BentoML

Right; that example isn't made to be run with flask, you should be running `bentoml serve`  instead!

ah right. do i then have to package it into a bento before i can serve it? or is there an example to serve it alongside the rest of my flask app?

Nope, you can simply serve it with `bentoml serve` without packaging if that's what you'd like. Because `bentoml` does a little bit more setup than a normal flask app, we don't have a good solution for running one without the `bentoml` CLI, though there is the `bentoml.HTTPServer` API (which uses the CLI under the hood).

i want to serve multiple different models in a single web server each on their own endpoints. whats the recommended approach for that?

If that's all you need, you can simply define multiple runners and use them in each endpoint!

because right now it seems like i have to define the service in a bentofile.yml file. would i then have to create one bentofile.yml for each service i create?

Er, a service is a single server, you would only have to define one.

oh i see so its not a single runner per service then? i can add every single runner i need into the service inside the runners array and use them individually in each of the endpoint?