Add clarifai as Inference provider

[mogith-pn]

Hi Team,
Added clarifai as provider for inference.
Related PR - huggingface/huggingface.js#1781

[Wauplin]

Thanks for opening this PR @mogith-pn! Implementation can be simplified, otherwise looks good :)

[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]

Could you also run make style locally and commit the changes please? It looks like you have some code quality issues (likely due to a "autoformat on save")

[mogith-pn]

Thanks for the quick review @Wauplin . Addressed the comments.

[Wauplin]

Looks ready! 🤗

Will keep it open until clarifai is integrated as staging so we can test it "for real".

[Wauplin]

Hi @mogith-pn, I just pushed 32baaad to fix the error you were getting. The URL / route was not set correctly (didn't catch that at first^^).

I can now run the example locally without issues:

from huggingface_hub import InferenceClient


client = InferenceClient(provider="clarifai")

completion = client.chat.completions.create(
    model="openai/gpt-oss-20b",
    messages=[{"role": "user", "content": "What is the capital of France?"}],
)

print("Reasoning:")
print(completion.choices[0].message.reasoning_content)

print("Content:")
print(completion.choices[0].message.content)

Our latest automated health check on model 'openai/gpt-oss-20b' for provider 'clarifai' did not complete successfully.  Inference call might fail.
Reasoning:
The user asks: "What is the capital of France?" The answer: Paris. Provide concise answer.
Content:
The capital of France is **Paris**.

[Wauplin]

(failing CI is unrelated)

[mogith-pn]

Thank you so much @Wauplin !

[Wauplin]

Merging it now!