It appears that version 0.0.3 of the synthetic-data-kit is designed to work exclusively with vLLM servers for local LLM inference. At present, it does not support external API endpoints, as the tool is hardcoded to expect a locally running vLLM server.

When I run the tool with a custom configuration file (using the --config flag), it still attempts to connect to http://localhost:8000/v1 and returns a 401 error, as if the configuration is being ignored entirely.

Could you please clarify when support for external API calls—per the documentation—will be properly supported?