fix(backend): implement step 1 — llama config from env vars

backend/src/main/kotlin/com/condado/newsletter/service/LlamaPreviewService.kt

@@ -10,8 +10,8 @@ import org.springframework.web.client.RestClient
 @Service
 class LlamaPreviewService(
     private val restClient: RestClient,
-    @Value("\${llama.base-url:http://localhost:11434}") private val baseUrl: String,
-    @Value("\${llama.model:gemma3:4b}") private val model: String
+    @Value("\${llama.base-url}") private val baseUrl: String,
+    @Value("\${llama.model}") private val model: String
 ) {
 
     /**

backend/src/main/resources/application.yml

@@ -49,8 +49,8 @@ openai:
   model: ${OPENAI_MODEL:gpt-4o}
 
 llama:
-  base-url: ${LLAMA_BASE_URL:http://localhost:11434}
-  model: ${LLAMA_MODEL:gemma3:4b}
+  base-url: ${LLAMA_BASE_URL}
+  model: ${LLAMA_MODEL}
 
 springdoc:
   swagger-ui: