Remotehost Docs

Pricing & Billing

How usage is metered and what each plan includes

Remotehost has two costs: a plan's base price, and usage — allocated resource time for the sandboxes you run. Sandboxes are metered by the second while running; stopping one stops its billing immediately.

Sandbox profiles

New sandboxes should be created with a workload profile. auto is the default profile: it starts from the smallest viable shape and can grow from measured CPU, memory, disk, and runtime pressure. Named profiles make heavier intent explicit when you already know the workload shape.

ProfileStarts atUse case
auto2 vCPU / 4 GBGeneral agent tasks with measured adjustment
agent-small2 vCPU / 4 GBLightweight editing, review, and small repos
agent-standard4 vCPU / 8 GBDefault coding tasks and medium repos
build-heavy8 vCPU / 16 GBBuild, install, and test-heavy workloads
browser-heavy4 vCPU / 12 GBBrowser automation and UI-heavy agent runs
repo-large4 vCPU / 12 GB / 160 GB diskLarge checkouts and dependency caches
long-running2 vCPU / 4 GB / 80 GB diskPersistent low-duty sandboxes

Machine sizes

Machine sizes are the current coarse allocation tiers behind profile starts and legacy clients:

SizeCPUMemoryMachine-time multiplier
small2 vCPU4 GB0.5x
default4 vCPU8 GB1x
large8 vCPU16 GB2x

An hour of a large sandbox running counts as 2 machine-hours against your plan's included allowance or overage; an hour of small counts as 0.5. Profile-aware billing will keep the same second-level metering discipline and rate usage from the resources actually allocated to the sandbox. Newer usage events record allocated vCPU, memory, and disk directly; older events fall back to sandbox profile or legacy machine size.

Plans

Three single-player plans for one engineer, then Business for a team, then Enterprise. Every plan bills additional usage by machine time; the rate depends on the infrastructure the plan runs on.

Individual plans

Starter, Pro, and Max are single-player: no team members, invitations, or shared-org billing. Each rung includes more machine-hours and concurrency than the one below. They are the plans for the person evaluating and advocating for Remotehost, or running it alone at scale — move to Business once you're ready to bring others on.

PlanPriceIncluded machine-hoursConcurrent sandboxesSessionsAlso
Starter$20/mo61035 included, up to 1,20024 hCommunity support
Pro$100/mo2,500100 included, up to 2,00024 hRepository templates
Max$200/mo6,000250 included, up to 2,00072 hCustom base images, repository templates, priority support

All three carry a 7-day free trial and are self-serve. Overage is $0.64 per standard machine-hour past the included allowance, with profile rates metered by the second.

Business — $100/seat/mo

7-day free trial, self-serve. Priced per seat with a 1-seat minimum, so a single engineer starts at $100/mo and the bill grows with the team. Two seat types, mixed freely and adjusted any time from Billing:

SeatPriceIncluded machine-hoursFor
Standard$100/seat/mo200 per seatEveryday agent work
Premium$300/seat/mo1,000 per seatEngineers who run agents all day

The org's included allowance is the sum across its seats: four standard and one premium seat include 1,800 machine-hours a month.

  • Members and pending invites up to your total seat count
  • Dedicated sandbox pool for production workloads
  • Shared org billing, limits, and controls
  • Audit logs for sandbox lifecycle events
  • Custom base images and repository templates
  • Up to 24-hour sandbox sessions
  • Priority support
  • Overage: $4.712 per standard machine-hour past the included allowance, with profile rates metered by second. Business runs on the dedicated pool, whose machine-hours cost several times a single-player plan's, which is why a seat includes fewer hours than Pro at the same price.

Enterprise — Custom

Contact sales@remotehost.ai.

  • Everything in Business
  • SSO and SCIM
  • Custom concurrency limits
  • Dedicated regions and capacity planning
  • Custom dedicated host pools
  • Security review and procurement support

Managing billing

Subscribe, change plans, and view usage against your included allowance from the console's Billing page.

Checkout requires acknowledgement that sandbox usage is measured from allocated resource-time and may be billed after the included allowance. The same disclosure is sent to Stripe Checkout for the subscription.

Usage guardrails

Subscriptions have sandbox usage guardrails:

  • sandboxMachineHourLimit is the hard cap for the current billing period.
  • sandboxOverageEnabled controls whether usage may continue past the included allowance.
  • When overage is disabled and no explicit cap is set, the effective cap is the plan's included machine-hour allowance.

The API checks this cap before creating or waking sandboxes, and current usage includes both ended interval events and still-running allocation intervals. The console surfaces warning states at 80%, 95%, and 100% of the effective cap.

Metronome prices usage by billing_profile, not legacy machine_size. agent-small, agent-standard, browser-heavy, repo-large, build-heavy, long-running, and custom-heavy each have explicit profile rates. auto maps to the billing profile that matches the actual allocation for that runtime interval.

On this page