diff --git a/examples/python/agency-swarm/local_open_source_models.py b/examples/python/agency-swarm/local_open_source_models.py
index 46e2898..bed5b45 100644
--- a/examples/python/agency-swarm/local_open_source_models.py
+++ b/examples/python/agency-swarm/local_open_source_models.py
@@ -6,9 +6,8 @@
 load_dotenv("./.env")
 load_dotenv("../../../.env")
 
-client = patch(OpenAI(default_headers={"LLM-PARAM-base-url": "http://localhost:11434"}))
-# if using docker-compose, pass custom header to point to the ollama container instead of localhost
-# client = patch(OpenAI(default_headers={"LLM-PARAM-base-url": "http://ollama:11434"}))
+# remember to set OLLAMA_API_BASE_URL="http://ollama:11434" and base_url="http://localhost:8000/v1" in your env
+client = patch(OpenAI())
 
 set_openai_client(client)
 
@@ -25,4 +24,4 @@
 print(assistant)
 
 completion = agency.get_completion("What's something interesting about language models?")
-print(completion)
\ No newline at end of file
+print(completion)