Balance, Billing & Usage
QuotaAPI records usage based on the requested model, input and output tokens, pricing multipliers and account balance. The console shows balance changes, request records and error details.
Balance
Balance pays for API calls. Before a request is served, QuotaAPI checks account state and available balance. If balance is insufficient, the request is rejected.
Common steps:
- Open the console and check current balance.
- Review recent request records for unexpected usage.
- Recharge if the balance is insufficient.
- If calls still fail after recharge, refresh the console and retry a short request.
Top-up methods
The console recharge page supports online payments and cryptocurrency. Available channels are whatever the recharge page currently shows.
Cryptocurrency top-up
Supported: USDT (TRC20 / ERC20), USDC, BTC, ETH. Top-ups are denominated in USD; the system generates a dedicated deposit address and an exact amount to pay. Please note:
- Pay the exact amount shown: underpayments are not credited automatically — they go to manual review and require contacting an administrator.
- The deposit address has an expiry. Payments made after the order expires are not credited automatically; start a new top-up instead.
- Make sure the transfer network matches what the page shows (e.g. TRC20 vs ERC20). Sending on the wrong network may lose funds.
- On-chain confirmation takes time (longer for BTC/ETH). Keep the page open after paying; it redirects automatically once confirmed.
- For underpaid or late payments, contact an administrator with your order number for manual verification.
Billing
Billing is usually affected by:
| Factor | Notes |
|---|---|
| Model | Different models can have different prices |
| Input tokens | User messages, system prompts and context can count |
| Output tokens | Model responses count toward usage |
| Image count and size | Image generation and edits are recorded by image count and size tier |
| Multiplier | The console may configure different multipliers per model |
| Streaming | Streaming and non-streaming requests are billed by final usage |
The console is the source of truth for current pricing.
Usage records
After a request completes, the console usually records:
- API Key or calling account.
- Target model and group.
- Request time, latency and status code.
- Input, output and total tokens.
- Image count, image size tier or image-related output details.
- Charge amount.
- Error message or failure reason.
These records help distinguish authentication, permission, balance, parameter and model-service errors.
Cost control
- Create separate API Keys for different projects.
- Set reasonable output token limits in server-side clients.
- Limit retry count when enabling client retries.
- Add queues or throttling for high-concurrency tasks.
- Review usage records regularly and disable abnormal keys.
Insufficient balance behavior
When balance is insufficient, OpenAI-compatible API, Anthropic Messages and Gemini v1beta return a clear error. Different clients may display it differently, but the console request records keep the failure reason.
If a client only shows a generic error, reproduce the same model and key with curl to see the original status code and message.