Set spend limits
Control usage and budget exposure for agent work.
Proliferate tracks two separate budgets: compute units for cloud workspace runtime, and LLM credits for model calls made through Proliferate's gateway. Both have a preview in Settings > Organization > Budgets, though right now that whole pane is a mocked preview of what's coming, not a live usage dashboard (more on that below).
How LLM spend is capped today
If you use Proliferate's built-in gateway (see Use your AI gateway), spend is capped at the gateway itself, not by a setting you tune per project:
- Every account gets a one-time free LLM credit at signup.
- Organizations on a paid plan get a shared team budget on top of that.
- The gateway enforces the budget as a hard stop: once it's spent, further gateway-routed calls fail closed until more credit is added.
- Organizations with overage enabled on their plan can let the team keep going past the included budget, funded through billing instead of a hard cap.
If you bring your own API keys or your own gateway instead, spend is capped wherever that provider or gateway already enforces it, Proliferate doesn't add a second limit on top.
Settings > Organization > Budgets
The Budgets pane: balance cards for Compute units and LLM credits, a usage-over-time chart, a per-source usage breakdown, a per-person usage table, and an Enterprise 'Monthly LLM budgets' section. Every one of these sections carries a visible 'Mocked UI' badge in the app today, and all of the save/top-up controls in the pane are disabled.
What the Budgets pane shows today
The whole Budgets pane is a mocked preview right now, not a live-tracking view. The balance cards, the usage-over-time chart, the per-source breakdown, the per-person usage table, and the per-member "Monthly LLM budgets" section are all placeholder numbers labeled "Mocked UI" in the product, and none of their controls (save, top-up, alert thresholds) do anything yet. It previews the shape of budget tracking that's coming; it isn't reporting what your organization has actually spent.
If your team needs real spend controls right now, use the org agent policy to limit which agents and routes are allowed, or route through your own AI gateway where you already have budget controls.
Your organization's actual included compute and LLM budget, overage setting, and auto top-up are real and live, separate from this preview pane. Manage them from Billing.