diff --git a/docs/projektplan.md b/docs/projektplan.md
index aba0416..73ed25d 100644
--- a/docs/projektplan.md
+++ b/docs/projektplan.md
@@ -536,7 +536,7 @@ OLLAMA_MODEL=gemma3:4b
 OLLAMA_KEEP_ALIVE=0
 
 # LLM Provider (OpenAI-kompatibel, z.B. Ionos)
-LLM_URL=https://openai.ionos.com/openai
+LLM_URL=https://openai.inference.de-txl.ionos.com
 LLM_API_KEY=...
 LLM_MODEL=...
 LLM_TEMPERATURE=0
diff --git a/server/.env.example b/server/.env.example
index 3e93a22..99b569c 100644
--- a/server/.env.example
+++ b/server/.env.example
@@ -24,7 +24,7 @@ OLLAMA_MODEL=gemma3:4b
 OLLAMA_KEEP_ALIVE=0
 
 # LLM provider (Ionos)
-LLM_URL=https://openai.ionos.com/openai
+LLM_URL=https://openai.inference.de-txl.ionos.com
 LLM_API_KEY=
 LLM_MODEL=
 LLM_TEMPERATURE=0
diff --git a/server/src/main.rs b/server/src/main.rs
index 8c40e5b..2a5ab08 100644
--- a/server/src/main.rs
+++ b/server/src/main.rs
@@ -3,7 +3,7 @@ use std::sync::Arc;
 use tokio::net::TcpListener;
 use tower_http::limit::RequestBodyLimitLayer;
 use tower_http::trace::TraceLayer;
-use tracing::info;
+use tracing::{info, warn};
 use tracing_subscriber::layer::SubscriberExt;
 use tracing_subscriber::util::SubscriberInitExt;
 use tracing_subscriber::EnvFilter;
@@ -40,6 +40,14 @@ async fn main() {
         )
         .init();
 
+    // Surface configuration gaps that silently disable features.
+    if !config.llm_configured() {
+        warn!(
+            "LLM not configured — 'Fall abschließen' is disabled. \
+             Set LLM_URL, LLM_API_KEY and LLM_MODEL in .env to enable."
+        );
+    }
+
     // Shared HTTP client for both downstream pipelines.
     let http_client = reqwest::Client::builder()
         .build()