[FEEDBACK] Inference Providers
Any inference provider you love, and that you'd like to be able to access directly from the Hub?
Love that I can call DeepSeek R1 directly from the Hub 🔥
from huggingface_hub import InferenceClient
client = InferenceClient(
provider="together",
api_key="xxxxxxxxxxxxxxxxxxxxxxxx"
)
messages = [
{
"role": "user",
"content": "What is the capital of France?"
}
]
completion = client.chat.completions.create(
model="deepseek-ai/DeepSeek-R1",
messages=messages,
max_tokens=500
)
print(completion.choices[0].message)
Is it possible to set a monthly payment budget or rate limits for all the external providers? I don't see such options in billings tab. In case a key is or session token is stolen, it can be quite dangerous to my thin wallet:(
@benhaotang you already get spending notifications when crossing important thresholds ($10, $100, $1,000) but we'll add spending limits in the future
@benhaotang you already get spending notifications when crossing important thresholds ($10, $100, $1,000) but we'll add spending limits in the future
Thanks for your quick reply, good to know!
Would be great if you could add Nebius AI Studio to the list :) New inference provider on the market, with the absolute cheapest prices and the highest rate limits...
Could be good to add featherless.ai
TitanML !!
Hi Hugging Face provider-onboarding team — following up on PyAI’s existing Hear/Speak request. Could you help us identify the current blocker and the next concrete qualification step?
Before we implement a native provider adapter, we need confirmation of API-only ASR/TTS eligibility, the required Hub model mappings, the ASR/TTS request and response contract, request-cost billing requirements, and any organization-plan prerequisites. Our working direct integrations are linked above; we understand they do not establish native Inference Providers approval or routing.
Is there a technical owner who can confirm the checklist and tell us what evidence or corrections are needed from PyAI? If audio-provider onboarding is currently unavailable or the request is simply queued, please let us know so we can plan appropriately. Thank you!
Runcrate: request to join Inference Providers when onboarding reopens
Hey Hugging Face team,
Jish here, founder of Runcrate (runcrate.ai). We run an OpenAI-compatible inference API for open-weight models, and we've built the provider side to your spec: the billing endpoint answers cost per request id in nano-USD, and our models list carries live pricing.
Tasks we'd bring: chat (LLM and VLM), feature extraction, text to image and speech to text.
We saw on huggingface.js#2698 that new-provider onboarding is paused for now. Could you keep us on the list for when it opens again? Happy to open the JS client PR whenever suits you.
Best,
Jish
Montagea — Request to join Hugging Face Inference Providers (FastH3 Omni)
Hi Hugging Face team, @Wauplin @SBrandeis ,
We are building Montagea, an AI video creation platform. Our API is deployed, and we'd like to join Inference Providers, initially with FastH3 Omni.
Website: https://montagea.com
API docs: https://montagea.com/docs/api
HF organization: https://huggingface.co/Montagea-com
Contact: tim@montagea.com
Initial model:
FastVideo/FastVideo-FastH3-Omni-8-Step-V1
Our API model ID: minimax-h3
Infrastructure:
We operate GPU instances on Google Cloud using NVIDIA RTX PRO 6000 Blackwell GPUs and UniServe.
Capabilities:
- Text-to-video and image-to-video
- Reference image, video and audio inputs
- Asynchronous video submission, polling and retrieval
Our implementation includes:
- Provider API: https://api.montagea.com/api/partners/huggingface/v1
- Inference-Id response headers
- Per-request billing lookup in integer nano-USD
- Local JavaScript/Python adapters; upstream SDK registration is pending
Could you please confirm:
- Are you currently onboarding new video inference providers?
- How should FastH3 Omni's reference inputs map to HF tasks and SDK parameters?
- What timeout and validation rules apply to asynchronous video generation?
- What are the current organization-plan, settlement and server-side registration requirements?
We can arrange a scoped evaluation key privately and prepare the required SDK PRs, model mappings and documentation.
Thank you!
Hi Hugging Face team,
I'm Olumide Ola, founder of Olaverse Lab. We publish open-weight models on the Hub (huggingface.co/olaverse), including MIST-Extract-1.0, a 4B document-to-JSON extractor with calibrated per-field confidence, and MIST-QG, multilingual question generation. Both are built for English and African-language documents.
We run a production API for these models (api.olaverselab.com) with an OpenAI-compatible /v1/chat/completions endpoint, on dedicated GPUs with failover.
We'd like to be listed as an Inference Provider for our own models, so users on our model pages can call them through the router. We're happy to do the huggingface_hub and huggingface.js pull requests, the model mappings and the docs.
Two questions before we start:
Do you onboard providers that serve only their own models?
What do you need from us on billing and scale?
Thanks,
Olumide Ola, Olaverse Ltd
Hello Hugging Face Team,
I'm Trista Tsai from IntelliTrain, a Taiwan-based AI technology company.
