Hugging Face Inference Endpoints vs Northflank

Hugging Face Inference Endpoints

OAuth scopes separate reading endpoints from writing them, both OpenAPI documents are public, and the unauthenticated /v2/provider route lists every instance with its hourly price. An account needs a payment method and credits before the first deployment, no rate limits or SLA were found for the management API, and the docs price table disagrees with the live list in places.

Northflank

OpenAPI 3.0 with over 100 paths, enums and per_page, page and cursor on every list. No scale to zero for services; minimum instances must be at least 1.

Hugging Face Inference Endpoints

1. Call GET https://api.endpoints.huggingface.cloud/v2/provider first and pick an instance whose status is available. The docs table lists types the API marks deprecated or not available
2. Send X-Scale-Up-Timeout: 600 on requests to an endpoint that scales to zero, or handle 503 while the first replica starts
3. Set scaleToZeroTimeout yourself. The docs give a default of 1 hour and the OpenAPI document says 15 minutes
4. Pause or delete an endpoint when the job is done. Billing covers every minute a replica is initialising or running
5. Give the agent a fine-grained token or the read-endpoints scope unless it must deploy. Endpoints are private by default and take the same Hugging Face token as a bearer

Northflank

1. Issue the agent a team token under an API role limited to one project, not a personal token
2. Read x-ratelimit-remaining and wait x-ratelimit-reset seconds on a 429; the default is 1,000 calls an hour
3. Page lists with per_page up to 100 and the returned cursor instead of numbered pages
4. Use a job, not a service, for anything that finishes; services bill until paused or deleted
5. Fetch any docs page with .md appended to get Markdown