diff --git a/plugins/festinger/festinger/main.py b/plugins/festinger/festinger/main.py
index 1362650..14c833d 100644
--- a/plugins/festinger/festinger/main.py
+++ b/plugins/festinger/festinger/main.py
@@ -1103,6 +1103,14 @@ async def openai_chat_completions(request: Request) -> Response:
         raise
 
 
+# Alias: some LiteLLM provider types (custom_openai, openai_like) omit the
+# /v1 prefix and post directly to /chat/completions.
+@app.post("/chat/completions")
+async def openai_chat_completions_no_prefix(request: Request) -> Response:
+    """Alias for /v1/chat/completions — handles LiteLLM providers that omit /v1."""
+    return await openai_chat_completions(request)
+
+
 # ---------------------------------------------------------------------------
 # /iknowthat — manual write path
 # ---------------------------------------------------------------------------
@@ -3070,6 +3078,14 @@ async def models_ui() -> str:
 @app.api_route("/{path:path}", methods=["GET", "POST", "PUT", "DELETE", "HEAD"])
 async def passthrough(path: str, request: Request) -> Response:
     cfg = request.app.state.yaml_config
+
+    # Some LiteLLM configurations build wrong paths like v1/messages/chat/completions
+    # (when api_base already includes /v1/messages) or v1/chat/completions without
+    # a leading slash.  Redirect all chat-completion variants to the proper handler.
+    if path.endswith("chat/completions") and request.method == "POST":
+        log.info("passthrough redirect  %s → /v1/chat/completions", path)
+        return await openai_chat_completions(request)
+
     if path.startswith("v1/"):
         upstream = cfg["upstream_anthropic"]
         relay_headers = ANTHROPIC_RELAY_HEADERS