Let ChatGPT Deploy Generated Files Directly to Vercel

Feature request: allow the ChatGPT/Vercel integration’s deploy action to accept a generated project folder or ZIP as input.
I use ChatGPT to generate complete static web builds for an existing Vercel project. The current integration can inspect deployments and expose a deploy action, but the deploy action does not accept a file bundle generated inside ChatGPT. This forces me to manually download the ZIP, open Vercel Drop, reconnect environment variables/storage, redeploy, and reconnect the production domain each time.

Vercel already supports programmatic deployments through the REST API and CLI. Could the ChatGPT Vercel connector expose a deployment action that accepts either:

  1. a ZIP/folder generated by ChatGPT, or
  2. a list of files plus project/team ID,
    and then create a preview or production deployment directly?

Ideally it would also allow deployment to an existing project while preserving existing environment variables, Blob bindings, Mux variables and domain settings.