XuanShu API
Make multi-model access more resilient, transparent, and manageable
XuanShu API is a new-generation AI model routing gateway for enterprises, technical teams, and individual developers, powered by an intelligent routing and fault-tolerance mechanism further developed from the Sub2API open-source project.
Unified access · Intelligent routing · Usage governance
Service positioning
XuanShu API is a new-generation AI model routing gateway for enterprises, technical teams, and individual developers. A unified entry point connects leading models and AI coding tools while reducing multi-provider integration and switching costs.
The platform continuously improves intelligent routing and service availability, providing stable, transparent, and manageable model infrastructure for development, teamwork, and product integration.
Who we serve
Enterprises
Integrate multiple AI models into products, operational workflows, and internal systems through one dependable service.
Technical teams
Reduce provider-specific integration work and collaborate through a consistent path for development, testing, and delivery.
Individual developers
Explore different models and AI coding tools conveniently, then turn experiments into working applications.
Technical boundaries
XuanShu API provides protocol compatibility and unified management without changing the underlying limits of third-party models, networks, or clients. Available models, context windows, tool use, streaming behavior, and other parameters follow the current console configuration and corresponding documentation.
Tool, model, and protocol names describe compatible access paths only. Users remain responsible for evaluating each use case against their own business, security, and compliance requirements.
How it works
The XuanShu operations team develops and operates XuanShu API on the Sub2API open-source foundation, with an intelligent routing and fault-tolerance mechanism further developed from the Sub2API open-source project. Model matching and account and channel health filtering select a suitable available path for each request.
When the request and current routing state permit, the service maintains session affinity; after a retryable error, it can switch to another available candidate. These mechanisms reduce the impact of localized failures but do not guarantee that every model or request will always succeed.
Support
Contact the XuanShu operations team through the channels below for general inquiries, technical questions, and account support. For enterprise onboarding, team usage governance, or invoicing, include your intended scenario so the team can respond efficiently.
Scan with your phone to add
Brand and open-source foundation
XuanShu API is independently operated by the XuanShu operations team and is further developed from an open-source project.