Deploy Anything.Control Everything.
A project-based container platform to run apps, APIs, and AI workloads with full infrastructure control, no server management required.
Deploy anything
Built for the runtimes teams already use.
Single containers, full compose stacks, GPU workloads, or static sites Hubfly space runs it all inside private project networks.
Compose stacks
Deploy full stacks, not just containers.
Or import your docker-compose.yml in the visual builder. Every service communicates privately over the project network.
Explore all stacksPlatform Model
One project. Every layer.
Deployments, networking, storage, team access, and billing live inside a single project not spread across separate tools.
Deployment model
Four paths, one project.
Each path lands in the same project structure same networking, volumes, domains, and team access.
GitHub repo → build → image → run
Any registry, public or private
PostgreSQL, Redis, Mongo & more
Multi-service YAML or visual builder
Hubfly space Project
Private network · Volumes · Domains · Team
Networking model
Private by default, public when you decide.
Services in the same project talk privately. Only selected ports touch the public edge.
Project private network
Private by default
All containers are unreachable from the internet unless you explicitly expose a port.
HTTP & TCP endpoints
Expose services over HTTP or raw TCP. SSL and force-SSL are one toggle away.
Load balancer groups
Route traffic across multiple containers with round-robin, least-conn, or IP-hash.
Custom domains
Verify a domain once, reuse it across any service with SSL and proxy templates.
HubTunnels
Reach databases and internal tools privately through the CLI no public port needed.
Storage model
Persistent volumes, built in.
Attach volumes to single containers or entire compose stacks. Pick a performance tier and go.
Managed Volume
Persistent block storage · survives redeploys
Postgres
Uploads
App state
Performance tier
Block storage
Volumes persist across container restarts and redeploys. Your data stays put.
Works in compose stacks
Define named volumes inside compose YAML and share them across services in a stack.
Read-only mode
Mount volumes as read-only for shared asset directories or audit safe workloads.
Cost estimation
See estimated volume cost before creation. Billed hourly at $0.00001536 / GB.
Performance tiers
Choose Standard, Balanced, or High Performance up to 16k IOPS for demanding databases.
Team + billing
Scoped access. Full cost visibility.
Permissions per resource category, usage charts, and projected spend all inside the project.
Permission matrix
| Role | Project | Containers | Storage | Network | Billing | Team |
|---|---|---|---|---|---|---|
| Owner | ||||||
| Infra Admin | ||||||
| Developer | ||||||
| Operator | ||||||
| Viewer |
Daily spend
This week$9.90
~$42 projected this month
Account balance
$48.20
GPU wallet
$15.00
Free hours used
240 / 360
Free storage
0.4 / 1 GB
GPU workflows
Accelerated compute, inside the platform.
Search offers, fund a wallet, provision a GPU instance, and launch Jupyter or SSH without leaving Hubfly space.
Offer example
NVIDIA A100 · 80 GB
4× GPU · EU-West · 500 GB disk
$3.20/hr
Instance lifecycle
Offer search
Browse GPU offers by price, region, and hardware. Sort and compare before committing.
Jupyter + SSH templates
Launch notebook-ready or SSH-capable instances from templates without manual setup.
Separate GPU wallet
Top up a GPU specific wallet. Transfers require approval keeping spend intentional.
Project or global scope
Bind a GPU instance to a project or keep it global at the owner level.
Full lifecycle control
Create, start, stop, sync, and destroy GPU instances from the dashboard.
Powerful CLI
Deploy locally. Tunnel privately. Automate the rest.
The Hubfly space CLI is built for teams that want fast deploys from the terminal, secure access to internal services, and enough scripting support to move the routine work out of the dashboard.
Deploy from your machine
hubfly deployBuild locally, upload the image to Hubfly space, and keep redeploys tied to the same service through `hubfly.build.json`.
Reach private services
hubfly tunnel createOpen short-lived local access to Postgres, Redis, internal APIs, and admin tools without exposing public ports.
hubfly-cli
AI integration
Give your coding agent the real Hubfly space deployment model.
Install one portable Markdown skill so your AI agent can inspect the repository, reconcile local and cloud configuration, review the Dashboard deployment plan, and follow the real rollout after you approve it.
Codex · Claude Code · Cursor · Agent Skills-compatible tools
Inspect
Read the app and its schema-v2 Hubfly space manifest without exposing secrets.
Reconcile
Compare managed local fields with the bound cloud container.
Plan
Review Dashboard's revision-guarded plan and destructive operations.
Follow
After approval, track the deployment session to its real terminal state.
Developer sub-accounts
Build your product on isolated Hubfly space tenants.
Create customer, agent, or automation environments from one parent account. Every tenant gets its own wallet, projects, resources, audit context, and rotatable machine credentials.
Parent platform
hfpk_… + X-HubFly-Subaccount
Customer A
sub_…
Wallet
$42.18
Agent fleet
sub_…
Wallet
$18.70
Preview API
sub_…
Wallet
$7.25
Scoped keys
Tenant isolation
Machine audit
Comparison
Built between typical PaaS and cloud infrastructure.
More control than a typical app host. Less infrastructure work than wiring raw cloud services yourself.
Typical PaaS
Hubfly space
Cloud Infra
Deployment paths
Git, image, template, compose stack
Git push only
Manual image setup
Private networking
Per-project, automatic
Shared or not available
VPC setup required
Persistent storage
Managed volumes, tiered performance
Limited or add-on only
Block storage, fully manual
Multi-service stacks
Compose with visual builder
Single container focus
Possible but no tooling
Custom domains + SSL
Included, custom proxy templates
Usually included
Configure Nginx, Certbot yourself
Load balancing
Built-in, weighted routing
Basic only
Provision a separate LB
Team permissions
5 role templates, 13 permissions
Basic or no team model
IAM roles, complex setup
GPU workloads
Built in, same platform
Not available
Separate GPU instance
Private tunnel access
HubTunnels via CLI
Not available
SSH + port forwarding
Usage-based billing
Per vCPU/GB/hr, free shared tier
Fixed tiers, often overpriced
Per instance, hard to predict
Pricing Snapshot
Usage-based pricing, directly tied to resources.
CPU
$0.004
per vCPU / hour
RAM
$0.002
per GB / hour
Storage
$0.00001536
per GB / hour
Shared tier
360h
free runtime monthly + 1GB storage
Start With Structure
Build, deploy, route, store, and operate in one product.
Hubfly space brings deployments, private networking, storage, billing, domains, tunnels, and GPU workflows into one coherent model.