UI: posodobi zastarel opis (Whisper → Soniox + Claude)

User feedback: 'tukaj imava cel kup stvari ki niso res, kako oblikujemo?'

Old text was misleading:
- 'Whisper, 3-sample voting' → not used since Soniox integration
- 'Model: medium' → irrelevant (Whisper not used)
- 'Whisper + energy → najde refren' → now Soniox + Claude LLM

New text reflects actual stack:
- STT: Soniox (primary) → ElevenLabs Scribe → Gemini fallback
- LLM: Claude Sonnet 4.6
- Energy profile + word-level timestamps + 15 reference examples
- Mention ✏️ Edit button for manual fine-tuning

templates/index.html

@@ -356,16 +356,17 @@
       <input type="hidden" id="model" value="large-v3">
 
       <div style="font-size: 12px; color: var(--text-dim); margin-top: 8px;">
-        🤖 Jezik: avtomatsko zaznan (Whisper, 3-sample voting) · Model: medium · LLM analiza: Claude
+        🤖 STT: <b>Soniox</b> (primary) → ElevenLabs Scribe → Gemini fallback · LLM analiza: <b>Claude Sonnet 4.6</b>
       </div>
 
       <label class="toggle" style="margin-top: 16px;">
         <input type="checkbox" id="auto-chorus" checked>
-        Pametna izbira odseka (Whisper + energy → najde refren)
+        Pametna izbira odseka (Soniox + Claude LLM → najde refren)
       </label>
       <div style="font-size: 12px; color: var(--text-dim); margin-top: 4px; margin-left: 26px;">
-        Sistem naredi <b>celoten transkript</b> in <b>energy profile</b>, najde refren in ga izreže.
-        Audio fade in/out je avtomatsko dodan na meje vokala.
+        Sistem naredi <b>celoten transkript</b> z word-level timestampi in <b>energy profile</b>,
+        nato Claude izbere najboljši refren po pravilih (15 referenčnih primerov).
+        Audio fade in/out na meje vokala. Ročno popravljanje preko ✏️ <b>Edit</b> gumba.
       </div>
 
       <label class="toggle" style="margin-top: 12px; margin-left: 26px;">