Product Overview
ZStack API Router is an enterprise AI model gateway. It provides unified model access, channel routing, token management, usage metering, audit tracking, and health observability. Administrators can configure providers, channels, models, Route Policies, organizations, permissions, and quotas from the console or control-plane APIs. Developers use OpenAI-compatible APIs to call models.
The typical usage flow is:
- Initialize the gateway and verify service readiness.
- Create organizations, users, and model access scopes.
- Create tokens.
- Configure channels and synchronize models.
- Enable models and configure Route Policies.
- Call models with an API Key.
- Review Usage Log records, Audit Log records, health status, and metrics.
- Troubleshoot requests from logs and metrics.
Main Objects
| Object | Description |
|---|---|
| User | A person who uses the console or calls models. |
| Organization | A management boundary for cost, model access, and usage ownership. |
| Token | Management object for call credentials. After a token is created, the system generates API Key plaintext. Applications use the API Key to call models. |
| Channel | Upstream access path to a model service. |
| Model | Model name used by applications. |
| Route Policy | Routing rule that selects channels for model requests. |
| Usage Log | Record of model requests, usage, status, and latency. |
| Audit Log | Record of administrative operations and resource changes. |
Note:
After a token is created, the system generates API Key plaintext. The API Key plaintext is displayed only once and should be delivered only to the caller. For troubleshooting, use the token ID, token name, user, and organization information. Do not paste real secrets into tickets, screenshots, or logs.