Frequently Asked Questions

The catalog spans OpenAI, Anthropic, Google, DeepSeek, Alibaba (Qwen), Moonshot (Kimi), Zhipu (GLM), ByteDance (Doubao), xAI (Grok), Tencent (Hunyuan), MiniMax, Meituan (LongCat), StepFun and Xiaomi (MiMo), plus a few image-generation models billed per image. Model IDs and availability change as providers update them, so the live model pricing page is the authoritative list — it is generated from the same billing table that charges your requests. A model that is not on that page cannot be called, and IDs are case-sensitive.

You pay per token used — input and output tokens are billed separately according to each model's published pricing. Image models may use a per-image rate. No monthly fees, no hidden costs. Check the live model pricing page for the current provider-grouped rates. Credits are deducted from your balance in real time as you make API calls.

A Standard key reaches every model in the catalog except jzs-max-3.0 — this is what most customers want. The other type is locked to jzs-max-3.0 alone and runs over its own dedicated route, which holds up better under heavy concurrent load. Pick Standard unless jzs-max-3.0 is the one model you plan to call. Both key types draw on the same account balance, so you can create one of each without splitting your credits.

No. Credits do not expire and there is no monthly commitment — you use them at your own pace. Top-ups are currently offered in fixed amounts starting at $1.50; the exact options are on the Buy Credits page.

We do not impose a per-key request-rate limit. What bounds throughput is your balance and the capacity of the upstream provider serving the model you call — under load, requests queue there and can return 429. The SDKs and clients we document retry 429 automatically. If you are planning a sustained high-volume workload, talk to us first so we can check capacity with you rather than have you discover it in production.

Visit your Dashboard — your current balance is displayed at the top of the page. You can also view your usage history and per-model spending in the Usage section.

Your credit purchases are listed with their order numbers under Orders in the dashboard. For an invoice document (for example, for company reimbursement), contact support with the order number — invoicing is handled manually, so allow for the support hours on the contact page.

jzs-max-3.0 is our flagship model ID, served over a dedicated upstream route with its own capacity rather than sharing the general pool — that is why it holds up better under heavy concurrent load. It is not part of the Standard key's scope: calling it requires a key created specifically for jzs-max-3.0. Its current rates are on the live catalog.

The fastest way to reach us is via WhatsApp. We answer during support hours (Daily, 09:00–22:00); how quickly depends on how busy the queue is, so we do not promise a response time.

We do not store or log your prompts or completions. Requests are handled in memory and the content is discarded once the response is delivered; the only record we keep is billing metadata (token counts, timestamp, model). That describes this platform. To answer a request we must forward its content to the upstream provider of the model you chose, and that provider handles it under its own policy — we do not control and cannot make claims about their retention. See the privacy policy for the full picture.

Yes. The endpoint speaks the OpenAI chat-completions format at https://jzstoken.com/api/v1, so it works as a drop-in base URL for the official OpenAI Python/Node.js SDK and for clients that support custom base URLs. One caveat we would rather state than let you discover: we do not independently certify every optional client feature — streaming, tool calling, structured output, vision — for every model. Check the guide for your tool, which states what has actually been exercised.