Support for OpenAI compatible LLM interfaces

[elasticdotventures]

I'm creating this ticket, hoping it will be straightforward to add support for OpenAI compatible API's & services and non-local GPU's.

add support for non openai providers to surfkit, the ability to specify the URL of the openAI API provider to support self-hosting
ollama, vllm, openrouter, etc.

I especially like how openAi Codex supports alternate providers (I would copy/emulate the codex config.json approach) -- see: https://github.com/openai/codex?tab=readme-ov-file#custom-ai-provider-configuration

examples of custom providers:

• openrouter
• LiteLLM
• vllm
• ollama
• ramalama
• azure openai service

[elasticdotventures]

I've submitted #130 pull request 129-support-for-openai-compatible-llm-interfaces