Cloudsome aggregates models from multiple AI providers through the documented interfaces. Available providers and models change over time — use the Models page or GET /v1/models for the current list.

1. Create an account 2. Make sure the current workspace has sufficient balance and top up if needed 3. Generate an API key in the Dashboard 4. Choose an available model and configure a documented endpoint 5. Start making requests See our Docs for the complete Quick Start guide.

If a request fails before the response begins because of validation, rate limiting, insufficient balance, or an upstream error, you are not charged and the pre-deducted amount is fully refunded. Errors after the response begins are billed for usage already generated. See our Billing Rules page for details.

If an error occurs before streaming begins, the pre-deducted amount is fully refunded. If you disconnect or an upstream error occurs after streaming begins, billing uses actual usage when available; otherwise, usage is estimated from the generated content.

When you make a request, Cloudsome routes it to the best available provider. If that provider fails (timeout, 5xx error, rate limit), the request is automatically retried on a fallback provider — transparently, with no extra charge for the failed attempt. You only pay for the successful response.

Cloudsome supports common OpenAI SDK workflows for the endpoints listed in our documentation. With the official OpenAI Python or Node.js SDK, you can usually get started by changing the base URL and API key. Check the API documentation for support in other tools.

Your Dashboard provides real-time visibility into: - Current balance and spending trends - Per-model usage breakdown (tokens, cost, requests) - Individual request logs with latency, status, and cost - Date-range filtering for billing periods

You can reach us through our support channels. We also have a Docs page with detailed API reference, error codes, and rate limiting information. For billing questions, see the Billing Rules page.