Skip to content
chalupa
How it worksGPU inferencePricingBlogDocumentation
Open console
Menu
How it worksGPU inferencePricingBlogDocumentationOpen console

Private compute. Your cloud account.

Remote GPU.
Local workflow.

Give your model more room. Connect a remote GPU to the tools on your laptop over SSH. Run temporary Compose environments and CI jobs with the same CLI.

Launch a model See how it works
Free local CLIAccess over SSH
install.sh install
curl -fsSL https://chalupa.run/install.sh | bash

Installs bun, go-task, and the CLI for you. Then run chalupa setup to connect your cloud account, SSH key, and console account.

Your laptopSSH tunnelHost Ollama / GPU
host Ollama · experimentalconfiguration
inference:
  model: qwen3.8:27b
  pull: true
  contextSize: 65536
  maxModelBytes: 25769803776

Configuration excerpt · requires a session and local expiry scheduler

Remote GPU. Local workflow.

Serve open weights on your own host and connect your local tools through SSH.

Inference command
chalupa launch opencode
Read the agent harnesses guide

Your infrastructure. Direct provider billing.

GPU modelsCompose environmentsCI jobs

GPU inference · experimental

More room for your model.
Keep your tools close.

Choose an open-weight model served by Ollama. Chalupa provisions the GPU host and connects your agent through a local endpoint.

Connect your agent
01

Choose the model and capacity

Set the model, GPU size and session duration in chalupa.yml. Preview the estimated hourly rate before launching.

chalupa up
02

Use the available context

Request a context window or negotiate one against the model ceiling and observed GPU placement. Weights, context and runtime all need memory.

Context guide
03

See what each session reports

With hosted reporting configured, inspect GPU memory and utilization alongside reported tokens, speed and turn outcomes. Then stop compute with chalupa down.

Open inference dashboard

A short-lived server. A familiar workflow.

From your project.
Back to your project.

Chalupa handles the environment lifecycle through the CLI. Your editor, tools, and cloud account stay yours.

01

Describe

Start with your Compose file, a CI suite, or an inference configuration. Preview the resolved plan offline.

chalupa preview
02

Connect

Confirm the launch, then open an SSH tunnel. Service ports bind to localhost on the remote host.

chalupa tunnel
03

Work

Use your environment. The optional hosted console keeps status, estimated costs, and reported results in view.

chalupa status
04

Leave cleanly

Tear down compute explicitly. Declared persistent data lives in its own stack and continues to be billed separately.

chalupa down

Clear boundaries

Temporary compute.
Lasting control.

Access, storage, and lifecycle each have a defined place. The console observes your fleet; launch and teardown stay in your CLI.

Explore the operating model

A private way in

Service ports stay on localhost. SSH is the entry point, scoped by your configured network allowlist.

Data has its own lifecycle

Protected volumes live in a separate stack. Stopping compute preserves declared persistent storage.

An explicit way out

Launch and teardown ask for confirmation. Inference expiry needs an awake Mac and unlocked credentials; it is not a hard spending cap.

Free CLI. Optional hosted visibility.

Pay for the view.
Own the compute.

Keep provider billing in your account. Hosted plans add fleet history, CI evidence, and cost visibility. GPU time and model downloads are separate.

Compare plans

Community

The complete local CLI and TUI for your own infrastructure.

$0/mo

Solo

Hosted operations and visibility for an individual developer.

$19/mo

Crew

More capacity, collaborators, and history for small teams.

$79/mo

Billing beta: usage is recorded; prepaid packs and overages are not enabled.

Start where you already work

Pick a model.
Keep your workflow.

Launch a model Start with Compose
chalupa

Compute for the work.
Control for what comes next.

DocumentationPricingBlogGet in touch

Bring your own cloud. Leave with your data.

Provider costs are estimates. Storage has its own lifecycle.