Simple, Transparent Pricing

Each API call to ModelRiver counts as one request, regardless of the AI model or provider used. This includes successful requests, rate-limited requests, and requests that return errors.

Yes, you can upgrade or downgrade your plan at any time. Upgrades take effect immediately, while downgrades take effect at the start of your next billing cycle. You'll be prorated for any mid-cycle upgrades.

Pay-as-you-go requests are billed at the per-request rate shown for your plan. You'll never be cut off mid-month - we'll automatically handle the pay-as-you-go usage and bill you at the end of your billing cycle. You can set up usage alerts in your console.

We support all major AI providers including OpenAI (GPT-5.1, o3, GPT-4.1), Anthropic (Claude Sonnet 4.5, Claude Opus 4.1), Google (Gemini 3, Gemini 2.5 Pro), Cohere (Command A), and more. You can switch between providers seamlessly using our unified API, and we handle automatic failover if a provider is unavailable.

When a request fails due to rate limits, errors, or downtime, ModelRiver automatically retries your request with an alternative provider you've configured. This happens transparently - you don't need to change your code. You can configure failover strategies in your workflow settings.

All plans include automatic failover, rate limiting, webhooks, websockets, structured data validation, analytics console, request logging, and access to all supported AI models. The main difference between plans is the number of requests and level of support.

Webhooks allow you to process AI requests asynchronously - we'll notify your endpoint when the response is ready. Websockets provide real-time streaming updates for long-running requests. Both support automatic retry logic and custom data passing for easy request tracking.

You connect using your own API keys for each provider you want to use. ModelRiver acts as a gateway that routes requests through your configured providers. We support all major AI providers (OpenAI, Anthropic, Google, Cohere) and you can also add custom providers with your own endpoints.

Yes! Every account starts with 10,000 free requests per month to test all features. No credit card required to get started. This lets you fully integrate and test ModelRiver in your development environment before committing to a paid plan.

All subscriptions are billed monthly and are non-refundable except as stated in our refund policy. We recommend starting with the free tier to test the service thoroughly before upgrading.