Plans & pricing

Every plan runs the same Firecracker microVMs — the same isolation and the same per-fork performance. Plans differ on concurrency, session duration, and support, never on the sandbox itself.

Billing is per-second metered CPU / RAM / disk while a sandbox is live, plus snapshot storage per GB-month while it is suspended. Suspending a sandbox stops the meter.

Plans

Demo — free

No card required.

• 2 concurrent sandboxes
• Fork wide: up to 1,000 per group
• 1-hour max session
• Shared capacity
• Community support

For trying every endpoint in these docs.

Pro — usage-based

A monthly base plus metered usage.

• No concurrency limit; run at extreme width, billed per second
• 24-hour max session
• Standard support

For teams running real agentic and rollout workloads.

Enterprise — custom

Custom terms, invoiced.

• Unlimited concurrency + session duration
• Dedicated capacity
• SSO, and a bring-your-own-cloud (BYOC) option
• Priority support + SLA

For labs and platforms at fleet scale.

What's metered

Dimension	When it accrues
CPU-core-hours	While a sandbox is running.
RAM-GB-hours	While a sandbox is running.
Disk-GB-hours	While a sandbox is running.
Snapshot-GB-months	While a sandbox is suspended.

A suspended sandbox accrues only snapshot storage — no CPU or RAM. Demo usage is metered for observability but priced at $0. Track your usage-to-date in the console or via GET /v1/usage.

Suspend to control spend

Because suspend stops the meter, long-lived agentic sessions cost only for the time they're actually working. Suspend between bursts and resume in place — you keep the exact live state and stop paying for idle CPU/RAM.

Sign up

• Create an account — start on Demo with no card, upgrade to Pro from the billing page.
• Talk to us — for Enterprise terms, dedicated capacity, self-hosting, BYOC, or an SLA.

Current per-unit rates are published on the Collimate site. The tier limits above are enforced by the API from your key's server-side record — see Authentication and Limits.