Pay for what you ship.

What counts as an LLM event?

One wrapped call into Anthropic / OpenAI / Gemini through the SDK. Streaming and tool-use round-trips count as a single event each. Cached cache-reads count as events but at much lower cost (and Cost surfaces the savings).

What happens if I go over my plan's event cap?

Soft cap. You won't be cut off mid-month. We send an alert at 80% and again at 100%, and ask you to upgrade or accept overage billing (currently no overage charges - launch grace).

Can I switch plans?

Yes - from the Customer Portal (signed in, click Manage subscription on your account). Plan changes prorate against the current cycle, so upgrading mid-month charges only the difference.

Do you store my prompts?

By default, no. The SDK sends metadata only (model, route, hashed user id, token counts, latency, cost, plus a SHA-256 hash of the prompt). Raw request and response bodies are sent only when you explicitly pass captureBodies: true to wrap() for a route, which is what powers verify-downgrade. When opted in, raw bodies are retained for 30 days on the free tier and 90 days on paid plans, then purged automatically. See the privacy page for the full data flow.

Cancel anytime?

Yes. Cancellations take effect at the end of your billing cycle - you keep access through what you've already paid for.

All SDKs are MIT-licensed and public. github.com/botzone-ai