fix(ai): prevent OpenAI credential error at module load time

- Moved OpenAI client initialization to a lazy getter function
- Added explicit check for NEBIUS_API_KEY in the getter
- Ensures Server Actions return structured errors instead of crashing when keys are missing

src/services/analyze-bottle-nebius.ts

@@ -1,6 +1,6 @@
 'use server';
 
-import { aiClient } from '@/lib/ai-client';
+import { getNebiusClient } from '@/lib/ai-client';
 import { SYSTEM_INSTRUCTION as GEMINI_SYSTEM_INSTRUCTION } from '@/lib/gemini';
 import { BottleMetadataSchema, AnalysisResponse, BottleMetadata } from '@/types/whisky';
 import { createServerActionClient } from '@supabase/auth-helpers-nextjs';
@@ -56,7 +56,8 @@ export async function analyzeBottleNebius(base64Image: string, tags?: string[],
         console.log(`[analyzeBottleNebius] Instruction prepared for AI: ${instruction.substring(0, 100)}...`);
 
 
-        const response = await aiClient.chat.completions.create({
+        const client = getNebiusClient();
+        const response = await client.chat.completions.create({
             model: "Qwen/Qwen2.5-VL-72B-Instruct",
             messages: [
                 {