Skip to main content

​
Hugging Face

Shanone’s Hugging Face integration gives your agent 87 tools covering nearly the entire Hub — models, datasets, spaces, serverless inference, inference endpoints, collections, discussions/PRs, papers, repository operations, security scans, and webhooks.

​
Getting Started

1

Get your Hugging Face Access Token

Generate a token from Settings → Access Tokens in your Hugging Face account.
2

Add it to Shanone

Open Integrations in the Shanone dashboard, select Hugging Face, and paste in your Access Token.
3

Retry your request

Once saved, shanone_execute_tool calls for huggingface_* tools will succeed.

​
Available Tools

Shanone provides 87 tools for Hugging Face, organized into these categories:
huggingface_list_models, huggingface_get_model_info, huggingface_list_user_models, huggingface_create_model_repo, huggingface_delete_model_repo, huggingface_update_model_settings
huggingface_list_datasets, huggingface_get_dataset_info, huggingface_create_dataset_repo, huggingface_delete_dataset_repo, huggingface_get_dataset_rows, huggingface_filter_dataset_rows, huggingface_search_dataset, huggingface_get_dataset_statistics, huggingface_get_dataset_splits, huggingface_get_dataset_size, huggingface_list_parquet_files
huggingface_list_spaces, huggingface_get_space_info, huggingface_create_space_repo, huggingface_delete_space_repo, huggingface_restart_space, huggingface_pause_space
huggingface_text_generation, huggingface_text_classification, huggingface_feature_extraction, huggingface_chat_completion, huggingface_image_classification, huggingface_summarization, huggingface_text_to_image, huggingface_image_to_text, huggingface_translation, huggingface_question_answering, huggingface_zero_shot_classification, huggingface_token_classification, huggingface_fill_mask, huggingface_automatic_speech_recognition
huggingface_list_inference_endpoints, huggingface_get_inference_endpoint, huggingface_create_inference_endpoint, huggingface_update_inference_endpoint, huggingface_delete_inference_endpoint, huggingface_pause_inference_endpoint, huggingface_resume_inference_endpoint, huggingface_scale_inference_endpoint
huggingface_create_branch, huggingface_delete_branch, huggingface_create_tag, huggingface_delete_tag, huggingface_list_repo_refs, huggingface_list_commits, huggingface_list_repo_files
huggingface_create_collection, huggingface_list_collections, huggingface_get_collection, huggingface_add_to_collection, huggingface_remove_from_collection, huggingface_delete_collection
huggingface_create_discussion, huggingface_list_discussions, huggingface_get_discussion, huggingface_create_discussion_comment, huggingface_change_discussion_status, huggingface_merge_pull_request
huggingface_search_papers, huggingface_get_daily_papers, huggingface_get_paper, huggingface_list_paper_comments, huggingface_create_paper_comment
huggingface_get_security_scan, huggingface_request_repo_access, huggingface_list_access_requests, huggingface_handle_access_request
huggingface_get_organization, huggingface_list_organization_members, huggingface_get_organization_repos, huggingface_get_user_info
huggingface_create_webhook, huggingface_list_webhooks, huggingface_get_webhook, huggingface_update_webhook, huggingface_enable_webhook, huggingface_disable_webhook, huggingface_delete_webhook, huggingface_get_billing_usage, huggingface_get_compute_usage, huggingface_get_storage_usage

​
Common Use Cases

Model/dataset research

Search models or datasets by task/tag and summarize the best fits for a use case

Serverless inference

Run text generation, classification, or image tasks directly without hosting your own model

Endpoint lifecycle management

Spin up a dedicated inference endpoint, scale it, and pause it when idle to control cost

Repo maintenance

Manage branches/tags, review discussions or PRs, and check security scan results before merging

​
Troubleshooting

Serverless Inference API models can be cold and need to load — retry after a short delay, or use huggingface_create_inference_endpoint for a model you call frequently to avoid cold starts entirely.
Prefer huggingface_filter_dataset_rows with a specific filter, or huggingface_list_parquet_files plus your own processing, rather than paging through an entire large dataset via rows.
Repos created under an organization namespace require the token’s account to be a member with write access to that organization — check huggingface_list_organization_members or create the repo under your personal namespace instead.