We are interested in joining Hugging Face Inference Providers as an AI API provider.
We have reviewed the official provider registration documentation and would like to connect with the team responsible for provider onboarding and commercial partnerships.
Could you please direct us to the appropriate contact person or email address?
We would also appreciate information about the onboarding process, expected launch timeline, commercial terms, and potential API traffic.
Thank you for your assistance!
Best regards,
Trista Tsai
IntelliTrain
https://intelli-train.ai/
Alfred – a voice agent for everyday users, built on Inference Providers + HF sign-in
Hi HF team and everyone!
I'd like to share an app built entirely around Inference Providers and say thanks for making it possible.
Alfred is a friendly, animated voice assistant for Android. Users sign in with their Hugging Face account (OAuth + PKCE). From then on, every model call goes through the Inference Providers router with their own token, so they need no API keys and no extra accounts.
What it does with the Hub:
- Chat with any Hub model, with a focus on regional models. For Polish, it automatically recommends the newest official Bielik (SpeakLeash) available on the router
- Image generation & editing: Z-Image-Turbo, Qwen-Image, Qwen-Image-Edit (fal-ai queue mode)
- Voice in/out with Whisper + Piper (sherpa-onnx), 10 UI languages, a voice translator, web search, PDF/Word documents and MCP servers
Alfred is in closed testing on Google Play. If anyone from the team wants to try it, send me a DM with the Google account email to add to the tester list.
We'd love to talk about collaboration too: a showcase or case study, feedback on the integration, or early access to new regional models. Who would be the right person to contact?
Thanks!
Marcin Wawrzyniak, DevPartner (Poland)
FlexAI: request to join Inference Providers (30+ open models on global and EU endpoints)
Hi Hugging Face team,
Sundar here, COO at FlexAI. Our platform at flex.ai offers a token factory with an OpenAI-compatible inference API for open models. We'd like to join your Inference Providers and are happy to do all the integration work on our side. I saw the "Register as an Inference Provider" guide was retired in late October. Is onboarding paused, or is there a new route for providers now that chat goes through the router? Any pointer would help, and in the meantime here is what we'd bring.
Endpoints:
Global: https://api.flex.ai/v1 (30+ models across chat/vision, embeddings, rerank, moderation, STT, TTS, image and video generation)
EU: https://eu.api.flex.ai/v1 (17 models served in the EU, for teams that need EU data residency)
Capabilities on the chat models: streaming, tool calling, structured outputs, reasoning with effort, image input on the vision models, prompt caching with a separate cached-input price, and contexts up to 1,048,576 tokens.
Scale and trust: 99% uptime, SOC 2/GDPR, and a DPA with zero data retention and no training on customer data.
Models: flex.ai/models
- DeepSeek: V4.1 Flash 1017, V4 Flash 0731
- Z.ai GLM: GLM-5.3 Flash, GLM-5.2, GLM-4.5 Air
- Xiaomi: MiMo V2.6 Flash
- MiniMax: M3.1 Flash, H3 (video)
- Qwen: Qwen3.8 27B, Qwen3.8 Flash Next, Qwen3.6 27B and 35B-A3B, Qwen3.5 9B, Qwen3 Coder 30B, Qwen3 30B Thinking, Qwen3 Reranker 8B
- Google: Gemma 4 31B and 26B-A4B
- OpenAI: gpt-oss-120b, gpt-oss-20b, Whisper large-v3 turbo
- Meta: Llama 3.3 70B, Llama 3.1 8B, Llama Guard 4
- NVIDIA: Nemotron 3.5 Lightning 30B, Parakeet TDT 0.6B v3
- Also: Mistral Nemo, Muse Glimmer 30B, PaddleOCR-VL, BAAI bge-m3, Kokoro 82M (TTS), FLUX.1 schnell (image)
Pricing: flex.ai/pricing
Sign-up/Onboarding: platform.flex.ai
Docs: docs.flex.ai/inference-api/overview
What we'll set up for Inference Providers: a Hub-shaped model list with pricing and context_length per model (our catalog already carries both), a per-request ID header on every response, the per-request billing endpoint, model mappings for any of the models above, and the client-side changes if any are still needed after the router change. Our catalog moves quickly: new open releases usually land on day0.
Look forward to hearing from you,
Thanks,
Sundar, FlexAI
Hey Julien,
I’m Hugo, CGO at aimlapi.com. We give developers access to 10,000+ models and MCP tools through one API with zero markup, and we’re trusted by 500k+ users.
We’d love to explore serving open-weight models through Hugging Face Inference Providers. We could start with a pilot on a handful of models through our OpenAI-compatible API, with streaming, tool calling, and automatic failover across providers.
Who’s the best person on the Hugging Face team to talk to about onboarding AIMLAPI as an inference provider?




