Fix LLM model list, Castopod API, and server runner

- Remove gpt-4o-realtime (WebSocket-only) from OpenRouter models
- Increase OpenRouter timeout to 60s and max_tokens to 150
- Handle empty LLM responses
- Fix publish_episode.py for current Castopod API fields
- Add port conflict check and graceful shutdown to run.sh

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/services/llm.py

@@ -7,10 +7,10 @@ from ..config import settings
 
 # Available OpenRouter models
 OPENROUTER_MODELS = [
-    "anthropic/claude-3-haiku",
-    "anthropic/claude-3.5-sonnet",
     "openai/gpt-4o-mini",
     "openai/gpt-4o",
+    "anthropic/claude-3-haiku",
+    "anthropic/claude-3.5-sonnet",
     "google/gemini-flash-1.5",
     "google/gemini-pro-1.5",
     "meta-llama/llama-3.1-8b-instruct",
@@ -114,7 +114,7 @@ class LLMService:
         """Call OpenRouter API with retry"""
         for attempt in range(2):  # Try twice
             try:
-                async with httpx.AsyncClient(timeout=30.0) as client:
+                async with httpx.AsyncClient(timeout=60.0) as client:
                     response = await client.post(
                         "https://openrouter.ai/api/v1/chat/completions",
                         headers={
@@ -124,12 +124,16 @@ class LLMService:
                         json={
                             "model": self.openrouter_model,
                             "messages": messages,
-                            "max_tokens": 100,
+                            "max_tokens": 150,
                         },
                     )
                     response.raise_for_status()
                     data = response.json()
-                    return data["choices"][0]["message"]["content"]
+                    content = data["choices"][0]["message"]["content"]
+                    if not content or not content.strip():
+                        print(f"OpenRouter returned empty response")
+                        return ""
+                    return content
             except (httpx.TimeoutException, httpx.ReadTimeout):
                 print(f"OpenRouter timeout (attempt {attempt + 1})")
                 if attempt == 0:

publish_episode.py

@@ -155,11 +155,12 @@ Respond with ONLY valid JSON, no markdown or explanation."""
     return metadata
 
 
-def create_episode(audio_path: str, metadata: dict, duration: int) -> dict:
+def create_episode(audio_path: str, metadata: dict, episode_number: int) -> dict:
     """Create episode on Castopod."""
     print("[3/5] Creating episode on Castopod...")
 
     headers = get_auth_header()
+    slug = re.sub(r'[^a-z0-9]+', '-', metadata["title"].lower()).strip('-')
 
     # Upload audio and create episode
     with open(audio_path, "rb") as f:
@@ -168,21 +169,25 @@ def create_episode(audio_path: str, metadata: dict, duration: int) -> dict:
         }
         data = {
             "title": metadata["title"],
-            "description_markdown": metadata["description"],
+            "slug": slug,
+            "description": metadata["description"],
             "parental_advisory": "explicit",
             "type": "full",
-            "created_by": "1"
+            "podcast_id": str(PODCAST_ID),
+            "created_by": "1",
+            "updated_by": "1",
+            "episode_number": str(episode_number),
         }
 
         response = requests.post(
-            f"{CASTOPOD_URL}/api/rest/v1/podcasts/{PODCAST_ID}/episodes",
+            f"{CASTOPOD_URL}/api/rest/v1/episodes",
             headers=headers,
             files=files,
             data=data
         )
 
     if response.status_code not in (200, 201):
-        print(f"Error creating episode: {response.text}")
+        print(f"Error creating episode: {response.status_code} {response.text}")
         sys.exit(1)
 
     episode = response.json()
@@ -312,7 +317,7 @@ def get_next_episode_number() -> int:
     headers = get_auth_header()
 
     response = requests.get(
-        f"{CASTOPOD_URL}/api/rest/v1/podcasts/{PODCAST_ID}/episodes",
+        f"{CASTOPOD_URL}/api/rest/v1/episodes",
         headers=headers
     )
 
@@ -323,7 +328,12 @@ def get_next_episode_number() -> int:
     if not episodes:
         return 1
 
