Feature: Support for Remote / Private Ollama Server (External LLM Backend)

[OxCom]

Summary
Allow the application to connect to and use a remote/private Ollama server instead of running local models.

Details
In many environments (e.g., corporate laptops, low-spec machines), running local LLMs is not feasible due to:

• Insufficient CPU/GPU resources
• Limited RAM
• Restricted environments (corporate policies)

At the same time, users may have access to a dedicated internal Ollama server capable of running models efficiently.

The application should:

• Allow configuration of a remote Ollama server endpoint
• Send inference requests to the remote server
• Avoid downloading or running models locally when remote is configured