Zhiqi Engine

Zhiqi Engine is an Enterprise-grade AI Model Aggregation gateway built for developers and SaaS provIDErs. It presents many domestic and overseas model APIs through a single OpenAI-compatible contrACT, aiming to reduce integration work and give teams centralized control over routing, credentials, quotas, and spend. It is a middleware layer rather than a model provider, so ouTPUt quality and upstream availability depend on the vendors behind it.

What it does

The platform sits between an APPlication and multiple model vendors. A request uses a unified endpoint and key; the gateway authenticates it, checks quotas, applies routing rules, forwards the request to the selected provider, logs token usage, latency, and status, then returns the result. The service describes support for text models, Image Generation, speech, OCR, and video interfaces, all exposed through an openai-style API surface so existing SDK-based code can often switch providers with minimal changes.

Key capabilities

• Unified model access: Aggregates hundreds of model and multimodal endpoints under one protocol and one key-management layer.
• Intelligent routing and failover: Supports weighted traffic distribution, scene-based model selection, priority policies for lower cost or higher stability, and automatic fallback when an upstream model times out or returns errors.
• Key and sub-account governance: Lets teams create project or developer subkeys under a master key, with per-key QPS liMITs and daily quotas to isolate usage and prevent overspend.
• Monitoring and billing: Tracks call counts, Token Consumption, success rates, and response latency, with bills split by subkey, model, or project and alerts for anomalies.
• Rate limits and cost controls: Provides custom frequency caps, daily quotas, spending warnings, and routing strategies that favor cost-effective models.
• Logging and traceability: Stores full-chain request logs with inputs and outputs to support debugging, reconciliation, and incident review.

Who it is for

It is aimed at independent developers, AI Startups, SaaS providers, and technical leads who need to integrate multiple large models and operate commercial AI Services with unified key management, traffic control, cost visibility, and stability safeguards. It is not a consumer-facing chat or image-generation interface; nontechnical users looking for a web app to talk to models or create images would need a different product.

How to get started

1. Register for the developer console and obtain a platfORM master key.
2. Enable the upstream models you want to use and configure routing and fallback policies.
3. Create project subkeys with QPS and quota limits, then assign them to services or team members.
4. Point application code at the unified endpoint using a subkey, and review usage, billing, and error logs in the dashboard.

Pricing and access caveat

Pricing is usage-based, with token or task billing that varies by model. The service mentions prepaid balances, test credits for development, and possible enterprise packages or high-concurrency channels; exact plans, rates, and availability may vary and should be confirmed with the provider. Because Zhiqi Engine is a transit and scheduling layer, it adds an extra network hop, sensitive business data passes through the gateway, and upstream rate limits, price changes, or model retirements can affect service. Teams with privacy or compliance requirements should evaluate data handling before production use.