Feat/swappable modes

[GoudaCouda]

Added a dropdown in the UI header to switch between different system prompt modes (like Conversation, Coder, etc.).

I was getting annoyed when I was trying to get info on something but it would only answer with a couple words and ran me in circles. Instead of needing to change via the txt file every time, this should be much easier.

I added a few just super quick example prompts.

• Prompt txt files go in code/prompts folder (which you can set in server.py).
• The server finds these automatically at startup.

UI Changes:

• Added a <select> dropdown to index.html (in the header).
• It gets populated by calling the /get_modes API endpoint.

Backend Logic:

• Moved all the prompt loading/switching logic into SpeechPipelineManager.
• server.py now just calls methods on the manager via /get_modes and /set_mode API endpoints.
• Cleaned up old code that loaded code/system_prompt.txt.

I also have a placeholder settings button in there for adding the ability to edit prompts, change models, or voices or such. I wasn't going to leave it in because I didn't do the logic yet, but then the UI looked weird so I just left it.

[KoljaB]

Looks very useful, thx a lot, will look into it soon

[NiaRiver]

You should not comment on unneeded code. If you want to recover it then you should use the commits history instead of commenting on unused rows.