Pricing
Flat pricing, zero token markup
You pay TokenRouter a flat subscription. Your providers bill your tokens at their exact rates — we never take a percentage. Compare every plan side-by-side.
Scale
How much gateway each plan includes
Gateway
One endpoint in front of every provider
Spend controls
Enforced at the gateway, before the invoice
Visibility
Metadata only — prompts are never stored
Support & billing
How we back you up
No surprise bills, ever
Budgets are hard-capped at the gateway. When a team hits its cap, requests stop — a looping agent can never turn into a five-figure invoice.
Zero token markup
Your providers bill your tokens at their exact rates, on your own keys. TokenRouter never takes a percentage of your traffic.
Cancel in two clicks
Self-serve from the billing portal, no email required. Your plan stays active until the end of the period you already paid for.
FAQ
Pricing, in plain English
Is there a free plan?
No. Every plan starts with a 14-day free trial instead — full product, card required, never charged if you cancel before day 14. We'd rather give you the real thing for two weeks than a hobbled free tier forever.
What happens on day 14 of the trial?
Your card is charged for the plan you picked at signup and your subscription starts. Nothing changes about your gateway, keys, or budgets — traffic keeps flowing. If you cancel before day 14, you are never charged.
Can I cancel anytime?
Yes. You can cancel from the billing portal in the console at any time. Your plan stays active until the end of the current billing period and you won't be charged again.
What counts as a request?
One API call to the gateway — a chat completion, a response, a message, or an embedding call. Streaming counts as a single request regardless of how many chunks are streamed. Failed requests that never reach a provider don't count against your quota.
Do you proxy my provider keys securely?
Yes. Provider keys are encrypted with AES at rest, decrypted only in memory at request time, and never written to logs. Prompts and completions are processed transiently and never stored — we keep only token counts, cost, and latency metadata.
Wondering how this stacks up against a usage-based gateway? See the OpenRouter comparison