Deployment
Plugin
skypilot
The AI Compute Platform for frontier teams. SkyPilot turns fragmented AI compute into one AI supercomputer, so frontier AI teams build custom intelligence faster.
Plugin Signal72/100Solid
maintained100
capable14
documented100
adopted86
Install on Claude Code
> /plugin marketplace add skypilot-org/skypilot> /plugin install skypilot@skypilot
Stats
10,466
Stars
1,180
Forks
Active
Maintenance
Python
Language
Apache-2.0
License
1h ago
Last commit
4y ago
Created
Repo: skypilot-org/skypilot
What's inside
Badges
5
README
1SkyPilot is a system to run, manage, and scale AI workloads on any AI infrastructure.
SkyPilot gives AI teams a simple interface to run jobs on any infra. Infra teams get a unified control plane to manage any AI compute — with advanced scheduling, scaling, and orchestration.
:fire: News :fire:
- [Jun 2026] Announcing SkyPilot Sandboxes: run untrusted, LLM-generated code on the Kubernetes clusters you already own. Learn more, join early access
- [May 2026] How Multiverse doubled their GPU utilization with SkyPilot: case study
- [Apr 2026] Introducing GPU Compass: One dashboard to browse, compare pricing, and launch across every GPU cloud. Try it at gpus.skypilot.co.
- [Apr 2026] Research-Driven Agents: Agents read arxiv papers before coding, landed 5 llama.cpp kernel fusions and +15% faster flash attention in ~3 hours for ~$29: blog, HackerNews
- [Mar 2026] Scaling Karpathy's Autoresearch: Autoresearch runs 1 experiment at a time. We gave it 16 GPUs and let it run in parallel: blog, HackerNews
- [Mar 2026] How H Company Unlocked Online RL and Unified their AI Platform: case study
- [Mar 2026] SkyPilot v0.12 released: Slurm Support, Job Groups for RL, Agent Skill, Recipes, Pool Autoscaling for Batch Inference, 7x Data Mounting, and More: Release notes
- [Mar 2026] SkyPilot Agent Skills: GPU access and job management for AI agents: docs
- [Jan 2026] Shopify case study: Shopify runs all AI training workloads on SkyPilot: case study
Overview
SkyPilot is easy to use for AI users:
- Quickly spin up compute on your own infra
- Environment and job as code — simple and portable
- Easy job management: queue, run, and auto-recover many jobs
SkyPilot makes Kubernetes easy for AI & Infra teams:
- Slurm-like ease of use, cloud-native robustness
- Local dev experience on K8s: SSH into pods, sync code, or connect IDE
- Turbocharge your clusters: gang scheduling, multi-cluster, and scaling
SkyPilot unifies multiple clusters, clouds, and hardware:
- One interface to use reserved GPUs, Kubernetes clusters, Slurm clusters, or 20+ clouds
- Flexible provisioning of GPUs, TPUs, CPUs, with smart failover
- Team deployment and resource sharing
SkyPilot maximizes GPU fleet utilization:
- Autostop: automatic cleanup of idle resources
- Binpacking: workload binpacking on shared clusters
- Intelligent scheduler: automatically schedule on the most available infra
SkyPilot supports your existing GPU, TPU, and CPU workloads, with no code changes.
Install with uv (also supported: pip, nightly, from source)
# Choose your clouds:
uv pip install "skypilot[kubernetes,aws,gcp,azure,oci,nebius,lambda,runpod,fluidstack,paperspace,cudo,ibm,scp,seeweb,shadeform,verda]"
To use SkyPilot directly with your agent (Claude Code, Codex, etc.), install the SkyPilot Skill. Tell your agent:
Fetch and follow https://github.com/skypilot-org/skypilot/blob/HEAD/agent/INSTALL.md to install the skypilot skill
Current supported infra: Kubernetes, Slurm, AWS, GCP, Azure, OCI, CoreWeave, Nebius, Lambda Cloud, RunPod, Fluidstack, Cudo, Digital Ocean, Paperspace, Cloudflare, Samsung, IBM, Vast.ai, VMware vSphere, Seeweb, Prime Intellect, Shadeform, Verda Cloud, VastData, Crusoe.
Getting started
Install SkyPilot in 1 minute. Then, launch your first cluster in 2 minutes in Quickstart.
