>>> curl -LsSf https://hf.co/cli/install.sh | bash

>>> powershell -ExecutionPolicy ByPass -c "irm https://hf.co/cli/install.ps1 | iex"

>>> curl -LsSf https://hf.co/cli/install.sh | bash -s -- --exclude-skill

>>> powershell -ExecutionPolicy ByPass -c "& ([scriptblock]::Create((irm https://hf.co/cli/install.ps1))) -ExcludeSkill"

>>> hf --help
Usage: hf [OPTIONS] COMMAND [ARGS]...

  Hugging Face Hub CLI

Options:
  --skills              Print the `hf-cli` SKILL.md to stdout (alias for `hf skills preview`).
  --install-completion  Install completion for the current shell.
  --show-completion     Show completion for the current shell, to copy it or customize the installation.
  -h, --help            Show this message and exit.

Main commands:
  auth                 Manage authentication (login, logout, etc.).
  buckets              Commands to interact with buckets.
  cache                Manage local cache directory.
  collections          Interact with collections on the Hub.
  datasets             Interact with datasets on the Hub.
  download             Download files from the Hub.
  endpoints            Manage Hugging Face Inference Endpoints.
  extensions           Manage hf CLI extensions.
  jobs                 Run and manage Jobs on the Hub.
  models               Interact with models on the Hub.
  papers               Interact with papers on the Hub.
  repos                Manage repos on the Hub.
  skills               Manage skills for AI assistants.
  spaces               Interact with spaces on the Hub.
  sync                 Sync files between local directory and a bucket.
  upload               Upload a file or a folder to the Hub.

Help commands:
  env      Print information about the environment.
  version  Print information about the hf version.

>>> uvx hf auth login
>>> uvx hf download
>>> uvx hf ...

>>> pip install -U "huggingface_hub"

>>> brew install hf

>>> hf update

# JSON output for scripting
>>> hf models ls --search bert --limit 2 --json | jq '.[].id'

# IDs only, one per line
>>> hf collections ls --owner nvidia -q
nvidia/nemotron-supervised-fine-tuning-69eab9824c9120a3a3b1e25e
nvidia/nvidia-nemotron-v3-69388dda16167bb1607171ea

>>> hf auth login

>>> hf auth login --force

? How would you like to log in?  [Use arrows, Enter to confirm]
> Log in with your browser
  Paste an access token

    Open this URL in your browser:
        https://huggingface.co/oauth/device

    And enter the code: ABCD-EFGH

    Waiting for authorization...
Token is valid.
The token `oauth-wauplin` has been saved to /home/wauplin/.cache/huggingface/stored_tokens
Your token has been saved to /home/wauplin/.cache/huggingface/token
Login successful.

# Or using an environment variable
>>> hf auth login --token $HF_TOKEN --add-to-git-credential
Token is valid (permission: write).
The token `token_name` has been saved to /home/wauplin/.cache/huggingface/stored_tokens
Your token has been saved in your configured git credential helpers (store).
Your token has been saved to /home/wauplin/.cache/huggingface/token
Login successful
The current active token is: `token_name`

>>> hf auth login --format agent
Ask the user to open https://huggingface.co/oauth/device in a browser and enter the code ABCD-EFGH. The code expires in 900 seconds. Waiting for authorization...
Login successful: logged in as wauplin (token saved as 'oauth-wauplin').

hf auth whoami
Wauplin
orgs:  huggingface,eu-test,OAuthTesters,hf-accelerate,HFSmolCluster

hf download --help

>>> hf download gpt2 config.json
downloading https://huggingface.co/gpt2/resolve/main/config.json to /home/wauplin/.cache/huggingface/hub/tmpwrq8dm5o
(…)ingface.co/gpt2/resolve/main/config.json: 100%|██████████████████████████████████| 665/665 [00:00<00:00, 2.49MB/s]
/home/wauplin/.cache/huggingface/hub/models--gpt2/snapshots/11c5a3d5811f50298f278a704980280950aedb10/config.json

>>> hf download HiDream-ai/HiDream-I1-Full text_encoder/model.safetensors

>>> hf download HuggingFaceH4/zephyr-7b-beta
Fetching 23 files:   0%|                                                | 0/23 [00:00<?, ?it/s]
...
...
/home/wauplin/.cache/huggingface/hub/models--HuggingFaceH4--zephyr-7b-beta/snapshots/3bac358730f8806e5c3dc7c7e19eb36e045bf720

>>> hf download gpt2 config.json model.safetensors
Fetching 2 files:   0%|                                                                        | 0/2 [00:00<?, ?it/s]
downloading https://huggingface.co/gpt2/resolve/11c5a3d5811f50298f278a704980280950aedb10/model.safetensors to /home/wauplin/.cache/huggingface/hub/tmpdachpl3o
(…)8f278a7049802950aedb10/model.safetensors: 100%|██████████████████████████████| 8.09k/8.09k [00:00<00:00, 40.5MB/s]
Fetching 2 files: 100%|████████████████████████████████████████████████████████████████| 2/2 [00:00<00:00,  3.76it/s]
/home/wauplin/.cache/huggingface/hub/models--gpt2/snapshots/11c5a3d5811f50298f278a704980280950aedb10

>>> hf download stabilityai/stable-diffusion-xl-base-1.0 --include "*.safetensors" --exclude "*.fp16.*"*
Fetching 8 files:   0%|                                                                         | 0/8 [00:00<?, ?it/s]
...
...
Fetching 8 files: 100%|█████████████████████████████████████████████████████████████████████████| 8/8 (...)
/home/wauplin/.cache/huggingface/hub/models--stabilityai--stable-diffusion-xl-base-1.0/snapshots/462165984030d82259a11f4367a4eed129e94a7b

# https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k
>>> hf download HuggingFaceH4/ultrachat_200k --repo-type dataset

# https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat
>>> hf download HuggingFaceH4/zephyr-chat --repo-type space

...

>>> hf download bigcode/the-stack --repo-type dataset --revision v1.1
...

