Review Pull Requests with locally hosted LLMs by sending the diff and additional context. The LLM will generate a thorough review, offering suggestions and sample code.