I don't get why a group of users that are willing to run their own LLMs locally and do not want to relay on centralized corporations like openAI or google prefer to discuss using a centralized site like Reddit
Decent enough for a model 50 times smaller than ChatGPT. I use orca_mini_3b.
Compile llama.cpp, download a small GGML LLM model and you will have a quite intelligent assiatant running into your phone.
I still don't understand why the developer does not consider the option of adding a custom API key
It is not that I think that are hard to refill, it is that I actually tried to refill them. Horrible experience. Also, even using a good coffee brand, the result is not optimal.
And you loose all the convenience of using capsules. They are dirty and difficult to refill, also the quality of the resulting coffee cup is horrible compared to nespresso capsules. A regular italian expresso machine is a better option: awesome quality, you can choose the coffee brand and it is easier to refill than a reusable capsule.
Do not do it. Get a cheap regular expresso instead. Nespresso is like printers, they seem cheap but the Cartridges/capsules are made of unicorn blood. Reusable capsules might be the solution but the quality of the coffee is awful if you use them and you lose the main point of the nespresso: convenience.