Billing & top-up
The service is pay-as-you-go. Site credits use the ¥1 paid = $1 credited numerical convention; this is a platform credit unit, not a foreign-exchange conversion rate.
Top-up availability
Top-up payments are enabled. Sign in and check the Recharge page for the payment methods currently offered, then follow the on-page instructions. Available methods and the status of your credited balance are shown there; do not assume a particular method or instant crediting. If no option appears or an order remains pending, contact support. Never send money to an off-site address or pay the same order twice. You may provide an order number for support, but never your API key.
Prices and final charges
The Model Plaza displays model/group prices. Check the unit (per-token, per-image, per-second, etc.) and the group attached to your API key; a displayed model is not necessarily available to every key. Actual charges depend on request usage, the effective group and billing mode. Confirm each charge in Usage; the dashboard shows balance and aggregates. If a quote and charge disagree, save the time, model and request ID for support.
When your balance is insufficient, new calls may fail. Check the dashboard and use the current Recharge options or contact support.
Understanding token charges
Text requests generally account separately for input, output and applicable cache usage, then apply the effective pricing rules. A price per one million tokens is not a per-request price. Tokens are not the same as characters.
| Item | Meaning |
|---|---|
| Input | Prompt, conversation history and other submitted input |
| Output | Generated output usage, as reported for the model and request |
| Cache read | Reused context; a cache hit may still incur a cache-read charge |
| Cache write | Establishing cached context; support and price depend on model/protocol |
Do not count all input and then add cache tokens again without checking whether the protocol includes them in input. Usage fields have different semantics across protocols. An absent cache field is not proof of free usage.
Groups, tiers and peak periods
- When Plaza details label prices as final charges including the group multiplier, do not apply that same multiplier again.
- A card labeled base-tier/off-peak reference is not a universal final quote. Open details for thresholds, schedules, timezone and tier prices.
- A context tier may apply to the entire request rather than only tokens above the threshold. Check the model's current rules and Usage. Conversation history and tool results increase subsequent input.
- Request-level options such as Fast may change charges or compatibility. Do not infer their cost from the base price; disable unnecessary extensions until the applicable rules are confirmed.
Images and video
For per-image billing, check image count and the applicable tier, not text-token rates. For per-second video, check duration, resolution and other pricing conditions. Rules for one media model do not automatically apply to another.
A listed price does not prove that a client supports the model's media protocol. Confirm endpoints and parameters before attempting repeated generation tests.
Reconcile a charge
- Locate the request in Usage by time, model and request ID when available.
- Compare actual usage, billing mode and any displayed group/multiplier fields. Do not invent missing fields.
- Check the rules applicable to that request. Current Plaza prices may differ from prices at the time of an older request.
- For discrepancies, provide redacted records as described in Troubleshooting.
Timeouts and broken streams do not prove that no charge occurred. Check records before retrying; records may be delayed. This page does not promise that all failed requests are free or automatically refunded.
Help
- Email: admin@nekocloud.vip
- Telegram: https://t.me/dnalec