# Equivalent to: hf download bigcode/the-stack --repo-type dataset --revision v1.1
>>> hf download hf://datasets/bigcode/the-stack@v1.1

# Download a single file from a specific revision
>>> hf download hf://datasets/HuggingFaceM4/FineVision@refs/pr/1/data/train.parquet

# Download a subfolder (note the trailing slash)
>>> hf download hf://datasets/HuggingFaceM4/FineVision/art/

# A bare id still works and defaults to a model repo
>>> hf download hf://openai-community/gpt2/config.json

>>> hf download adept/fuyu-8b model-00001-of-00002.safetensors --local-dir fuyu
...
fuyu/model-00001-of-00002.safetensors

>>> hf download openai-community/gpt2 --dry-run
[dry-run] Fetching 26 files: 100%|█████████████| 26/26 [00:04<00:00,  6.26it/s]
[dry-run] Will download 11 files (out of 26) totalling 5.6G.
File                              Bytes to download
--------------------------------- -----------------
.gitattributes                    -
64-8bits.tflite                   125.2M
64-fp16.tflite                    248.3M
64.tflite                         495.8M
README.md                         -
config.json                       -
flax_model.msgpack                497.8M
generation_config.json            -
merges.txt                        -
model.safetensors                 548.1M
onnx/config.json                  -
onnx/decoder_model.onnx           653.7M
onnx/decoder_model_merged.onnx    655.2M
onnx/decoder_with_past_model.onnx 653.7M
onnx/generation_config.json       -
onnx/merges.txt                   -
onnx/special_tokens_map.json      -
onnx/tokenizer.json               -
onnx/tokenizer_config.json        -
onnx/vocab.json                   -
pytorch_model.bin                 548.1M
rust_model.ot                     702.5M
tf_model.h5                       497.9M
tokenizer.json                    -
tokenizer_config.json             -
vocab.json                        -

>>> hf download adept/fuyu-8b --cache-dir ./path/to/cache
...
./path/to/cache/models--adept--fuyu-8b/snapshots/ddcacbcf5fdf9cc59ff01f6be6d6662624d9c745

>>> hf download gpt2 config.json --token=hf_****
/home/wauplin/.cache/huggingface/hub/models--gpt2/snapshots/11c5a3d5811f50298f278a704980280950aedb10/config.json

>>> hf download gpt2 --quiet
/home/wauplin/.cache/huggingface/hub/models--gpt2/snapshots/11c5a3d5811f50298f278a704980280950aedb10

`httpx2.TimeoutException: (TimeoutException("HTTPSConnectionPool(host='cdn-lfs-us-1.huggingface.co', port=443): Read timed out. (read timeout=10)"), '(Request ID: a33d910c-84c6-4514-8362-c705e2039d38)')`

export HF_HUB_DOWNLOAD_TIMEOUT=30

>>> hf upload --help

# Usage:  hf upload [repo_id] [local_path] [path_in_repo]

>>> hf upload my-cool-model . .
https://huggingface.co/Wauplin/my-cool-model/tree/main/

>>> hf upload my-cool-model ./models .
https://huggingface.co/Wauplin/my-cool-model/tree/main/

>>> hf upload my-cool-model ./path/to/curated/data /data/train
https://huggingface.co/Wauplin/my-cool-model/tree/main/data/train

>>> hf upload Wauplin/my-cool-model ./models/model.safetensors
https://huggingface.co/Wauplin/my-cool-model/blob/main/model.safetensors

>>> hf upload Wauplin/my-cool-model ./models/model.safetensors /vae/model.safetensors
https://huggingface.co/Wauplin/my-cool-model/blob/main/vae/model.safetensors

# Sync local Space with Hub (upload new files except from logs/, delete removed files)
>>> hf upload Wauplin/space-example --repo-type=space --exclude="/logs/*" --delete="*" --commit-message="Sync local Space with Hub"
...

>>> hf upload Wauplin/my-cool-dataset ./data /train --repo-type=dataset
...

>>> hf upload MyCoolOrganization/my-cool-model . .
https://huggingface.co/MyCoolOrganization/my-cool-model/tree/main/

# Upload files to a PR
>>> hf upload bigcode/the-stack . . --repo-type dataset --revision refs/pr/104
...

# Equivalent to: hf upload Wauplin/my-cool-dataset ./train.csv data/train.csv --repo-type dataset --revision my-branch
>>> hf upload hf://datasets/Wauplin/my-cool-dataset@my-branch/data/train.csv ./train.csv

# Upload a whole folder to the root of a model repo
>>> hf upload hf://Wauplin/my-cool-model ./models

# Create a PR and upload the files to it
>>> hf upload bigcode/the-stack . . --repo-type dataset --revision refs/pr/104
https://huggingface.co/datasets/bigcode/the-stack/blob/refs%2Fpr%2F104/

# Upload new logs every 10 minutes
hf upload training-model logs/ --every=10

>>> hf upload Wauplin/my-cool-model ./models . --commit-message="Epoch 34/50" --commit-description="Val accuracy: 68%. Check tensorboard for more details."
...
https://huggingface.co/Wauplin/my-cool-model/tree/main

>>> hf upload Wauplin/my-cool-model ./models . --token=hf_****
...
https://huggingface.co/Wauplin/my-cool-model/tree/main

>>> hf upload Wauplin/my-cool-model ./models . --quiet
https://huggingface.co/Wauplin/my-cool-model/tree/main

>>> hf buckets create my-bucket

>>> hf buckets create my-bucket --private

>>> hf buckets list
ID                   PRIVATE       SIZE TOTAL_FILES CREATED_AT
-------------------- ------- ---------- ----------- ----------
username/my-bucket                   32           5 2026-02-16
username/checkpoints         117609095         700 2026-02-13
username/logs                321757477        2000 2026-02-13

# Human-readable sizes
>>> hf buckets list -h
ID                   PRIVATE     SIZE TOTAL_FILES CREATED_AT
-------------------- ------- -------- ----------- ----------
username/my-bucket               32 B           5 2026-02-16
username/checkpoints         117.6 MB         700 2026-02-13
username/logs                321.8 MB        2000 2026-02-13

# List buckets in a specific namespace
>>> hf buckets ls my-org

# Filter buckets by name
>>> hf buckets list --search "checkpoint"

>>> hf buckets info username/my-bucket
{
  "id": "username/my-bucket",
  "private": false,
  "created_at": "2026-02-16T15:28:32+00:00",
  "size": 32,
  "total_files": 5
}

# Make a bucket private
>>> hf buckets settings username/my-bucket --private

# Make it public again
>>> hf buckets settings username/my-bucket --public

>>> hf buckets delete username/my-bucket --yes

>>> hf buckets rm username/my-bucket/old-model.bin

>>> hf buckets rm username/my-bucket/logs/ --recursive

>>> hf buckets rm username/my-bucket --recursive --include "*.tmp"

>>> hf buckets rm username/my-bucket/checkpoints/ --recursive --dry-run

>>> hf buckets list username/my-bucket
        2048  2026-01-15 10:30:00  big.bin
           5  2026-01-15 10:30:00  file.txt
              2026-01-15 10:30:00  sub/

