Use ollama tool calling support to make local llm talk to internet API for more accurate response - View it on GitHub