-    max_num = max(ep.get("number", 0) for ep in episodes)
+    # Filter to our podcast
+    our_episodes = [ep for ep in episodes if ep.get("podcast_id") == PODCAST_ID]
+    if not our_episodes:
+        return 1
+
+    max_num = max(ep.get("number", 0) or 0 for ep in our_episodes)
     return max_num + 1
 
 
@@ -373,7 +383,7 @@ def main():
         return
 
     # Step 3: Create episode
-    episode = create_episode(str(audio_path), metadata, transcript["duration"])
+    episode = create_episode(str(audio_path), metadata, episode_number)
 
     # Step 4: Publish
     episode = publish_episode(episode["id"])

run.sh

@@ -1,6 +1,7 @@
 #!/bin/bash
 # AI Radio Show - Server Runner with restart support
 
+PORT=8000
 LOG_FILE="/tmp/ai-radio-show.log"
 RESTART_FLAG="/tmp/ai-radio-show.restart"
 STOP_FLAG="/tmp/ai-radio-show.stop"
@@ -13,16 +14,46 @@ source venv/bin/activate
 # Cleanup old flags
 rm -f "$RESTART_FLAG" "$STOP_FLAG"
 
+# Check if port is already in use
+if lsof -i ":$PORT" -sTCP:LISTEN -t >/dev/null 2>&1; then
+    EXISTING_PID=$(lsof -i ":$PORT" -sTCP:LISTEN -t 2>/dev/null | head -1)
+    echo "ERROR: Port $PORT is already in use by PID $EXISTING_PID"
+    echo "Run: kill $EXISTING_PID"
+    exit 1
+fi
+
+kill_server() {
+    local pid=$1
+    if ! kill -0 "$pid" 2>/dev/null; then
+        return
+    fi
+    kill "$pid" 2>/dev/null
+    # Wait up to 5 seconds for graceful shutdown
+    for i in $(seq 1 10); do
+        if ! kill -0 "$pid" 2>/dev/null; then
+            return
+        fi
+        sleep 0.5
+    done
+    # Force kill if still alive
+    echo "[$(date)] Server didn't stop gracefully, force killing..." | tee -a "$LOG_FILE"
+    kill -9 "$pid" 2>/dev/null
+    wait "$pid" 2>/dev/null
+}
+
 echo "AI Radio Show Server Runner"
 echo "Log file: $LOG_FILE"
 echo "Press Ctrl+C to stop"
 echo ""
 
-while true; do
+# Handle Ctrl+C
-    echo "[$(date)] Starting server..." | tee -a "$LOG_FILE"
+trap 'echo ""; echo "[$(date)] Interrupted" | tee -a "$LOG_FILE"; kill_server $SERVER_PID; exit 0' INT TERM
 
-    # Start uvicorn with output to both console and log file
+while true; do
-    python -m uvicorn backend.main:app --host 0.0.0.0 --port 8000 2>&1 | tee -a "$LOG_FILE" &
+    echo "[$(date)] Starting server on port $PORT..." | tee -a "$LOG_FILE"
+
+    # Start uvicorn directly (not through tee pipe so we get the real PID)
+    python -m uvicorn backend.main:app --host 0.0.0.0 --port $PORT >> "$LOG_FILE" 2>&1 &
     SERVER_PID=$!
 
     # Wait for server to exit or restart signal
@@ -30,8 +61,7 @@ while true; do
         if [ -f "$RESTART_FLAG" ]; then
             echo "[$(date)] Restart requested..." | tee -a "$LOG_FILE"
             rm -f "$RESTART_FLAG"
-            kill $SERVER_PID 2>/dev/null
+            kill_server $SERVER_PID
-            wait $SERVER_PID 2>/dev/null
             sleep 1
             break
         fi
@@ -39,8 +69,7 @@ while true; do
         if [ -f "$STOP_FLAG" ]; then
             echo "[$(date)] Stop requested..." | tee -a "$LOG_FILE"
             rm -f "$STOP_FLAG"
-            kill $SERVER_PID 2>/dev/null
+            kill_server $SERVER_PID
-            wait $SERVER_PID 2>/dev/null
             echo "[$(date)] Server stopped." | tee -a "$LOG_FILE"
             exit 0
         fi