# Recursive with human-readable sizes
>>> hf buckets list username/my-bucket -R -h
      2.0 KB         Jan 15 10:30  big.bin
         5 B         Jan 15 10:30  file.txt
        14 B         Jan 15 10:30  sub/nested.txt
         4 B         Jan 15 10:30  sub/deep/file.txt

# Tree with human-readable sizes
>>> hf buckets list username/my-bucket --tree -h -R
2.0 KB  Jan 15 10:30  ├── big.bin
   5 B  Jan 15 10:30  ├── file.txt
                      └── sub/
                          ├── deep/
   4 B  Jan 15 10:30  │       └── file.txt
  14 B  Jan 15 10:30  └── nested.txt

# Clean tree without metadata
>>> hf buckets list username/my-bucket --tree --quiet -R
├── big.bin
├── file.txt
└── sub/
    ├── deep/
    │   └── file.txt
    └── nested.txt

>>> hf buckets list username/my-bucket/sub -R

# To a repository
>>> hf cp ./model.safetensors hf://username/my-model/model.safetensors

# To a bucket (uses the local filename when the destination ends with /)
>>> hf cp ./data.csv hf://buckets/username/my-bucket/logs/

# From a repository
>>> hf cp hf://datasets/username/my-dataset@refs/pr/1/data.csv ./data.csv

# From a bucket, to the current directory (destination omitted)
>>> hf cp hf://buckets/username/my-bucket/config.json

# Download to stdout
>>> hf cp hf://buckets/username/my-bucket/config.json - | jq .

# Upload from stdin
>>> echo "hello" | hf cp - hf://username/my-model/hello.txt

# Repo to repo
>>> hf cp hf://username/source-model/config.json hf://username/dest-model/config.json

# Repo to bucket
>>> hf cp hf://datasets/username/my-dataset/data/train/ hf://buckets/username/my-bucket/datasets/train/

# Bucket to bucket
>>> hf cp hf://buckets/username/source-bucket/logs/ hf://buckets/username/archive-bucket/logs/

# Without trailing slash: "logs" dir is nested => archive/logs/...
>>> hf cp hf://buckets/username/my-bucket/logs hf://buckets/username/archive-bucket/

# With trailing slash: only contents of "logs" are copied => archive/...
>>> hf cp hf://buckets/username/my-bucket/logs/ hf://buckets/username/archive-bucket/

>>> hf buckets sync ./data hf://buckets/username/my-bucket

>>> hf buckets sync hf://buckets/username/my-bucket ./data

>>> hf buckets sync ./data hf://buckets/username/my-bucket --delete

>>> hf buckets sync ./data hf://buckets/username/my-bucket --include "*.safetensors" --exclude "*.tmp"

>>> hf buckets sync ./data hf://buckets/username/my-bucket --existing
>>> hf buckets sync ./data hf://buckets/username/my-bucket --ignore-existing

# Generate a plan
>>> hf buckets sync ./data hf://buckets/username/my-bucket --plan sync-plan.jsonl

# Review and apply the plan
>>> hf buckets sync --apply sync-plan.jsonl

>>> hf buckets sync ./data hf://buckets/username/my-bucket --dry-run | jq .

# List trending models
>>> hf models ls

# Search for models
>>> hf models ls --search "lora"

# Filter by author
>>> hf models ls --author Qwen

# Filter by pipeline tag
>>> hf models ls --pipeline-tag summarization

# Filter by parameter count
>>> hf models ls --num-parameters min:6B,max:128B

# Only non-gated models
>>> hf models ls --no-gated --author google

# Only models runnable by a given app
>>> hf models ls --apps llama.cpp

# Sort by downloads
>>> hf models ls --sort downloads --limit 10

# List files in a model repo
>>> hf models ls meta-llama/Llama-3.2-1B-Instruct

# List files recursively
>>> hf models ls meta-llama/Llama-3.2-1B-Instruct -R

# Tree view with human-readable sizes
>>> hf models ls meta-llama/Llama-3.2-1B-Instruct --tree -h

# List files at a specific revision
>>> hf models ls meta-llama/Llama-3.2-1B-Instruct --revision main

>>> hf models info Lightricks/LTX-2

# Full card (metadata + text)
>>> hf models card google/gemma-4-31B-it

# Just the metadata (from the YAML frontmatter)
>>> hf models card google/gemma-4-31B-it --metadata

# Metadata as JSON (useful for scripting and agents)
>>> hf models card google/gemma-4-31B-it --metadata --format json

# Just the text body (no YAML frontmatter)
>>> hf models card google/gemma-4-31B-it --text

# List trending datasets
>>> hf datasets ls

# Search for datasets
>>> hf datasets ls --search "code"

# List official benchmark datasets
>>> hf datasets ls --filter benchmark:official

# Sort by downloads
>>> hf datasets ls --sort downloads --limit 10

# List files in a dataset repo
>>> hf datasets ls HuggingFaceFW/fineweb

# List files recursively with human-readable sizes
>>> hf datasets ls HuggingFaceFW/fineweb -R -h

# Tree view
>>> hf datasets ls HuggingFaceFW/fineweb --tree

>>> hf datasets leaderboard SWE-bench/SWE-bench_Verified
>>> hf datasets leaderboard SWE-bench/SWE-bench_Verified --limit 5 --format json

>>> hf datasets info HuggingFaceFW/fineweb

# Full card (metadata + text)
>>> hf datasets card HuggingFaceFW/fineweb

# Just the metadata (from the YAML frontmatter)
>>> hf datasets card HuggingFaceFW/fineweb --metadata

# Metadata as JSON (useful for scripting and agents)
>>> hf datasets card HuggingFaceFW/fineweb --metadata --format json

# Just the text body (no YAML frontmatter)
>>> hf datasets card HuggingFaceFW/fineweb --text

>>> hf datasets parquet cfahlgren1/hub-stats
>>> hf datasets parquet cfahlgren1/hub-stats --subset models
>>> hf datasets parquet cfahlgren1/hub-stats --split train
>>> hf datasets parquet cfahlgren1/hub-stats --format json

>>> hf datasets sql "SELECT COUNT(*) AS rows FROM read_parquet('https://huggingface.co/api/datasets/cfahlgren1/hub-stats/parquet/models/train/0.parquet')"
>>> hf datasets sql "SELECT * FROM read_parquet('https://huggingface.co/api/datasets/cfahlgren1/hub-stats/parquet/models/train/0.parquet') LIMIT 5" --format json

