With the latest update to the @huggingface CLI you have hf cache ls/prune/rm/verify Want to delete everything in cache? `hf cache rm $(hf cache ls -q)` does the trick. Thank you @Wauplin @hanouticelina 🐐 PS: Call me the cache janitor now.

Tired of cloudflare issues and don't want to rely on public APIs? Don't forget you are always 1 command line away from deploying a private instance of any open-source LLM: `uvx hf endpoints catalog deploy --repo openai/gtp-oss-20b`

@astral_sh A big thank you to @thorwhalen1 for generously transferring the hf package name to us on PyPI. This will make the CLI much more accessible for all Hugging Face users. 🤗

Using the hf CLI just got insanely easy: uvx hf auth login uvx hf download uvx hf ... Always runs the latest version in an isolated environment. No install or updates required thanks to @astral_sh 's awesome tooling!

Who said we should stop shipping after v1.0 ?

We're definitely not slowing down - it's only been 2 weeks since v1.0, but 𝚑𝚞𝚐𝚐𝚒𝚗𝚐𝚏𝚊𝚌𝚎_𝚑𝚞𝚋 v1.1 is already here packed with a lot of nice features! 🔥 This update brings a faster and cleaner download experience and new powerful features to the Hugging Face CLI! $

Feels really great to be partnering with @wavespeed_ai on this! 🤗

Earlier this week, we released v1.0 of huggingface_hub. But to get there, it took 5 years, 35+ releases, and 280+ contributors. So we wrote a blog post to summarize the journey:

We finally stopped being scared of v1.0 😅 Hugging Face Hub just hit v1.0, a testament to how far the ecosystem (and API!) have come. Here's to the next decade of open ML 🚀 https://t.co/M9RwcqGbc9

🔥 We're thrilled to announce 𝚑𝚞𝚐𝚐𝚒𝚗𝚐𝚏𝚊𝚌𝚎_𝚑𝚞𝚋 v1.0! After five years of development, this foundational release is packed with A fully modernized HTTP backend and a complete, from-the-ground-up CLI revamp! $ pip install huggingface_hub --upgrade 🧵highly recommend

Have you noticed different performance between open model inference providers? You can now run evals across providers with the @huggingface inference providers integration in InspectAI.

Last Stop Before 1.0! We've shipped huggingface_hub v0.36.0 today with performance optimizations in HfFileSystem. Streaming from the Hub is now the go-to solution for large distributed training! And yes, you've heard it. This is the last release before the long-awaited v1.0!

huggingface_hub v0.35.0 is here! Major highlights: ⏰ Scheduled Jobs: Run GPU cron jobs on the Hub with full CLI support 🎥 Image-to-video generation powered by @FAL ⚡ New providers: welcome @Scaleway_fr & PublicAI! More details in

Awesome work from @hanouticelina and the team to use SoTA open-source LLMs with GitHub Copilot in @code without vendor lock-in! 🔥

After 5 years and 1.4B+ installs, huggingface_hub 1.0 is coming! Expect a smoother UX with httpx & typer, plus removal of legacy code while keeping migration as painless as possible. Want to be part of it? Share your comments on https://t.co/vraUTFMq0v

👋 Hello ML community! Trackio is here - the free, local-first experiment tracking library from @huggingface 🤗 ✅ Drop-in replacement for wandb ✅ Track locally, share via Spaces ✅ Zero setup, zero cost ✅ <1000 lines of Python Ready to simplify your ML experiments?

Get started at https://t.co/hGt8cTCY6D

Vibecoding is great but it’s often a pain to quickstart a project. So I made a Hugging Face Space template (FastAPI + React + SQLite + HF login), all wired up for local dev + Spaces The kicker: a PROMPT file so Claude/Cursor can take over the vibecoding of UI + API. 🧵

4/ All details and examples on

3/The new Responses API is where things get wild: ✅ Adjustable reasoning effort levels ✅ Built for agentic workflows (function calling, structured output, etc.) ✅ Remote MCP calls to external services This isn't your typical chat endpoint.