In my previous post I discussed the benefits of using locally hosted open weights LLMs, like data privacy and cost savings. By using mostly free models and occasionally switching to GPT-4, my monthly expenses dropped from 20 USD to 0.50 USD. For mobile access, setting up a port-forward to your local LLM server is a free solution. There are many open-source tools for hosting open weights LLMs local
![50+ Open-Source Options for Running LLMs Locally](https://cdn-ak-scissors.b.st-hatena.com/image/square/cdeb0651562d0998f7d8ce6465c147e906499ac6/height=288;version=1;width=512/https%3A%2F%2Fmiro.medium.com%2Fv2%2Fresize%3Afit%3A1200%2F1%2ASmFROaOLlZDXxARO99QGyA.png)