# Python package
>>> pip install duckdb

# or standalone DuckDB CLI (Homebrew: macOS/Linux)
>>> brew install duckdb

# List trending Spaces
>>> hf spaces ls

# Search for Spaces
>>> hf spaces ls --search "3d"

# Sort by likes
>>> hf spaces ls --sort likes --limit 10

# List files in a Space repo
>>> hf spaces ls victor/deepsite

# List files recursively with tree view
>>> hf spaces ls victor/deepsite --tree -R -h

>>> hf spaces info victor/deepsite

# Full card (metadata + text)
>>> hf spaces card mteb/leaderboard

# Just the card metadata (from the YAML frontmatter)
>>> hf spaces card mteb/leaderboard --metadata

# Card metadata as JSON
>>> hf spaces card mteb/leaderboard --metadata --format json

# Just the text body (no YAML frontmatter)
>>> hf spaces card mteb/leaderboard --text

>>> hf spaces pause username/my-space

>>> hf spaces restart username/my-space
>>> hf spaces restart username/my-space --factory-reboot

>>> hf spaces wait username/my-space

# With a timeout
>>> hf spaces wait username/my-space --timeout 5m

# Chain with restart
>>> hf spaces restart username/my-space && hf spaces wait username/my-space

>>> hf spaces hardware

>>> hf spaces zero-gpu quota
✓ ZeroGPU quota (in GPU-seconds)
  remaining: 1810
  base: 2400
  resets_at: 2026-09-30T18:12:03+00:00
  overquota_used: 0

>>> hf spaces zero-gpu quota --format json
{"remaining": 1810, "base": 2400, "resets_at": "2026-09-30T18:12:03+00:00", "overquota_used": 0}

>>> hf spaces settings username/my-space --sleep-time 3600
>>> hf spaces settings username/my-space --hardware t4-medium

>>> hf spaces secrets ls username/my-space
>>> hf spaces secrets add username/my-space -s OPENAI_API_KEY=sk-...
>>> hf spaces secrets add username/my-space --secrets-file .env.secrets
>>> hf spaces secrets delete username/my-space OPENAI_API_KEY --yes

>>> hf spaces variables ls username/my-space
>>> hf spaces variables add username/my-space -e MODEL_ID=gpt2 -e MAX_TOKENS=512
>>> hf spaces variables add username/my-space --env-file .env
>>> hf spaces variables delete username/my-space MAX_TOKENS --yes

# SSH into a Space
>>> hf spaces ssh username/my-space

# Print the SSH command without running it
>>> hf spaces ssh username/my-space --dry-run

# Auto-enable Dev Mode if disabled
>>> hf spaces ssh username/my-space --auto

# Use a specific SSH key
>>> hf spaces ssh username/my-space -i ~/.ssh/id_ed25519

# List most recent daily papers
>>> hf papers ls

# List trending papers
>>> hf papers ls --sort=trending

# List papers from a specific date
>>> hf papers ls --date=2025-01-23

# List today's papers
>>> hf papers ls --date=today

# List papers from a specific week
>>> hf papers ls --week=2025-W09

# List papers from a specific month
>>> hf papers ls --month=2025-02

# List papers submitted by a specific user
>>> hf papers ls --submitter=akhaliq

# Limit results
>>> hf papers ls --sort=trending --limit=5

# Search papers by keyword
>>> hf papers search "vision language"

# Limit search results
>>> hf papers search "diffusion models" --limit=10

# Output as JSON
>>> hf papers search "attention" --format=json

# Get structured metadata for a paper (returns JSON)
>>> hf papers info 2601.15621

# Read the full paper content as markdown
>>> hf papers read 2601.15621

>>> hf discussions list username/my-model

>>> hf discussions list username/my-model --kind pull_request --status merged
>>> hf discussions list username/my-model --author alice

>>> hf discussions list username/my-model --format json
>>> hf discussions ls username/my-model --quiet

>>> hf discussions info username/my-model 5

>>> hf discussions create username/my-model --title "Bug report"
>>> hf discussions create username/my-model --title "Feature request" --body "Please add X"
>>> hf discussions create username/my-model --title "Report" --body-file report.md

>>> hf discussions create username/my-model --title "Fix typo" --pull-request

>>> hf discussions comment username/my-model 5 --body "Thanks for reporting!"
>>> hf discussions comment username/my-model 5 --body-file review.md
>>> echo "LGTM" | hf discussions comment username/my-model 5 --body-file -

>>> hf discussions edit username/my-model 5 abc123 --body "Updated comment."
>>> hf discussions edit username/my-model 5 abc123 --body-file fixed.md

>>> hf discussions create username/my-model --title "Fix rendering" --pr --body "Before: ![Before the fix](./before.png)" --attach ./before.png --attach "./after.png#After the fix"
>>> hf discussions comment username/my-model 5 --body "Here is a demo:" --attach demo.mp4

>>> hf discussions close username/my-model 5
>>> hf discussions close username/my-model 5 --yes --comment "Resolved"

>>> hf discussions reopen username/my-model 5 --yes

>>> hf discussions merge username/my-model 5 --yes

>>> hf discussions rename username/my-model 5 "Updated title"

>>> hf discussions diff username/my-model 5

# List all your repos (first 30)
>>> hf repos ls
REPOSITORY                TYPE     UPDATED      VISIBILITY   STORAGE  % OF TOTAL
------------------------  -------  ----------   ----------  --------  ----------
username/bucket-raw       bucket   2026-04-29   public        1.7 TB       72.3%
username/my-model         model    2026-05-06   public        4.8 GB       18.1%
username/my-dataset       dataset  2024-09-14   private     598.4 MB        5.2%
Hint: Showing 30 of 42 repos. Use `--limit 0` to list all.

# List only models
>>> hf repos ls --type model

# Search by name
>>> hf repos ls --search "bert"

# List repos from an organization
>>> hf repos ls --namespace my-org

# Combine filters
>>> hf repos ls --namespace my-org --type dataset --search "train"

# List all repos (no limit)
>>> hf repos ls --limit 0

>>> hf repos ls --limit 0 --format json | jq '.[].id'
>>> hf repos ls -q

>>> hf repos create Wauplin/my-cool-model
Successfully created Wauplin/my-cool-model on the Hub.
Your repo is now available at https://huggingface.co/Wauplin/my-cool-model

>>> hf repos create my-cool-dataset --repo-type dataset --private
>>> hf repos create my-gradio-space --repo-type space --sdk gradio

>>> hf repos create my-model --region us

>>> hf repos duplicate openai/gdpval --type dataset
>>> hf repos duplicate openai/gdpval myorg/my-gdpval --type dataset --private

>>> hf repos delete Wauplin/my-cool-model

>>> hf repos delete my-cool-dataset --repo-type dataset
>>> hf repos delete my-gradio-space --repo-type space

>>> hf repos move old-namespace/my-model new-namespace/my-model

>>> hf repos settings Wauplin/my-cool-model --gated auto
>>> hf repos settings Wauplin/my-cool-model --private true
>>> hf repos settings Wauplin/my-cool-model --private false

>>> hf repos delete-files Wauplin/my-cool-model folder/
Files correctly deleted from repo. Commit: https://huggingface.co/Wauplin/my-cool-mo...

