Ready to get started?
Deploy a fully functional AI Gateway locally in just 5 minutes and experience multi-model proxy, Token rate limiting, and more.
Product / AI Gateway
AI Gateway
Product Introduction
Designed for AI scenarios, unified proxy for LLM APIs and MCP Servers, supporting centralized access and management of models, tools (MCP), and Agents.
Star on GitHub Supports 100+ LLM models
Product Advantages
Provides multi-model unified routing and protocol conversion, semantic caching, Token-level rate limiting, content security protection (such as prompt injection detection, sensitive content recognition), multi-dimensional authentication and observability, and supports hot-pluggable plugins and intelligent tool routing.
Unified Multi-Model Access
Provides a unified gateway entry for OpenAI-compatible interfaces and mainstream LLM services, shielding authentication and interface differences across vendors, facilitating rapid switching and multi-active deployment. Also supports model load balancing and Fallback to improve model service reliability.
Unified MCP Management
Supports HTTP-to-MCP protocol conversion and proxying of native MCP services, converging scattered tool capabilities into a unified entry point, reducing access and maintenance costs across multiple teams and environments.
Token Traffic Management
In addition to traditional QPS rate limiting, provides quota management and rate limiting capabilities based on Token consumption. Combined with API Key pool rotation and consumer authentication mechanisms, it effectively prevents model overload and controls client call quotas.
Scalability
Quickly integrate AI ecosystem components such as RAG, vector databases, log observability, and caching via plugins for on-demand capability combination; also supports custom plugin development to meet complex business logic scenarios.
Higress supports unified access and governance for 100+ mainstream large models
Higress vs Other AI Gateways
Comprehensive comparison from architecture, features to enterprise capabilities
查看完整对比分析
深入了解 Higress 相比其他 AI 网关的技术优势和企业级特性
Choose the Right Version for You
Community edition is free, Enterprise offers fully managed service with commercial support
Community
Higress Community
Local deployment, verify core capabilities in 5 minutes
- Full AI Gateway core capabilities
- Community support & documentation
- Self-managed stability & performance
开源免费，社区支持
Free Trial
Higress Enterprise
Fully managed cloud service, production-grade guarantee
- Fully managed, 99.95% SLA
- Software-hardware integrated performance optimization 90%+
- Multi-AZ deployment, auto failover
- Deep integration with Alibaba Cloud
Ticket, DingTalk and other commercial support
Premium Service
Feitian Exclusive Edition
Match business scale with ultimate SLA guarantee
- Performance & Resources: Match business scale with elastic scalability
- SLA: Negotiable on demand with ultimate guarantee
- Out-of-the-box advantages in stability, performance, cloud product integration, and usability over community edition
- Unified proxy for AI model API and MCP Server with rich integration and governance capabilities
Ticket, DingTalk & offline commercial support
Learn about Enterprise Edition
Looking for deeply integrated API management, traffic gateway, and microservice gateway capabilities? The Enterprise Edition offers more powerful unified governance and security extensions. Learn more.