How to beginner · 3 min read

How to use Hugging Face dedicated endpoints

Quick answer

Use Hugging Face dedicated endpoints by sending HTTP POST requests to the model-specific URL with your API key in the Authorization header. Use Python's requests library or Hugging Face's huggingface_hub SDK to interact with these endpoints securely and efficiently.

PREREQUISITES

Python 3.8+
Hugging Face API key
pip install requests huggingface_hub

Setup

Install the required Python packages and set your Hugging Face API key as an environment variable for secure access.

bash

pip install requests huggingface_hub

Step by step

Use Python to call a Hugging Face dedicated endpoint by sending a POST request with your input data and API key.

python

import os
import requests

API_URL = "https://api-inference.huggingface.co/models/gpt2"
API_TOKEN = os.environ["HF_API_KEY"]

headers = {"Authorization": f"Bearer {API_TOKEN}"}

def query(payload):
    response = requests.post(API_URL, headers=headers, json=payload)
    response.raise_for_status()
    return response.json()

if __name__ == "__main__":
    data = {"inputs": "Can you explain how to use Hugging Face dedicated endpoints?"}
    output = query(data)
    print(output)

output

[{'generated_text': 'Can you explain how to use Hugging Face dedicated endpoints? Hugging Face provides dedicated endpoints for models that allow you to send requests directly to a hosted model for inference.'}]

Common variations

You can also use the huggingface_hub Python SDK for easier integration, or switch models by changing the endpoint URL.

python

from huggingface_hub import InferenceClient
import os

client = InferenceClient(token=os.environ["HF_API_KEY"])

response = client.text_generation(
    model="gpt2",
    inputs="Explain Hugging Face dedicated endpoints."
)
print(response)

output

[{'generated_text': 'Explain Hugging Face dedicated endpoints. Hugging Face offers dedicated endpoints to access models via API calls for inference.'}]

Troubleshooting

If you receive a 401 Unauthorized error, verify your API key is set correctly in the environment variable HF_API_KEY. For 503 errors, the model might be overloaded or unavailable; retry after some time.

✅

Key Takeaways

Use the Hugging Face model-specific URL with your API key in the Authorization header for dedicated endpoints.
The requests library or huggingface_hub SDK are the recommended ways to interact with these endpoints in Python.
Set your API key securely in environment variables to avoid exposing credentials in code.
Switch models by changing the endpoint URL or model parameter in the SDK calls.
Handle common HTTP errors by checking API key validity and model availability.

Verified 2026-04 · gpt2

Verify ↗