Install fails because API server fails and there is no graceful fallback

[DingoOz]

The tool onboard path is brittle. Currently the nvidia api key server is overloaded so no API key is available and this install fails without a api key.

Need a workaround/local install non API key.

[kjw3]

Latest install can take advantage of local ollama service if you are interested to test

[DingoOz]

Is it correct to assume it is this branch?: feature/local-nim-support

[DingoOz]

Actually I think it is feat/ollama-local-inference

[DingoOz]

I couldn't see any documentation to trigger local (API key free) onboarding. Tried to specify ollama with a flag and the onboard script is 100% the same as main, needs an API key

[chandler-barlow]

This was addressed in #1136 and by local ollama support.