[CLI] Add Inference Endpoints Commands

[hanouticelina]

This PR implements a CLI to manage Inference Endpoints, this provides "one liners" to deploy/delete/update/etc. endpoints, which could be handy in many cases. The DX intentionally mirrors a bit the UI instead of the API, to quote @ErikKaum :

we're renaming things in the UI quite fast to adapt and make things make more sense. And in many cases in the UI things are configured with slightly different names/groupings that in the API. Just because it's faster than in the API.

I explored a few layouts (e.g. a single deploy command with --catalog), but the cleanest UX ended up being two explicit paths:

• hf endpoints deploy ... – minimal set of hardware/task configs for Hub models.
• hf endpoints catalog deploy ... – one liner using optimized configs from the model catalog.
• delete and update endpoints currently live under the "Settings" group in the UI, but feels more natural to keep them top-level in the CLI 🤷‍♀️

> hf endpoints --help
Usage: hf endpoints [OPTIONS] COMMAND [ARGS]...

  Manage Hugging Face Inference Endpoints.

Options:
  --help  Show this message and exit.

Commands:
  catalog        Interact with the Inference Endpoints catalog.
  delete         Delete an Inference Endpoint permanently.
  deploy         Deploy an Inference Endpoint from a Hub repository.
  describe       Get information about an existing endpoint.
  ls             Lists all Inference Endpoints for the given namespace.
  pause          Pause an Inference Endpoint.
  resume         Resume an Inference Endpoint.
  scale-to-zero  Scale an Inference Endpoint to zero.
  update         Update an existing endpoint.


> hf endpoints catalog --help
Usage: hf endpoints catalog [OPTIONS] COMMAND [ARGS]...

  Interact with the Inference Endpoints catalog.

Options:
  --help  Show this message and exit.

Commands:
  deploy  Deploy an Inference Endpoint from the Model Catalog.
  ls      List available Catalog models.

happy to more iterate if there are more suggestions to make the DX better (and simpler?)

[hanouticelina]

splitted the deploy into two subcommands instead of using a flag to deploy from the Model Catalog because Typer doesn't easily allow conditional requirements (i.e., "these parameters are required unless that flag (e.g. --from-catalog) is set"), which makes validation and type hints messy so using subcommands lets Typer enforce required options cleanly for each case

[Wauplin]

what do you think of having

# deploy from Hub
hf endpoints deploy

# list catalog
hf endpoints catalog ls

# deploy from catalog
hf endpoints catalog deploy

?

[Wauplin]

otherwise if keeping the same I'd be more inclined for

hf endpoints deploy from-repo
hf endpoints deploy from-catalog

(since both are deployed from Hub + the from- makes it more explicit)

(but still slight preference for the one above)

[ErikKaum]

Sorry for late response but I'll propose a 3rd option 😄

for example simply:

hf endpoints deploy Qwen/Qwen3-Next-80B-A3B-Instruct

if it's in the catalog --> uses catalog configuration just deploys as such
if not --> then the user is prompted to configure the deployment themselves

It's a bit more logic but might be more clean 🤔 wdyt?

Honestly in the web UI as well, we're looking to make this a bit more streamlined to be just "deploy model" but have the catalog configurations as the top recommendations and highlighted. But not per se as completely separate deployment flows. If that makes sense.

[Wauplin]

The issue with "deploy" + prompt user for additional configuration is that it cannot be added to an automated script (or much more difficult). Same for documentation where it's harder to say "copy-paste and execute this command" to deploy the model.

[ErikKaum]

Could we make it more clear by not prompting but returning to std err? 🤔

Treat it in the same way as passing a configuration which is missing some fields for example:

hf endpoints deploy Qwen/Qwen3-Next-80B-A3B-Instruct

• if in catalog: succeeds with the catalog config
• fails if not in catalog, since there's no config

hf endpoints deploy Qwen/Qwen3-Next-80B-A3B-Instruct --hardware=H200

• if in catalog: overrides the hardware recommendation
• if not: fails also since it's not a complete config

---

But take my suggestion with a grain of salt. I don't know the ins and outs as well as you both do. I also think your earlier suggestion is a good improvement:

# deploy from Hub
hf endpoints deploy

# list catalog
hf endpoints catalog ls

# deploy from catalog
hf endpoints catalog deploy

[hanouticelina]

@Wauplin @ErikKaum
getting back to this as it would be cool to ship this for the next release! i went with the first suggestion because it feels the cleanest and the simplest to implement, with a clear behavior and no hidden logic. Since this CLI is still more of an exploration at this stage, i think it's better to ship like this (unless you have a strong opinion against 😄) but i guess we can always reassess later based on how people use it, wdyt?

> hf endpoints --help
Usage: hf endpoints [OPTIONS] COMMAND [ARGS]...

  Manage Hugging Face Inference Endpoints.

Options:
  --help  Show this message and exit.

Commands:
  catalog        Interact with the Inference Endpoints catalog.
  delete         Delete an Inference Endpoint permanently.
  deploy         Deploy an Inference Endpoint from a Hub repository.
  describe       Get information about an existing endpoint.
  ls             Lists all Inference Endpoints for the given namespace.
  pause          Pause an Inference Endpoint.
  resume         Resume an Inference Endpoint.
  scale-to-zero  Scale an Inference Endpoint to zero.
  update         Update an existing endpoint.


> hf endpoints catalog --help
Usage: hf endpoints catalog [OPTIONS] COMMAND [ARGS]...

  Interact with the Inference Endpoints catalog.

Options:
  --help  Show this message and exit.

Commands:
  deploy  Deploy an Inference Endpoint from the Model Catalog.
  ls      List available Catalog models.

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[Wauplin]

Thanks for working on this! Left some high level feedback mostly geared towards the CLI syntax. Let me know what you think :)

[Wauplin]

what do you think of having

# deploy from Hub
hf endpoints deploy

# list catalog
hf endpoints catalog ls

# deploy from catalog
hf endpoints catalog deploy

?

[Wauplin]

otherwise if keeping the same I'd be more inclined for

hf endpoints deploy from-repo
hf endpoints deploy from-catalog

(since both are deployed from Hub + the from- makes it more explicit)

(but still slight preference for the one above)

[Wauplin]

Left a comment about the try/except parts, apart from that looks good to me :) Thanks for getting back to this PR!