# Hugging Face Transformers

> Hugging Face Transformers is an open-source machine learning library providing thousands of pretrained models and pipelines for NLP, computer vision, audio, and multimodal tasks. This index also covers the surrounding Hugging Face APIs developers use to run inference, manage models on the Hub, deploy demos on Spaces, and serve LLMs at scale. (Generated by API Evangelist from the catalog; the provider does not publish a /llms.txt — /.well-known/openapi.md is the provider's agent-facing spec surface.)

## APIs

- [Hugging Face Transformers Library](https://huggingface.co/docs/transformers/index): Open-source Python library — pretrained models, tokenizers, and pipelines for inference and fine-tuning.
- [Hugging Face Inference API](https://huggingface.co/docs/api-inference/index): Serverless inference across thousands of Hub-hosted models (base https://api-inference.huggingface.co).
- [Hugging Face Hub API](https://huggingface.co/docs/hub/api): REST API to upload, download, and manage models, datasets, and Spaces (base https://huggingface.co/api).
- [Hugging Face Spaces API](https://huggingface.co/docs/hub/spaces-overview): Deploy and manage ML apps and demos (Gradio, Streamlit, Docker).
- [Text Generation Inference (TGI)](https://huggingface.co/docs/text-generation-inference/index): High-performance LLM inference server with OpenAI-compatible chat completions.

## Specs

- [Hub API OpenAPI 3.1](https://raw.githubusercontent.com/api-evangelist/hugging-face-transformers/refs/heads/main/openapi/hugging-face-transformers-openapi.json)
- [Provider markdown OpenAPI](https://huggingface.co/.well-known/openapi.md)
- [OpenID configuration](https://huggingface.co/.well-known/openid-configuration)

## Auth

- [Security tokens](https://huggingface.co/docs/hub/security-tokens): Bearer User Access Tokens (hf_...), fine-grained or classic.
- [OAuth / Sign in with HF](https://huggingface.co/docs/hub/oauth): OAuth 2.0 / OIDC (PKCE, device code, token exchange). Scopes: openid, profile, email, read/write/manage-repos, inference-api, jobs, webhooks, write-discussions, and more.

## Tools

- [Hugging Face MCP Server](https://huggingface.co/mcp): Official remote MCP server (stdio + streamable-http) — model/dataset/Space/paper search, docs search, jobs, repo details.
- [hf CLI](https://huggingface.co/docs/huggingface_hub/guides/cli): Agent-optimized Hub command line.
- Packages: `pip install transformers`, `pip install huggingface_hub`, `npm i @huggingface/transformers`, `@huggingface/inference`, `@huggingface/hub`.

## Docs

- [Documentation home](https://huggingface.co/docs)
- [Changelog](https://huggingface.co/changelog)
- [Status](https://status.huggingface.co)
- [Blog](https://huggingface.co/blog)