Proposal: multiple AI providers, auto-detect models

Currently, Joplin Desktop only allows 1 model in its configuration.

The only current workaround to switch provider or model is to change configuration. It is not very comfortable, because:

  • When changing provider, you have to be creating new API keys all the time, as they are usually only displayed once.
  • Also requires dealing too often with technical things like the model slug or the API provider URL.

Would it be possible to have multiple inference providers and to auto-detect available models when connected, please?

Thanks!