[New Feature] Add direct support for Local Model Servers (Ollama, Lemonade, LM Studio)

[huhlig]

It would be great to add support for local/self-hosted LLM providers such as:

Ollama
LM Studio
Lemonade

This would allow NeverWrite to work with fully local models in addition to cloud APIs.

Why This Would Help

Many users are moving toward local inference for:

Privacy and data ownership
Offline usage
Lower API costs
Faster experimentation
Self-hosted/enterprise environments

NeverWrite looks to be an excellent upgrade to Obsidian and local model support would make it even more flexible for those with privacy concerns.

[jsgrrchg]

This will be as open as it gets. We are just getting started. I've been very busy with code, I will upload a proper roadmap.

Contributors Guys please feel free to open a PR on this topic so we can itinerate, my hands are full with other ACP, I need help to make this one happen.