Pricing & billing
ApiGo prices every call in USD with separate input and output rates. Top-ups and invoices happen inside the workspace via Stripe.
Pricing rules
Open Pricing for the product explanation. The short version:
- Every model is billed in USD per 1M tokens.
- Input and output are billed separately.
- Cache hits use each model’s discounted cache rate when the model supports caching.
- There are no promo credits that inflate the posted balance — paid amount equals balance posted.
Browse live unit prices on Models.
Stream cancellation and billing
ApiGo currently handles every streaming request as if its upstream does not support stream cancellation. Aborting the client request or closing the connection stops the downstream response, but it does not guarantee that upstream generation stops.
The upstream may continue generating to completion. Your final ApiGo charge follows the settled upstream usage and may therefore include tokens generated after you disconnect. Streaming usage also commonly arrives only in the final chunk, so the usage visible before a disconnect is not a billing boundary.
Set max_tokens, max_output_tokens, or the equivalent parameter for your protocol on every production request. Treat that value as the hard ceiling for generated output and the corresponding output-token charge exposure. Do not rely on aborting a stream as a cost-control mechanism.
Top up
- Open Workspace → Billing.
- Start a top-up.
- Enter amount, billing email, and Tax ID when prompted.
- Complete Stripe Checkout. Balance updates when payment succeeds.
Stripe issues a separate invoice for each new top-up. You can open invoices from Billing and export a non-tax consolidated statement across multiple top-ups when needed.
Balance alerts and empty balance
- Configure balance alerts so email and in-workspace banners fire before you run dry.
- Pair alerts with Routing budget guards / fallback chains for production.
- When available funds are not positive, new calls return
402 insufficient_balance. API keys stay intact and work again after funds are available and the admission snapshot refreshes.
Balance is a Workspace-wide start-of-request gate. API key spend caps and runtime limits are separate checks. See Limits and request enforcement for their order and interaction.
FAQ
What is the smallest billable unit?
Each request is billed pro-rata against the per-1M-token rate, rounded to cents, and posted in real time.
When are invoices issued?
Stripe issues an invoice per top-up under Workspace → Billing.
Do model prices change?
Yes — ApiGo mirrors upstream changes, announces them in-workspace, and keeps at least a 7-day transition window.