SkyPilot is BYOC: Everything is launched within your cloud accounts, VPCs, and clusters.
Benefits of SkyPilot on Kubernetes
SkyPilot makes Kubernetes AI-native.
It turbocharges your existing Kubernetes clusters by accelerating AI/ML velocity:
- AI-friendly interface to launch jobs and deployments
- Much simplified interactive dev for K8s (SSH / sync code / connect IDE to pods)
...and optimizing GPU scheduling, utilization, and scaling:
- Advanced scheduling: Gang scheduling, multi-node jobs, and queueing
- Multi-cluster support: Bring all your clusters under one control plane
- Multi-cloud support: One consistent interface to manage many providers
See SkyPilot vs Vanilla Kubernetes and this blog post for more details.
SkyPilot in 1 minute
A SkyPilot task specifies: resource requirements, data to be synced, setup commands, and the task commands.
Once written in this unified interface (YAML or Python API), the task can be launched on any available infra (Kubernetes, Slurm, cloud, etc.). This avoids vendor lock-in, and allows easily moving jobs to a different provider.
Paste the following into a file my_task.yaml:
resources:
accelerators: A100:8 # 8x NVIDIA A100 GPU
num_nodes: 1 # Number of VMs to launch
# Working directory (optional) containing the project codebase.
# Its contents are synced to ~/sky_workdir/ on the cluster.
workdir: ~/torch_examples
# Commands to be run before executing the job.
# Typical use: pip install -r requirements.txt, git clone, etc.
setup: |
cd mnist
pip install -r requirements.txt
# Commands to run as a job.
# Typical use: launch the main program.
run: |
cd mnist
python main.py --epochs 1
Prepare the workdir by cloning:
git clone https://github.com/pytorch/examples.git ~/torch_examples
Launch with sky launch (note: access to GPU instances is needed for this example):
sky launch my_task.yaml
SkyPilot then performs the heavy-lifting for you, including:
- Find the cheapest & available infra across your clusters or clouds
- Provision the GPUs (pods or VMs), with auto-failover if the infra returned capacity errors
- Sync your local
workdirto the provisioned cluster - Auto-install dependencies by running the task's
setupcommands - Run the task's
runcommands, and stream logs
See Quickstart to get started with SkyPilot.
Runnable examples
See SkyPilot examples that cover: development, training, serving, LLM models, AI apps, and common frameworks.
Latest featured examples:
| Task | Examples |
|---|---|
| Training | Verl, Finetune Llama 4, TorchTitan, PyTorch, DeepSpeed, NeMo, Ray, Unsloth, Jax/TPU, OpenRLHF |
| Serving | vLLM, SGLang, Ollama |
| Models | DeepSeek-R1, Llama 4, Llama 3, CodeLlama, Qwen, Kimi-K2, Kimi-K2-Thinking, Mixtral |
| AI apps | RAG, vector databases (ChromaDB, CLIP) |
| Common frameworks | Airflow, Jupyter, marimo |
Source files can be found in llm/ and examples/.
Learn more
To learn more, see SkyPilot Overview, SkyPilot docs, and SkyPilot blog.
SkyPilot adopters: Testimonials and Case Studies
Follow updates:
- Slack
- X
- YouTube
- SkyPilot Blog
Questions and feedback
We are excited to hear your feedback:
- For issues and feature requests, please open a GitHub issue.
- For questions, please use GitHub Discussions.
For general discussions, join us on the SkyPilot Slack.
Contributing
We welcome all contributions to the project! See CONTRIBUTING for how to get involved.
