UNPKG

openclaw

Version:

Multi-channel AI gateway with extensible messaging integrations

github.com/openclaw/openclaw

openclaw/openclaw

22 lines (15 loc) • 763 B
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
# Hugging Face Use hosted models through Hugging Face Inference Providers. OpenClaw connects to Hugging Face's router and discovers available models; it does not download or run the model weights locally. ## Get started Create a Hugging Face token with permission to call Inference Providers, then run: ```bash openclaw onboard --auth-choice huggingface-api-key ``` The Gateway also accepts `HUGGINGFACE_HUB_TOKEN` or `HF_TOKEN`. Browse the catalog with `openclaw models list --provider huggingface`, then choose a model your account can access. Model availability and supported inputs depend on the selected inference provider. See the [Hugging Face guide](https://docs.openclaw.ai/providers/huggingface) for token setup, routing, and model configuration.