Your spend dashboard

1. Per-model spending

The Usage page aggregates your LLM costs by model. Every call to Claude, GPT-4, Gemini, or any other model records its cost—sum them up and you see exactly where your wallet is going. Compare Basic, Pro, and Frontier usage to find cost optimization opportunities.

2. Where to find it

Open your dashboard and look for the Usage tab. You'll see a graph of cumulative spend and a breakdown by model. To top up your wallet and set Frontier access, visit /docs/topup.

3. Web search & voice

The model price isn't the only thing that draws on your wallet. When you turn them on, web search and voice each bill their own upstream usage—charged to the same wallet, with the same platform markup and welcome-credit offset, and shown on the same Usage dashboard alongside your model spend.

• Web search — billed per search the agent runs (web-provider cost).
• Voice input — speech-to-text, billed per second of audio you dictate.
• Voice output — text-to-speech, billed per ~1,000 characters spoken back.

These only cost anything when you use them: leave web search Off and skip voice, and you pay model tokens alone. Exact rates can change—the Usage dashboard always reflects what was actually charged.