Inquiry About Becoming an Inference Provider on Hugging Face [GMI Cloud]

🤗Hub
1

Hi Hugging Face team,

I’m Kyle Shao from GMI Cloud. Our team has been working on building a stable inference engine designed to make people’s lives more efficient through a single, unified API. To reach more users, we’re very interested in becoming one of the inference providers on Hugging Face.

I’ve cloned the official repository and added the required components. However, I noticed that providers need to expose an API for users to check which models are currently supported. Could you please advise on the next steps or documentation needed to obtain the endpoint for adding model mappings?

Thank you for your time and support!

Best regards,

Kyle Shao

GMI Cloud

1 Like
2

Hmm… Maybe try these steps?

3

Hi, I followed the steps outlined on the page.

For Step 3, it says we need to use the endpoint /api/partners/{provider}/models to update the model mapping. I’m wondering whether our organization needs to apply for access to this endpoint, since I assume this API is created and managed by Hugging Face.

Could you clarify how we can get permission to use this endpoint?
Thanks for the help:)

1 Like
4
  1. Submit a PR for JS client integration - Add your provider to huggingface.js (see JS Client Integration).

Once the PR for Step 2 is merged, it should become available at that URL. If you want a faster merge, I think the most reliable approach is to consult the Hugging Face team via the Inference Providers channel on the Hugging Face Discord.

1 Like
5

So, just to confirm—does this mean we can only update the model mappings after our PR is merged by the Hugging Face team?

Thanks!

1 Like
6

does this mean we can only update the model mappings after our PR is merged by the Hugging Face team?

Yeah. I think that’s how it was when I read a similar inquiry before. I’m not sure if it’s changed or not…

1 Like
7

I see.
Appreciate for the explanation!
I will submit the PR later.

1 Like