</>MCP Agents Market
MCP Server

higress

by higress-group9.2kGoUpdated 2026-08-25

🤖 AI Gateway | AI Native API Gateway

Claude Desktop

Higress is an AI-native API gateway built on Istio and Envoy that functions as an MCP server platform, enabling AI agents to access tools and services through the Model Context Protocol. Originally developed at Alibaba and now a CNCF sandbox project, it provides unified management for both LLM APIs and MCP APIs with built-in authentication, rate limiting, and observability. Developers can convert OpenAPI specifications into remote MCP servers using the included openapi-to-mcp tool and host them through Higress's plugin system. The platform offers a web console at mcp.higress.ai for immediate experimentation with hosted MCP servers.

Key Features

Hosts MCP servers through a plugin mechanism with unified authentication and authorization
Converts OpenAPI specifications to remote MCP servers via openapi-to-mcp tool
Supports all mainstream LLM model providers with unified protocol and AI observability
Provides fine-grained rate limiting, comprehensive audit logs, and rich monitoring for MCP tool calls
Extends functionality with Wasm plugins written in Go, Rust, or JavaScript
Includes out-of-the-box web console for managing both LLM and MCP APIs
Enables dynamic updates without connection disruption or traffic jitter
Offers true streaming processing for SSE and other streaming protocols with reduced memory overhead

Use Cases

  • 01Hosting remote MCP servers that AI agents can call to access external tools and APIs
  • 02Converting existing REST APIs documented with OpenAPI into MCP-compatible tool endpoints
  • 03Building AI gateway infrastructure with unified access to multiple LLM providers
  • 04Managing authentication, rate limiting, and audit logs for AI agent tool invocations
  • 05Deploying production-grade MCP endpoints with comprehensive observability and monitoring
  • 06Creating secure AI agent environments with WAF protection and fine-grained access control

Related MCP Servers

View more

higress — FAQ

What is the Higress MCP server?+

Higress is an AI-native API gateway that hosts MCP (Model Context Protocol) servers through its plugin system, enabling AI agents to access various tools and services. It provides unified management, authentication, rate limiting, and observability for MCP APIs alongside LLM API management.

How do I install and run Higress as an MCP server?+

You can start Higress locally using Docker with a single command that exposes the UI console on port 8001 and gateway endpoints on ports 8080/8443. For Kubernetes deployments, install via Helm chart. Configuration files are stored in the working directory where you launch the container.

Which AI clients work with Higress MCP servers?+

Higress hosts remote MCP servers that are accessible to any AI client supporting the Model Context Protocol. Clients connect to the hosted MCP endpoints through the gateway's HTTP/HTTPS ports with configured authentication.

Do I need API keys or prerequisites to use Higress?+

For basic local deployment you only need Docker installed. To connect to LLM providers or convert OpenAPI specs to MCP servers, you'll need API keys for those specific services. Kubernetes deployment requires a K8s cluster and Helm.

Is Higress free and open source?+

Yes, Higress is free and open source under the Apache 2.0 license. It's a CNCF sandbox project developed as a vendor-neutral solution, originally created at Alibaba and now community-maintained.

How do I convert an OpenAPI specification to an MCP server with Higress?+

Use the openapi-to-mcp tool included with Higress to convert OpenAPI specifications into remote MCP servers. Once converted, host them through Higress's plugin mechanism for unified authentication, rate limiting, and monitoring.

How do I install higress?+

Open the source repository on GitHub and follow its README. higress is a mcp server — MCP Agents Market links you directly to the official repo.

Is higress free?+

higress is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches