diff --git a/litellm/caching/redis_semantic_cache.py b/litellm/caching/redis_semantic_cache.py
index b892e4c3a1d..bf0cba434e5 100644
--- a/litellm/caching/redis_semantic_cache.py
+++ b/litellm/caching/redis_semantic_cache.py
@@ -81,8 +81,10 @@ class RedisSemanticCache(BaseCache):
                 embedding; defaults to the Router deployment's configured max_input_tokens
             embedding_timeout: Seconds a cache lookup may spend embedding the prompt before it
                 gives up and lets the request continue to the LLM
-            ttl: Default time-to-live for cache entries in seconds
-            **kwargs: Additional arguments passed to the Redis client
+            **kwargs: Additional constructor arguments are accepted but currently
+                unused. Cache-entry TTL is not a constructor setting: _get_ttl reads
+                a `ttl` from the kwargs of each set_cache/async_set_cache call, so a
+                TTL applies to that individual cache write only
 
         Raises:
             Exception: If similarity_threshold is not provided or required Redis