I started self-hosting LLMs and absolutely loved it

Raghav Sethi self-hosted LLMs using Ollama as the core engine and Open WebUI for the interface, running DeepSeek R1's 7B parameter model on a MacBook Air with an M1 chip and 16GB of memory. The setup worked well for lighter tasks like debugging code offline but did not fully replace cloud-based tools like ChatGPT for complex queries.