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

Hosted plans and separate compute

Pay for the view. Choose your compute.

The plan does not include GPU time or droplets. It includes fleet history, cost estimates, and a way to shut them down.

Billing beta is observation-only

  • Provider invoice is authoritative.
  • Estimates are labeled estimated.
  • Powered-off GPU droplets keep billing until destroyed.
  • Persistent volumes and explicitly retained model caches keep billing after chalupa down.
  • Billing beta: observation only; extras are not charged yet.

Chalupa records managed-hour usage and shows balance alerts, but an empty balance does not block signed activity today. Hosted-hour packs and overages are not enabled. Managed GPU uses a separate USD balance.

On your machine

Community

The complete local CLI and TUI for your own infrastructure.

$0/ month

  • Local CLI and TUI
  • Bring your own cloud
  • 1 organization seat
  • Local history
Install the CLI

Hosted · individual

Solo

Hosted operations and visibility for an individual developer.

$19/ month

  • 500 managed hours
  • 3 concurrent environments
  • 1 organization seat
  • 7-day retention
Choose Solo

Hosted · team

Crew

More capacity, collaborators, and history for small teams.

$79/ month

  • 2,500 managed hours
  • 10 concurrent environments
  • 5 organization seats
  • 30-day retention
Choose Crew

Managed hours are hours the control plane observed signed activity; not GPU hours, not droplet hours.

Managed GPU

Or let us run the GPU.

Prepaid balance, Chalupa's cloud account, destroyed at the deadline. Requires Solo or Crew.

Managed GPU hourly rates in USD
TierGPUVRAM$/h
gpu-smallRTX 4000 Ada20 GiB$0.95
gpu-largeRTX 6000 Ada48 GiB$1.97
gpu-maxH10080 GiB$5.52
gpu-ultraH100 x8640 GiB$44.10

Models use temporary host storage by default and download on each launch. GPU time includes downloads. Optional retained model cache: $0.13/GiB-month.

  • Rates include the provider price and Chalupa's margin.
  • Balance is checked at launch; reservations and automatic usage debits are not yet implemented.
  • Balance is not a spending cap. Sessions stop at their configured deadline or after a successful chalupa down.
  • Powered-off managed hosts are destroyed, not parked.
  • Balance is refundable within 30 days.

What a managed hour means

Managed hours are hours the control plane observed signed activity; not GPU hours, not droplet hours. Sunk compute does not consume hours. Safety signals and teardown remain available even when the included balance is exhausted.

Compute is separate

BYOC servers and persistent volumes run in your own cloud account, with direct provider billing and no Chalupa markup. Managed GPU runs in Chalupa’s account at the separate rates shown above.

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.