>>> hf repos delete-files Wauplin/my-cool-model file.txt folder/pytorch_model.bin
Files correctly deleted from repo. Commit: https://huggingface.co/Wauplin/my-cool-mo...

>>> hf repos delete-files Wauplin/my-cool-model "*.txt" "folder/*.bin"
Files correctly deleted from repo. Commit: https://huggingface.co/Wauplin/my-cool-mo...

>>> hf repos delete-files --token=hf_**** Wauplin/my-cool-model file.txt

# Create a branch
>>> hf repos branch create Wauplin/my-cool-model dev

# Create a branch from a specific revision
>>> hf repos branch create Wauplin/my-cool-model release-1 --revision refs/pr/104

# Delete a branch
>>> hf repos branch delete Wauplin/my-cool-model dev

# List cached repositories
>>> hf cache ls

# List cached revisions
>>> hf cache ls --revisions

# Remove specific items from cache
>>> hf cache rm model/gpt2

# Remove unreferenced revisions
>>> hf cache prune

# Verify cached file checksums
>>> hf cache verify gpt2

>>> hf cache ls
ID                          SIZE     LAST_ACCESSED LAST_MODIFIED REFS
--------------------------- -------- ------------- ------------- -----------
dataset/nyu-mll/glue          157.4M 2 days ago    2 days ago    main script
model/LiquidAI/LFM2-VL-1.6B     3.2G 4 days ago    4 days ago    main
model/microsoft/UserLM-8b      32.1G 4 days ago    4 days ago    main

Found 3 repo(s) for a total of 5 revision(s) and 35.5G on disk.

>>> hf cache ls --filter "size>30g" --revisions
ID                        REVISION                                 SIZE     LAST_MODIFIED REFS
------------------------- ---------------------------------------- -------- ------------- ----
model/microsoft/UserLM-8b be8f2069189bdf443e554c24e488ff3ff6952691    32.1G 4 days ago    main

Found 1 repo(s) for a total of 1 revision(s) and 32.1G on disk.

>>> hf cache rm $(hf cache ls --filter "accessed>1y" -q) -y
About to delete 2 repo(s) totalling 5.31G.
  - model/meta-llama/Llama-3.2-1B-Instruct (entire repo)
  - model/hexgrad/Kokoro-82M (entire repo)
Deleted 2 repo(s) and 2 revision(s); freed 5.31G.

>>> hf cache rm model/LiquidAI/LFM2-VL-1.6B
About to delete 1 repo(s) totalling 3.2G.
  - model/LiquidAI/LFM2-VL-1.6B (entire repo)
Proceed with deletion? [y/N]: y
Deleted 1 repo(s) and 2 revision(s); freed 3.2G.

>>> hf cache rm hf://models/openai-community/gpt2 --dry-run
About to delete 1 repo(s) totalling 1.1G.
  - model/openai-community/gpt2 (entire repo)
Dry run: no files were deleted.

>>> hf cache rm hf://models/unsloth/gemma-3-27b-it-GGUF/gemma-3-27b-it-Q4_K_M.gguf --dry-run
About to delete 1 file(s) totalling 16.5G.
  - model/unsloth/gemma-3-27b-it-GGUF@3f4b5c1d2e6a7b8c9d0e1f2a3b4c5d6e7f8a9b0c/gemma-3-27b-it-Q4_K_M.gguf
Dry run: no files were deleted.

>>> hf cache rm model/t5-small 8f3ad1c --dry-run
About to delete 1 repo(s) and 1 revision(s) totalling 1.1G.
  - model/t5-small:
      8f3ad1c [main] 1.1G
Dry run: no files were deleted.

>>> hf cache prune
About to delete 3 unreferenced revision(s) and 2 incomplete download(s) (2.4G total).
  - model/t5-small:
      1c610f6b [refs/pr/1] 820.1M
      d4ec9b72 [(detached)] 640.5M
  - dataset/google/fleurs:
      2b91c8dd [(detached)] 937.6M
Proceed? [y/N]: y
Deleted 3 unreferenced revision(s) and 2 incomplete download(s); freed 2.4G.

# Verify main revision of a model in cache
>>> hf cache verify deepseek-ai/DeepSeek-OCR

# Verify a specific revision
>>> hf cache verify deepseek-ai/DeepSeek-OCR --revision refs/pr/5
>>> hf cache verify deepseek-ai/DeepSeek-OCR --revision ef93bf4a377c5d5ed9dca78e0bc4ea50b26fe6a4

# Verify a private repo
>>> hf cache verify me/private-model --token hf_***

# Verify a dataset
>>> hf cache verify karpathy/fineweb-edu-100b-shuffle --repo-type dataset

# Verify files in a local directory
>>> hf cache verify deepseek-ai/DeepSeek-OCR --local-dir /path/to/repo

>>> hf cache verify deepseek-ai/DeepSeek-OCR --fail-on-missing-files --fail-on-extra-files

✅ Verified 13 file(s) for 'deepseek-ai/DeepSeek-OCR' (model) in ~/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B-Instruct/snapshots/9213176726f574b556790deb65791e0c5aa438b6
  All checksums match.

# Create a tag
>>> hf repos tag create my-model v1.0

# List tags
>>> hf repos tag list my-model

# Delete a tag
>>> hf repos tag delete my-model v1.0

>>> hf repos tag create Wauplin/my-cool-model v1.0
You are about to create tag v1.0 on model Wauplin/my-cool-model
Tag v1.0 created on Wauplin/my-cool-model

>>> hf repos tag create Wauplin/my-cool-model v1.0 --revision refs/pr/104
You are about to create tag v1.0 on model Wauplin/my-cool-model
Tag v1.0 created on Wauplin/my-cool-model

>>> hf repos tag create bigcode/the-stack v1.0 --repo-type dataset
You are about to create tag v1.0 on dataset bigcode/the-stack
Tag v1.0 created on bigcode/the-stack

>>> hf repos tag list Wauplin/gradio-space-ci --repo-type space
Tags for space Wauplin/gradio-space-ci:
0.2.2
0.2.1
0.2.0
0.1.2
0.0.2
0.0.1

>>> hf repos tag delete Wauplin/my-cool-model v1.0
You are about to delete tag v1.0 on model Wauplin/my-cool-model
Proceed? [Y/n] y
Tag v1.0 deleted on Wauplin/my-cool-model

>>> hf env

Copy-and-paste the text below in your GitHub issue.

