Do I pay for AI tokens through ModelRiver?
No. You connect your own provider API keys and each provider bills you directly for the tokens you use. Your ModelRiver plan covers the platform: routing, failover, structured outputs, logs and the rest.
Do I need my own provider accounts?#
Yes. ModelRiver doesn't supply provider keys. Create an account with each provider you want to use, then add its key under Settings → Providers. Keys are stored encrypted and only the last four characters are ever shown.
Why bring your own keys?#
- You keep your provider's prices, discounts and credits.
- Usage and rate limits stay on your own provider account.
- Your agreement with each provider, including its data terms, stays in place.
- You can leave at any time without losing a provider account.
The cost in my request logs doesn't match my provider's bill. Why?#
Costs in Request Logs are estimates, based on the model prices set in your project. Your provider's invoice is the final amount. Use the estimates to compare models and spot expensive workflows, and the provider's bill for accounting.
Can I cap what I spend on AI?#
Yes. Set spend limits per project and per workflow model in US dollars. When a cap is reached, requests stop with a budget error, or move to a cheaper backup model if you've set one up.
Can I try ModelRiver without spending on tokens?#
Yes. Test mode returns sample responses without calling any provider. See Can I try ModelRiver without paying for AI calls?