support local llm using ollama, lm-studio or any other OpenAPI-compatible local servers

[b3nab]

hardcoded local server URL is: http//localhost:1234 - default from LM Studio

• imported partial working code from Yhozen:feat/ai-sdk
• add new model-context
• extension-model ui component
• use model context inside chat and summary
• add new background port to auto-retrieve available models from local server
• improve select ui adding tabs for Online and Local models

Briefly I took the work you were already doing with @Yhozen from the Yhozen:feat/ai-sdk branch and extended it to support a local server. I didn't went down the path to fully create an interface to manage settings like the url of the server (or other things like editing the prompts with proper {variables} for transcript, video title etc..) so I hardcoded it to the default server url from LM Studio.

I wanted to share this, feel free to do whatever you want with this code. 😆

Great work done with this extension @PaoloJN , this local llm feature is just the cherry on top 🍒 .

[PaoloJN]

Thank you for your contribution @b3nab! The updates you've made, the integration of the local server support, look great. I really appreciate the way you've extended the work from the @Yhozen Yhozen:feat/ai-sdk branch to enhance the model management capabilities.

I'll take a closer look at the changes soon and plan to merge them afterward. You've done a fantastic job with this great work! 🍒.

Thanks @Yhozen & @b3nab : )