question: Local REST API functionality clarification

[dylbarne]

potentially related to other issue question, is it intended for the Local REST API documented here that it can act as a proxy forwarding service any external api models that are available via the AI Toolkit Playground (example: github provided openai chatgpt gpt-4o) (not necessarily only local/downloaded/onyx model)

that was my assumption originally, although I am getting 400 Bad Request code back on requests to hosted models like this (suggestion if this is not the case, improve documentation to make explicitly more clear what local Rest Api capabilities are, or provide some additional error details in 400 response)

I had also noticed in other issue comment verbiage of "remote party requests renegotiation" so it made me think something to this affect

please advise, thanks

[microsoft-github-policy-service[bot]]

Thank you for contacting us! Any issue or feedback from you is quite important to us. We will do our best to fully respond to your issue as soon as possible. Sometimes additional investigations may be needed, we will usually get back to you within 2 days by adding comments to this issue. Please stay tuned.

[a1exwang]

@dylbarne Currently the local inference endpoint only support local ONNX models so passing in a remote model name will cause 400 error which is by design. This is a feature request.

[dylbarne]

awesome thank you for the clarification!

yes if this could be triaged as Feature Request and have this functionality developed, I think that would be great in assisting towards the usecase stated in the linked documentation of "local REST API web server allows you to build-and-test your application locally"