File tree
1651skypilot-org/skypilotView on GitHub
.buildkite/
generate_pipeline.py
test_buildkite_pipeline_generation.py
.claude-plugin/
marketplace.json
.cursor/
worktrees.json
.dockerignore
.github/
actions/
run-python-tests/
action.yaml
issue_template.md
pull_request_template.md
workflows/
buildkite-trigger-wait.yml
compile-protos-check.yml
dashboard.yml
docker-build.yaml
extract-buildkite.yml
format.yml
fusermount-server-image.yaml
go-reviewable.yaml
helm-docker-release.yaml
helm-unittest.yaml
helm-values-schema.yaml
image-vulnerability.yml
mypy.yml
nightly-build.yml
publish-and-validate-both.yml
publish-and-validate.yml
publish-helm.yml
pylint.yml
pytest-generic.yml
pytest-optimizer.yml
pytest.yml
readthedocs-manual-pr-build.yml
release-build.yml
release-publish.yml
skill-references-reviewable.yaml
smoke-tests-trigger.yaml
stale.yml
test-doc-build.yml
wait-for-buildkite.yml
.gitignore
.pre-commit-config.yaml
.pylintrc
.readthedocs.yml
.trivyignore
addons/
fuse-proxy/
.gitignore
cmd/
fusermount-server/
main.go
fusermount-shim/
main.go
fusermount-wrapper/
main.go
Dockerfile
go.mod
go.sum
Makefile
pkg/
client/
client.go
common/
common.go
server/
flags.go
server.go
README.md
agent/
.claude-plugin/
plugin.json
INSTALL.md
scripts/
generate_references.py
skills/
skypilot/
evals/
evals.json
references/
advanced-patterns.md
cli-reference.md
examples.md
python-sdk.md
troubleshooting.md
yaml-spec.md
SKILL.md
test/
.gitignore
diagnose.py
eval_set.json
test_trigger.py
test.sh
AGENTS.md
charts/
skypilot/
.gitignore
.helmignore
Chart.yaml
DCGM_MONITORING.md
developer.md
manifests/
api-server-overview.json
autoscaling-dashboard.json
compute-overview-dashboard.json
dcgm-cluster-dashboard.json
dcgm-cluster-filter-dashboard.json
vllm-serving-dashboard.json
README.md
skypilot.svg
templates/
_helpers.tpl
api-configmap.yaml
api-dashboard-grafana-configmap.yaml
api-deployment.yaml
api-secrets.yaml
api-service.yaml
auth.yaml
autoscaling-dashboard-grafana-configmap.yaml
compute-overview-dashboard-grafana-configmap.yaml
datasource.yaml
db-secrets.yaml
dcgm-cluster-dashboard-grafana-configmap.yaml
dcgm-cluster-filtering-dashboard-grafana-configmap.yaml
dcgm-prometheus-scrape-service.yaml
grafana-ingress.yaml
ingress.yaml
initial-auth.yaml
NOTES.txt
oauth2-proxy-deployment.yaml
oauth2-proxy-ingress.yaml
oauth2-proxy-redis.yaml
oauth2-proxy-service.yaml
pvc.yaml
rbac.yaml
server-config.yaml
system-rbac.yaml
vllm-serving-dashboard-grafana-configmap.yaml
tests/
api_secrets_test.yaml
api_service_test.yaml
db_secrets_test.yaml
deployment_test.yaml
fixtures/
oauth2-scheduling-values.yaml
ingress_test.yaml
metrics_test.yaml
oauth2_test.yaml
values.schema.json
values.yaml
CLAUDE.md
CONTRIBUTING.md
Dockerfile
Dockerfile_k8s
Dockerfile_k8s_gpu
docs/
.gitignore
.nojekyll
build.sh
make.bat
Makefile
README.md
repo-images/
managed-job-schedule-state-diagram.png
managed-job-status-diagram.png
README
requirements-docs.txt
source/
_static/
custom.css
custom.js
favicon.ico
intro.gif
intro.mp4
rtd-data.js
sandbox-chat.mp4
sandboxes-diagram.html
SkyPilot_wide_dark.svg
SkyPilot_wide_light.svg
skypilot-icon-dark.svg
skypilot-icon.png
_templates/
author.html
layout.html
main-sidebar-home.html
main-sidebar.html
navbar-skypilot-logo.html
try-platform-button.html
.gitignore
admin/
workspaces.rst
cloud-setup/
airgap.rst
cloud-permissions/
aws-eks-iam-roles.rst
aws.rst
gcp.rst
index.rst
kubernetes.rst
nebius.rst
vsphere.rst
external-logging.rst
policy.rst
quota.rst
compute/
cloud-vm.rst
gpus.rst
show-gpus-all.txt
show-gpus-h100-8.txt
conf.py
developers/
CONTRIBUTING.md
index.rst
docs/
