How to stream results back with using an LLM api directly?

[Godrules500]

Since the vercel ai sdk does not have the ability yet to handle other types of grounding provided by vertex ai, I am going to call the Vertex AI api directly. With that, I want to still stream the response back to the front end. How can I do this with the tools provided via the ai sdk?

So the workflow would be:
user asks a question --> pre processing determines that we have a vector db for this question --> call vertex ai api directly --> stream the response back to the ai sdk UI.