TokenSense
Start Free Trial

Budgets

TokenSense enforces budget caps server-side. When a budget is exhausted, TokenSense returns a clean 402 to your tool — not a surprise on next month's bill. Think of budgets as spending insurance for your AI workflows.

Three levels of budget

  • Workspace budget — total spend cap across all projects and workflows (all tiers)
  • Per-project budget — spend cap for a single project/client (Pro and Agency tiers)
  • Per-key budget — spend cap for a single API key (Pro and Agency tiers)

Tier matrix

Budget typeStarterProAgency
Workspace budget✓✓✓
Per-project budget—✓✓
Per-key budget—✓✓

See plans and pricing for full tier comparison including request limits and team seats.

How enforcement works

TokenSense checks budget status on every request. When a budget approaches its cap, TokenSense can warn before it's reached. Once the configured cap is exhausted, subsequent requests receive a 402 response — your workflow stops cleanly rather than running up an unexpected bill. Exact thresholds and grace behavior depend on your workspace settings, plan tier, and budget type.

Budget alerts

Configure alerts to get notified before a budget is fully exhausted. Set thresholds (e.g., 80% of budget) from the Alerts page in your dashboard. Alerts fire via email — you choose which budgets to watch.

Subscription quotas vs workflow budgets

These are two distinct mechanisms:

  • Subscription quotas — workspace-level request count per billing cycle (e.g., 10,000 requests/month on Starter). Enforced at the subscription layer.
  • Workflow budgets — per-key or per-project spend caps in dollars. Enforced at the proxy layer. These are what you configure to control individual workflow or client spend.

A request can be blocked by either mechanism: quota exhaustion (you've used all your plan's requests) or budget exhaustion (the specific workflow/project has hit its spend cap).

Related