First identify whether you are buying a subscription, receiving an account, or calling a model API. These products have different deliverables, billing, and troubleshooting paths, even when they share a model brand.
XiuStore purchases and delivery
AI subscription
Membership benefits for a website or application such as ChatGPT or Claude. A subscription is not API credit. Treat API credit as included only when the product page and order explicitly say so.
Activation on your own account
The subscription is activated on an account you already use. Existing conversations and account identity usually remain, but the account must be accessible and meet the order and platform’s current requirements.
Ready-to-use account
The order delivers a new account and sign-in instructions. It does not move the subscription to your existing account. On the delivery day, check the identity, plan, term, and which security details the order permits you to change.
One-time service
SMS verification or identity-verification assistance delivers one service, without an ongoing subscription or a guarantee of third-party approval, verification, or account status.
Before buying, read Choose a product and delivery method and the AI subscription buying guide.
XiuRouter model APIs
Model API
An HTTP interface that applications use to call a model. It is separate from web memberships such as ChatGPT Plus or Claude Pro and is generally billed by actual request usage.
API key
A credential used to call XiuRouter. It can restrict models, service groups, quota, and expiration. It is not a XiuAI sign-in token. Keep it out of repositories, screenshots, chats, and public configurations.
Model ID
The exact model name in a request. Copy it from Models and pricing or the target key’s /v1/models response rather than constructing it from a display name.
Token
A unit of text processing. Input and output are counted separately. Tokenization varies by model, so character counts do not convert to a fixed token count.
Context
The input, conversation history, tool results, and reserved output a model can process together in one request. Limits depend on the model and protocol; there is no single limit shared by every XiuRouter model.
Cache reads and writes
Some models record cache-write usage and subsequent cache-read usage separately. These dimensions are billed only when a request actually produces them. Check current rates on the pricing page and actual charges in usage logs.
Protocol compatibility
A client’s request format can connect to a gateway endpoint, such as Chat Completions, Responses, Anthropic Messages, or Gemini GenerateContent. Compatibility does not migrate every provider-hosted tool, file, feature, or cloud session.
Service tier and service group
The pricing page uses service tiers, while API keys and usage logs use service groups. They refer to the same group value used to select a model source and price. A group name does not itself guarantee speed, quality, availability, or an SLA.
Usage logs
Evidence of requests that have occurred. Use them to reconcile the key, model, service group, status, tokens, and cost. The pricing page describes current rates; usage logs show actual results.
Continue with API compatibility, models, pricing, and usage, and errors and production checks.
Sources and review date
Purchase options, protocol endpoints, pricing, and usage fields were reviewed on 2026-08-30. This translation preserves that review date. Models, prices, plans, and third-party features change; the current product page takes precedence.