Plans & usage
Usage metering and billing behavior for Proliferate Cloud.
Local Proliferate is always free. The desktop app, local checkouts, and worktrees never touch billing, on any plan. Cloud workspaces are different: they run on real, metered compute and can call real models through Proliferate's managed gateway, so cloud usage is tracked against your plan.
Bringing your own agent subscription (Claude Code, Codex, OpenCode, and so on) and running it locally costs nothing beyond what you already pay that provider. Billing only applies to Proliferate's managed cloud sandboxes and managed model access.
Plans
Proliferate Cloud has three plans.
- Free: $0, forever. Personal use and evaluation: bring your own agents, run one cloud sandbox at a time, and use Proliferate's hosted model gateway with a starter credit.
- Core: an organization plan billed monthly. Adds organization-wide cloud billing, unlimited workflows and team members, shared role-based access, and beta features.
- Enterprise: custom pricing and usage caps. Adds SSO/SAML, org-wide secrets and audit trails, bringing your own model provider keys, and private VPC or custom instance types.
Manage your plan from Settings → Billing. Personal accounts see their own usage; organization owners and admins can also manage the organization's plan and billing details from there.
Settings → Billing
The Billing settings pane: current plan badge and Manage button, compute and LLM credit usage cards, the auto top-up toggle, and a link to the Stripe billing portal.
Two credit pools
Cloud usage is metered on two separate ledgers instead of one blended number:
- Compute credits pay for sandbox runtime: the machine your cloud workspace actually runs on. Usage is tracked in Proliferate Credit Units (PCUs), which track sandbox hours.
- LLM credits pay for calls made through Proliferate's managed model gateway, the built-in way to use models like Claude and GPT without bringing your own API key. Every new account gets a one-time free credit (currently $5) to try it. If you'd rather use your own model subscription or API key, you can, and it draws down nothing from this pool.
Keeping the two separate means a model-heavy chat session and a long build don't compete for the same budget, and switching to your own model keys (available on Enterprise) stops LLM credit usage without touching your compute balance.
Choosing or changing your plan
Click Manage on the Billing pane to open the plan dialog. From there you can start a Core subscription through Stripe checkout, or reach out for an Enterprise trial.
Settings → Billing → Choose your plan dialog
The plan management dialog comparing Core and Enterprise: included compute units and LLM credits per month, plan features, a coupon code field, and the link to the Stripe billing portal.
Upgrading or downgrading doesn't interrupt work in progress: cloud sandboxes that are already running keep running under the plan they started on. New limits and included credits apply going forward, from your next sandbox start or the start of your next billing period.
Buying more
If you're going to need more than your plan includes, turn on top up from the Billing pane (organization owners and admins only). With it enabled, going over your included compute or LLM credits bills the overage automatically instead of pausing your work, up to a cap you set. See Billing and limits for exactly how the cap and the pause behavior work.
Where to see exact numbers
Plan pricing and included credits are easiest to get wrong in writing and easiest to get right in the product. Your current plan's exact compute credit balance always shows on the Billing pane and in the plan dialog, so check there before relying on a specific number.
The LLM credits balance on the Billing pane is still placeholder data while usage rollups for the model gateway are being wired up. Treat any LLM credit number you see there as illustrative, not a real balance, until this note is removed.