Choose a transcription engine.',
-            'ollama_url.required' => 'Enter the URL of the Ollama host (OpenAI-compatible Whisper endpoint).',
-            'ollama_url.url' => 'Enter a valid URL, e.g. http://192.168.1.50:8000',
-            'ollama_url.regex' => 'The host URL must start with http:// or https://',
-        ];
+        return [];
     }
 }