index.rst
examples/
agents/
autonomous-code-optimization.md
autoresearch.md
gpu-job-management.rst
index.rst
slime.md
applications/
batch_inference.md
deepseek-ocr.md
index.rst
localgpt.md
pools_batch_inference.md
rag.md
redisvl-vector-search.md
sam3-video-segmentation.md
stable_diffusion.md
streamlit.md
tabby.md
vector_database.md
auto-failover.rst
batch/
custom-formats.rst
index.rst
checkpointing.rst
docker-containers.rst
frameworks/
dvc.md
index.rst
jupyter.md
marimo.md
mpi.md
spyder.md
index.rst
interactive-development.rst
job-groups.rst
managed-jobs.rst
models/
codellama.md
dbrx.md
deepseek-janus.md
deepseek-r1-distilled.md
deepseek-r1.md
gemma.md
gemma3.md
gpt-2.md
gpt-oss.md
index.rst
kimi-k2-thinking.md
kimi-k2.md
llama-2.md
llama-3_1.md
llama-3_2.md
llama-3.md
llama-4.md
mixtral.md
pixtral.md
qwen.md
vicuna.md
yi.md
orchestrators/
airflow.md
cron.md
github_actions.md
index.rst
prefect.md
temporal.md
performance/
aws_efa.md
coreweave_infiniband.md
gcp_gpu_direct_tcpx.md
index.rst
nebius_infiniband.md
together_infiniband.md
pools.rst
ports.rst
serving/
cog.md
index.rst
lorax.md
nvidia-dynamo.md
ollama.md
sglang.md
tgi.md
vllm.md
spot/
spot-jobs.rst
bert_qa.rst
lightning_cifar10.rst
resnet.rst
syncing-code-artifacts.rst
training/
axolotl.md
cosmos3-finetuning.md
deepspeed.md
distributed-pytorch.md
distributed-tensorflow.md
fairseq2.md
gpt-oss-finetuning.md
index.rst
llama-2-finetuning.md
llama-3_1-finetuning.md
llama-4-finetuning.md
nanochat.md
nemo.md
nemorl.md
openrlhf.md
ray.md
skyrl.md
torchtitan.md
tpu.md
unsloth.md
verl.md
vector_database.rst
extension/
dynamic_llms_txt.py
linting.py
markdown_export.py
frontier-ai.rst
gallery/
applications/
localgpt.rst
rag.rst
tabby.rst
vector_database.rst
frameworks/
lorax.rst
ollama.rst
sglang.rst
tgi.rst
vllm.rst
index.rst
llms/
codellama.rst
dbrx.rst
deepseek-janus.rst
deepseek-r1-distilled.rst
deepseek-r1.rst
gemma.rst
gpt-2.rst
index.rst
llama-2.rst
llama-3_1.rst
llama-3_2.rst
llama-3.rst
mixtral.rst
pixtral.rst
qwen.rst
vicuna.rst
yi.rst
tutorials/
finetuning.rst
generate_examples.py
getting-started/
installation.rst
quickstart.rst
skill.rst
tutorial.rst
images/
admin-policy.png
ai-gallery-cover.png
client-server/
arch.png
auth-proxy-internals.svg
basic-auth-user-flow.png
cluster-users.png
executor.png
google-auth-initial-setup.png
google-auth-setup.png
high-level-arch.png
local.png
login.png
microsoft-auth-initial.png
oauth-user-flow.svg
okta-setup.png
remote.png
service-account-architecture.svg
service-account.png
users.png
cloud-logos-dark.png
cloud-logos-light.png
config-cheatsheet-dark.svg
... 1600 moreShowing a partial view of a very large repo.
FAQ
skypilot is a Claude Code plugin with 1 hand-picked skill for deployment work, indexed on Flowy. Install it with the command on its page. It includes skypilot. Its skills do not fire on their own yet. Request auto-invocation to have Flowy route them as you prompt. Free and open source.
- What is skypilot?
- The AI Compute Platform for frontier teams. SkyPilot turns fragmented AI compute into one AI supercomputer, so frontier AI teams build custom intelligence faster.
- How do I install skypilot?
- Run these in Claude Code: /plugin marketplace add skypilot-org/skypilot then /plugin install skypilot@skypilot. Then prompt normally.
- Does skypilot auto-invoke its skills?
- Not yet. It is indexed on Flowy as a plain plugin. Request auto-invocation on its page and Flowy will route its skills for you as you prompt.
- Is skypilot free?
- Yes. Flowy is free and open source, with nothing gated. You can read every skill in full before you install.