- huggingface_hub version: 1.0.0.rc6
- Platform: Linux-6.8.0-85-generic-x86_64-with-glibc2.35
- Python version: 3.11.14
- Running in iPython ?: No
- Running in notebook ?: No
- Running in Google Colab ?: No
- Running in Google Colab Enterprise ?: No
- Token path ?: /home/wauplin/.cache/huggingface/token
- Has saved token ?: True
- Who am I ?: Wauplin
- Configured git credential helpers: store
- Installation method: unknown
- Torch: N/A
- httpx2: 2.0.0
- hf_xet: 1.1.10
- gradio: 5.41.1
- tensorboard: N/A
- pydantic: 2.11.7
- ENDPOINT: https://huggingface.co
- HF_HUB_CACHE: /home/wauplin/.cache/huggingface/hub
- HF_ASSETS_CACHE: /home/wauplin/.cache/huggingface/assets
- HF_TOKEN_PATH: /home/wauplin/.cache/huggingface/token
- HF_STORED_TOKENS_PATH: /home/wauplin/.cache/huggingface/stored_tokens
- HF_HUB_OFFLINE: False
- HF_HUB_DISABLE_TELEMETRY: False
- HF_HUB_DISABLE_PROGRESS_BARS: None
- HF_HUB_DISABLE_SYMLINKS_WARNING: False
- HF_HUB_DISABLE_EXPERIMENTAL_WARNING: False
- HF_HUB_DISABLE_IMPLICIT_TOKEN: False
- HF_HUB_DISABLE_XET: False
- HF_HUB_DISABLE_SHARED_BLOBS: False
- HF_HUB_ETAG_TIMEOUT: 10
- HF_HUB_DOWNLOAD_TIMEOUT: 10

# Directly run Python code
>>> hf jobs run --name hello-world python:3.12 python -c 'print("Hello from the cloud!")'

# Use GPUs without any setup
>>> hf jobs run --flavor a10g-small pytorch/pytorch:2.6.0-cuda12.4-cudnn9-devel \
... python -c "import torch; print(torch.cuda.get_device_name())"

# Run in an organization account
>>> hf jobs run --namespace my-org-name python:3.12 python -c 'print("Running in an org account")'

# Run from Hugging Face Spaces
>>> hf jobs run hf.co/spaces/lhoestq/duckdb duckdb -c 'select "hello world"'

# Run a Python script with `uv` (experimental)
>>> hf jobs uv run --name my-script my_script.py

# Run a simple Python script
>>> hf jobs run python:3.12 python -c 'print("Hello from HF compute!")'

>>> JOB_ID=$(hf jobs run -dq python:3.12 python train.py)
>>> hf jobs wait "$JOB_ID"

# List your running jobs
>>> hf jobs ls
# List all jobs
>>> hf jobs ls -a

# Inspect the status of a job
>>> hf jobs inspect <job_id>

# View logs from a job
>>> hf jobs logs <job_id>

# View resources usage stats and metrics of running jobs
>>> hf jobs stats
# View resources usage stats and metrics of some jobs
>>> hf jobs stats [job_ids]...
# Follow live stats until the jobs complete
>>> hf jobs stats -f [job_ids]...

# Cancel a job
>>> hf jobs cancel <job_id>

# Wait until one or more jobs finish (exit code 0 only if all jobs completed successfully)
>>> hf jobs wait <job_id> [<job_id>...]

# Wait for all currently running jobs
>>> hf jobs ls -q | xargs hf jobs wait

>>> hf jobs run python:3.12 python train.py && echo "training succeeded"

# Use an A10G GPU to check PyTorch CUDA
>>> hf jobs run --flavor a10g-small pytorch/pytorch:2.6.0-cuda12.4-cudnn9-devel \
... python -c 'import torch; print(f"This code ran with the following GPU: {torch.cuda.get_device_name()}")'

This code ran with the following GPU: NVIDIA A10G

>>> hf jobs run --flavor cpu-basic python:3.12 -- python --help

# Pass environment variables
>>> hf jobs run -e FOO=foo -e BAR=bar python:3.12 python -c 'import os; print(os.environ["FOO"], os.environ["BAR"])'

# Pass an environment from a local .env file
>>> hf jobs run --env-file .env python:3.12 python -c 'import os; print(os.environ["FOO"], os.environ["BAR"])'

# Pass secrets by name - the value is read from your environment, never from the command line
>>> hf jobs run -s MY_SECRET python:3.12 python -c 'import os; print(os.environ["MY_SECRET"])'

# Pass secrets from a local .env.secrets file - they will be encrypted server side
>>> hf jobs run --secrets-file .env.secrets python:3.12 python -c 'import os; print(os.environ["MY_SECRET"])'

# Or pipe them in, so the values touch neither the command line nor the disk
>>> printf 'MY_SECRET=psswrd\n' | hf jobs run --secrets-file - python:3.12 python -c 'import os; print(os.environ["MY_SECRET"])'

# Access job environment information
>>> hf jobs run python:3.12 python -c "import os; print(f'Job: {os.environ.get(\"JOB_ID\")}, CPU: {os.environ.get(\"CPU_CORES\")}, Mem: {os.environ.get(\"MEMORY\")}')"

# Set timeout in seconds (default unit)
>>> hf jobs run --timeout 7200 python:3.12 python train.py

# Use time units: s (seconds), m (minutes), h (hours), d (days)
>>> hf jobs run --timeout 2h pytorch/pytorch:2.6.0-cuda12.4-cudnn9-devel python train.py
>>> hf jobs run --timeout 90m python:3.12 python process_data.py
>>> hf jobs run --timeout 1.5h python:3.12 python train.py  # floats are supported

# UV script with timeout
>>> hf jobs uv run --timeout 2h training_script.py

# Scheduled job with timeout
>>> hf jobs scheduled run @daily --timeout 4h python:3.12 python daily_task.py

>>> hf jobs run --attempts 3 --detach python:3.12 python train.py
>>> hf jobs rerun <job_id>

# Docker Job with a mounted volume as input
>>> hf jobs run -v hf://datasets/HuggingFaceFW/fineweb:/dataset \
...     duckdb/duckdb duckdb -c "SELECT * FROM '/dataset/**/*.parquet' LIMIT 5"

# UV Job with a mounted volume to save checkpoints when training a model
>>> hf jobs uv run -v hf://buckets/username/my-bucket:/training-outputs \
...     sft.py --output-dir /training-outputs/training-v3-final ...

