Skip to content

Spendware

Spendware is usage-based infrastructure for AI applications. It gives product and platform teams a control plane for provider usage, policy checks, cost attribution, billing workflows, guardrails, and outbound event delivery.

Developers usually integrate Spendware in one of two ways:

  • Wrap an existing AI provider client with the Python or TypeScript SDK. The wrapped client keeps the provider's normal interface while Spendware performs preflight checks, extracts usage, and logs activity.
  • Call Spendware APIs directly from a backend service. This is useful for management automation, direct usage ingestion, billing workflows, and webhooks administration.

What Spendware Does

  • Organizes usage by organizations, projects, and customers.
  • Issues scoped user API keys for management automation.
  • Issues scoped customer keys for runtime preflight and usage logging.
  • Runs preflight policy checks before provider calls when enabled.
  • Logs raw token and request usage from SDKs or direct API calls.
  • Tracks usage analytics for finalized and eligible partial usage.
  • Connects usage to pricing rules, billing cycles, and invoice previews.
  • Enforces guardrails for usage or cost thresholds.
  • Emits signed outbound webhooks for integration events.

Service Model

Spendware provides a dashboard, backend API, webhooks API, Python SDK, and TypeScript SDK. Developers integrating Spendware mostly interact with the APIs and SDKs from their own backend services.