When the end user is typing, each keystroke is passed to the autocomplete endpoint. As the end user continues to type, keep calling autocomplete, passing the entered text and the same token from step 1.
See the Autocomplete documentation for full details.