>>> hf jobs uv run -v ./my-data:/data process.py

>>> hf jobs uv run -v ./pdfs:/input -v ./md-out:/output:rw ocr.py
...
Hint: Volume '/output' is mounted read-write. Once the job is over, pull back its data with:
  hf buckets sync hf://buckets/username/jobs-artifacts/md-out-a1b2c3d4 ./md-out

>>> hf jobs run -l my-label --label foo=bar ubuntu echo "This Job has multiple labels"

>>> hf jobs run --name training-v2 python:3.12 python train.py
>>> hf jobs labels <job_id> --name training-v2

# Show completed Jobs
>>> hf jobs ls -a --status completed

# Show running or scheduling Jobs
>>> hf jobs ls --status running,scheduling

# Show Jobs named `training-v2`
>>> hf jobs ls -a --name training-v2

# Show Jobs with the `model=Qwen3-06B` label
>>> hf jobs ls -a --label model=Qwen3-06B

# Combine filters: running Jobs labelled both `env=prod` and `team=ml`
>>> hf jobs ls --status running --label env=prod --label team=ml

# Show up to 500 Jobs
>>> hf jobs ps -a --limit 500

# Show all Jobs (no limit)
>>> hf jobs ps -a --limit 0

# Start a Job with a public HTTP server
>>> hf jobs run --detach --expose-public 8000 python:3.12 python -m http.server 8000

# Replace the configuration on a running Job: port 8000 is private, 9000 is public
>>> hf jobs expose <job_id> 8000 --public 9000

# Close all exposed ports
>>> hf jobs expose <job_id> --clear

# Start a job with SSH enabled
>>> hf jobs run --ssh --detach python:3.12 sleep infinity

# Open an SSH session into it
>>> hf jobs ssh <job_id>

# Print the SSH command instead of running it
>>> hf jobs ssh <job_id> --dry-run

# Use a specific identity file
>>> hf jobs ssh <job_id> -i ~/.ssh/id_ed25519

# Start a server, reachable by the other members of the group as "master"
>>> hf jobs run --detach --network-group train --network-alias master python:3.12 python -m http.server 8000

# Start a client in the same group
>>> hf jobs run --detach --network-group train python:3.12 sh -c 'curl --retry 10 --retry-connrefused "http://${HF_NETWORK_GROUP_PREFIX}master:8000/"'

# Run a UV script (creates temporary repo)
>>> hf jobs uv run my_script.py

# Run with persistent repo
>>> hf jobs uv run --repo my-uv-scripts my_script.py

# Run with GPU
>>> hf jobs uv run --flavor t4-small ml_training.py

# Pass arguments to script
>>> hf jobs uv run process.py input.csv output.parquet

# Add dependencies
>>> hf jobs uv run --with transformers --with torch train.py

# Run a script directly from a URL
>>> hf jobs uv run https://huggingface.co/datasets/username/scripts/resolve/main/example.py

# Run a command
>>> hf jobs uv run --with lighteval python -c 'import lighteval'

>>> hf jobs uv run --flavor t4-small train.py -- --help

# /// script
# requires-python = ">=3.11"
# dependencies = ["vllm", "datasets"]
#
# [tool.hf-jobs]
# name    = "unlimited-ocr"
# image   = "vllm/vllm-openai:unlimited-ocr"
# flavor  = "l4x1"
# python  = "/usr/bin/python3"
# timeout = "2h"
# env     = { PYTHONPATH = "/usr/local/lib/python3.12/dist-packages", BATCH_SIZE = "64" }
# secrets = ["HF_TOKEN"]
# labels  = { task = "ocr" }
# volumes = ["hf://datasets/org/pdfs:/input"]
# ///

# Same script, on bigger hardware, with one extra env var
>>> hf jobs uv run --flavor a10g-large -e BATCH_SIZE=64 ocr.py in_ds out_ds

>>> hf jobs uv run --dry-run --flavor a10g-large ocr.py in_ds out_ds
Warning: The script's [tool.hf-jobs] table requests HF_TOKEN: the value(s) from your local environment would be sent to the Job.
Job configuration:
  script   ocr.py
  args     in_ds out_ds
  image    vllm/vllm-openai:unlimited-ocr (from script)
  flavor   a10g-large
  python   /usr/bin/python3 (from script)
  timeout  2h (from script)
  env      PYTHONPATH=/usr/local/lib/python3.12/dist-packages (from script)
           BATCH_SIZE=64 (from script)
  secrets  HF_TOKEN=*** (from script)
  volumes  hf://datasets/org/pdfs:/input (from script)
  labels   task=ocr (from script)
           name=unlimited-ocr (from script)
(dry run) Job not submitted.

# Schedule a job that runs every hour
>>> hf jobs scheduled run @hourly --name hourly-task python:3.12 python -c 'print("This runs every hour!")'

# Use the CRON syntax
>>> hf jobs scheduled run "*/5 * * * *" python:3.12 python -c 'print("This runs every 5 minutes!")'

# Schedule with GPU
>>> hf jobs scheduled run @hourly --flavor a10g-small pytorch/pytorch:2.6.0-cuda12.4-cudnn9-devel \
... python -c 'import torch; print(f"This code ran with the following GPU: {torch.cuda.get_device_name()}")'

# Schedule a UV script
>>> hf jobs scheduled uv run @hourly --name hourly-script my_script.py

# List your active scheduled jobs
>>> hf jobs scheduled ls

# Same filters as `hf jobs ls`: --status, --label and --name
>>> hf jobs scheduled ls --all
>>> hf jobs scheduled ls --status suspended
>>> hf jobs scheduled ls --name hourly-task --label env=prod

# Inspect the status of a job
>>> hf jobs scheduled inspect <scheduled_job_id>

# Suspend (pause) a scheduled job
>>> hf jobs scheduled suspend <scheduled_job_id>

# Resume a scheduled job
>>> hf jobs scheduled resume <scheduled_job_id>

# Change future run times (Mondays at 9:00)
>>> hf jobs scheduled reschedule <scheduled_job_id> "0 9 * * 1"

# Trigger a scheduled job to run right now (does not change the schedule)
>>> hf jobs scheduled trigger <scheduled_job_id>

# Delete a scheduled job
>>> hf jobs scheduled delete <scheduled_job_id>

# Create a sandbox (waits until it is ready, prints its id)
>>> hf sandbox create
✓ Sandbox ready id=687f911eaea852de79c4a50a image=python:3.12 elapsed=6.0s

# Attach labels to the underlying Job
>>> hf sandbox create --label controller-run=run-42 --label team=data-infra

# Run commands inside it (output is streamed, exit code is propagated)
>>> hf sandbox exec 687f911eaea852de79c4a50a -- python -c "print('hi')"
hi

