Having spent the last couple of months tinkering with LLMs, I have to admit that these local models have boosted my productivity once I began pairing them with FOSS tools. So far, I’ve hooked my Ollama models to Blinko, Paperless-GPT, Karakeep, VS Code (via Continue.dev), and other self-hosted services, and their superior reasoning capabilities make LLMs perfect for tackling monotonous tasks.