# Copy files in and out (docker-style)
>>> hf sandbox cp data.csv 687f911eaea852de79c4a50a:/data/data.csv
>>> hf sandbox cp 687f911eaea852de79c4a50a:/app/results.json results.json

# Terminate a sandbox
>>> hf sandbox kill 687f911eaea852de79c4a50a

# Plain environment variables
>>> hf sandbox create -e LOG_LEVEL=debug --env-file .env

# Secrets, read from your environment by name (nothing sensitive on the command line)
>>> hf sandbox create -s OPENAI_API_KEY -s HF_TOKEN

# Secrets from a file, or piped in on stdin
>>> hf sandbox create --secrets-file .env.secrets
>>> printf 'OPENAI_API_KEY=sk-...\n' | hf sandbox create --secrets-file -

>>> hf webhooks ls
ID           URL                                DISABLED  DOMAINS  WATCHED
wh-abc123    https://example.com/hook            False     repo     model:bert-base-uncased
wh-def456    https://example.com/other-hook      False     repo     org:HuggingFace

>>> hf webhooks ls --format json
>>> hf webhooks ls -q

>>> hf webhooks info wh-abc123

>>> hf webhooks create --url https://example.com/hook --watch model:bert-base-uncased
>>> hf webhooks create --url https://example.com/hook --watch org:HuggingFace --watch model:gpt2 --domain repo

>>> hf webhooks create --job-id 687f911eaea852de79c4a50a --watch user:julien-c

>>> hf webhooks create --job-id 687f911eaea852de79c4a50a --watch bucket:my-org/my-bucket --secrets HF_TOKEN

>>> hf webhooks update wh-abc123 --url https://new-url.com/hook
>>> hf webhooks update wh-abc123 --watch model:gpt2 --domain repo

>>> hf webhooks update wh-abc123 --job-id 687f911eaea852de79c4a50a --secrets HF_TOKEN

>>> hf webhooks enable wh-abc123
>>> hf webhooks disable wh-abc123

>>> hf webhooks delete wh-abc123
Are you sure you want to delete webhook 'wh-abc123'? [y/N]: y
Webhook deleted: wh-abc123

>>> hf webhooks delete wh-abc123 --yes

# Lists endpoints in your namespace
>>> hf endpoints ls

# List the hardware you can deploy on
>>> hf endpoints hardware

# Deploy an endpoint from Model Catalog
>>> hf endpoints catalog deploy --repo openai/gpt-oss-120b --name my-endpoint

# Deploy an endpoint from the Hugging Face Hub
>>> hf endpoints deploy my-endpoint --repo gpt2 --framework pytorch --accelerator cpu --instance-size x2 --instance-type intel-icl

# List catalog entries
>>> hf endpoints catalog ls

# Show status and metadata
>>> hf endpoints describe my-endpoint

# Pause the endpoint
>>> hf endpoints pause my-endpoint

# Delete without confirmation prompt
>>> hf endpoints delete my-endpoint --yes

>>> hf endpoints hardware --vendor aws --region eu-west-1
VENDOR REGION    ACCELERATOR INSTANCE_TYPE INSTANCE_SIZE MEMORY_GB GPU_MEMORY_GB PRICE_PER_HOUR QUOTA STATUS
------ --------- ----------- ------------- ------------- --------- ------------- -------------- ----- ---------
aws    eu-west-1 cpu         intel-spr     x1                  2.0                        0.033 0/60  available
aws    eu-west-1 cpu         intel-spr     x2                  4.0                        0.067 0/60  available
aws    eu-west-1 cpu         intel-spr     x4                  8.0                        0.134 0/60  available
aws    eu-west-1 cpu         intel-spr     x8                 16.0                        0.268 0/60  available
aws    eu-west-1 cpu         intel-spr     x16                32.0                        0.536 0/60  available
aws    eu-west-1 gpu         nvidia-a10g   x1                 30.0            24            1.0 0/16  available
aws    eu-west-1 gpu         nvidia-t4     x1                 15.0            16            0.5 1/30  available
Hint: Deploy on one of these, e.g.: hf endpoints deploy my-endpoint --repo <repo> --framework <framework> --vendor aws --region eu-west-1 --accelerator cpu --instance-type intel-spr --instance-size x1

>>> hf endpoints deploy nex-n2-pro \
      --repo nex-agi/Nex-N2-Pro \
      --framework custom \
      --accelerator gpu --vendor aws --region us-east-1 \
      --instance-type nvidia-h200 --instance-size x8 \
      --custom-image nexagi/sglang:v0.5.12 \
      --health-route /health --port 30000 \
      --container-args "--reasoning-parser qwen3 --tool-call-parser qwen3_coder --mamba-scheduler-strategy extra_buffer --tp 8" \
      --env MODEL_ID=/repository \
      --type authenticated

>>> hf endpoints update my-endpoint \
      --container-args "--enable-auto-tool-choice --tool-call-parser lfm2"

>>> hf endpoints deploy gpt-oss-120b-vllm \
      --repo openai/gpt-oss-120b \
      --framework custom \
      --accelerator gpu --vendor aws --region us-east-1 \
      --instance-type nvidia-h200 --instance-size x8 \
      --engine vllm --custom-image vllm/vllm-openai:v0.23.0 \
      --tensor-parallel-size 8

>>> hf endpoints update gpt-oss-120b-vllm --tensor-parallel-size 4 --data-parallel-size 2

# List available catalog models
>>> hf endpoints catalog ls
REPO_ID                             TASK               LICENSE    ACCELERATOR ENGINE   GGUF_FILE                     RECIPE_ID
----------------------------------- ------------------ ---------- ----------- -------- ----------------------------- -------------------------
meta-llama/Llama-3.1-8B-Instruct    text-generation    Llama 3.1  gpu         vllm                                   sizzling-biryani-g4xsi1ac
meta-llama/Llama-3.1-8B-Instruct    text-generation    Llama 3.1  neuron      vllmNeuron                             artisanal-quinoa-yz9ynamx
bartowski/QwQ-32B-Preview-GGUF      text-generation    Apache 2.0 gpu         llamacpp QwQ-32B-Preview-Q8_0.gguf     baked-orange-m863gx7d

# Narrow it down
>>> hf endpoints catalog ls --engine vllm --task text-generation --search llama --limit 10

# Deploy the default recipe of a model
>>> hf endpoints catalog deploy --repo meta-llama/Llama-3.2-1B-Instruct

# Deploy an exact recipe listed above
>>> hf endpoints catalog deploy --recipe artisanal-quinoa-yz9ynamx

# Deploy with a custom name
>>> hf endpoints catalog deploy --repo meta-llama/Llama-3.2-1B-Instruct --name my-llama-endpoint