Higress open source Remote MCP Server hosting solution, and will launch the MCP market

Introduction

The implementation of the MCP Server faces numerous challenges, particularly regarding authentication, authorization, service reliability, and observability. Higress, as an AI-native API gateway, offers a complete open-source MCP Server hosting solution. The upcoming MCP market will significantly reduce the time and manpower costs for developers building MCP Servers.

1. From De Facto Standards to Industry Standards

MCP (Model Context Protocol) is rapidly evolving into a thriving ecosystem. Recently, two key events have shifted MCP from a de facto standard to an industry standard: on one hand, OpenAI officially announced its intention to adopt Anthropic's MCP protocol; on the other hand, Anthropic released a new version of the MCP protocol, which brings significant improvements in the Remote MCP Server context.

OpenAI CEO Sam Altman confirmed on social media that OpenAI will integrate Anthropic's MCP protocol into its products, including the ChatGPT desktop application. Altman stated, "The market response to MCP has been very positive, and we are delighted to support this protocol in our own products. This protocol is already available in the Agents SDK, and support for the ChatGPT desktop application and the Responses API will be released soon!" This move means developers can more easily build smart assistants capable of accessing real-time data, like enterprise-level chatbots or automated workflows.

At the same time, Anthropic released a new revision of the MCP protocol on March 26, 2025. The key highlight of this update is the Streamable HTTP transmission mechanism, which not only inherits the real-time message transmission capabilities of HTTP+SSE but also optimizes stability and data transmission resilience. The new protocol supports both HTTP POST and GET requests through a single MCP endpoint, enforces session management using the Mcp-Session-Id header, and supports batching requests, responses, notifications, as well as the resiliency of SSE streams.

The combined effect of these two events is accelerating the flourishing development of the MCP ecosystem. As of February 2025, more than 1,000 community-built MCP Servers are already available. With the involvement of OpenAI, the MCP protocol is expected to become the universal standard connecting AI models to the external world, similar to the status of USB, HTTP, or ODBC in their respective fields.

2. MCP Server Application Scenarios

MCP Server is redefining the capabilities of AI applications, transforming AI from a simple conversational machine into a smart assistant capable of understanding user needs and assisting in completing real tasks. Here are typical application scenarios of the MCP Server in different fields:

Smart Office Scenario

In the corporate office environment, MCP Server can connect various internal systems such as email servers, calendars, and document management systems, enabling AI assistants to:

● Meeting Management: Automatically record meeting content, generate meeting minutes, and create to-do lists based on discussions.

● Email Processing: Classify important emails, draft replies, and set reminders.

● Document Collaboration: Search for information in team documents, provide editing suggestions, and track changes.

For example, a manager might ask the AI assistant, "Summarize the key points from all sales meetings last week and create an action item list." The AI assistant accesses the meeting records and project management tools via the MCP Server to complete this task automatically.

IoT Integration

In smart home and industrial IoT environments, the MCP Server can connect various smart devices and sensors to achieve:

● Device Status Monitoring: Real-time tracking of device operation and predicting maintenance needs.

● Cross-Device Collaboration: Coordinating multiple devices to accomplish complex tasks.

● Environmental Optimization: Automatically adjusting environmental parameters based on user preferences and usage patterns.

Users can say, "When I wake up at 7 AM tomorrow, turn on the coffee machine 20 minutes in advance and adjust the living room temperature to 22 degrees." The AI assistant communicates with the smart home system via the MCP Server to schedule these tasks.

Developer Tool Integration

Software development teams can use the MCP Server to connect code repositories, CI/CD pipelines, and project management tools to enhance development efficiency:

● Code Assistance: Generate code based on project context and provide refactoring suggestions.

● Automated Testing: Generate test cases, execute tests, and analyze results.

● Project Management: Track task progress, allocate resources, and predict delivery timelines.

Developers can request, "Analyze our codebase to find all unhandled exceptions and provide fix suggestions." The AI assistant accesses the code repository via the MCP Server to perform static analysis and generate a detailed report.

3. Local MCP Server vs Remote MCP Server

With the development of the MCP ecosystem, the evolution from Local MCP Server to Remote MCP Server has become a trend. To better understand this transition, we first need to clarify the concepts and distinctions between these two types of MCP Servers.

Concept of Local MCP Server and Remote MCP Server

Local MCP Server: Refers to MCP servers running on the user's local device. In this mode, MCP clients (such as Claude Desktop or Cursor) interact with the MCP server via local process communication (stdin/stdout), while the server connects to various APIs and services over the internet. This architecture is straightforward and suitable for individual developers but has some limitations.

Remote MCP Server: Refers to MCP servers deployed in the cloud that users can access over the internet. In this model, the MCP client can be broader web applications or mobile apps that communicate with the remote MCP server via HTTP protocol. Remote MCP Servers typically integrate enterprise-level features such as authentication, authorization, state management, and database access, enabling multi-user services.

This change addresses many limitations of Local MCP Servers in enterprise environments.

Limitations of Local MCP Server

Although Local MCP Server is simple and easy to use, it faces numerous challenges in enterprise applications:

● Local Environment Dependency: It depends on the user's local environment, such as needing to install execution environments like Python or Docker to run the MCP Server, which can be unfriendly to non-technical users.

● Security Risks: It is impractical for companies to configure sensitive database credentials, API keys, or other critical access tokens into each employee's local environment. This not only violates the principle of least privilege but also greatly increases the risk of credential leakage.

● Consistency Issues: When multiple users need access to the same enterprise resources, ensuring consistency in configuration and permissions becomes difficult, leading to potential data inconsistencies or permission chaos.

● Maintenance Costs: Deploying and maintaining MCP Servers on each user's device requires significant IT resources, with version updates, security patches, and configuration changes needing to be executed individually on each device.

Advantages of Remote MCP Server

Remote MCP Server addresses these issues through centralized deployment and management:

● Broadened Usage Scenarios: Non-technical users can access MCP capabilities via web or mobile applications anytime and anywhere over the internet.

● Centralized Security Control: Enterprises can enforce stringent access controls, encryption, and auditing mechanisms on remote servers to ensure the security of sensitive credentials.

● Unified Permission Management: Through centralized identity verification and authorization systems, businesses can precisely control each user's access rights to different resources.

● Simplified Deployment and Maintenance: Only central server maintenance is needed, greatly reducing operational costs and complexity.

4. Higress Open Source Remote MCP Server Hosting Solution

While Remote MCP Servers have many advantages, their implementation also presents new challenges, particularly in terms of authentication, authorization, service reliability, and observability. This is exactly where API gateways excel, and Higress, as an AI-native API gateway, provides a complete open-source MCP Server hosting solution.

This solution has been updated in the official Anthropic MCP Server project documentation, which can be accessed here: https://github.com/modelcontextprotocol/servers?tab=readme-ov-file#for-servers

Higress' MCP Server hosting solution adopts a layered architecture design, providing complete service access capabilities for AI Agents.

At the top layer, various AI Agents (such as Claude, Cursor, Cline, etc.) interact with Higress via the standard MCP protocol. Higress implements a complete security and management layer, including core functionalities like MCP session persistence, OAuth2 authentication, audit logs, rate limiting, and MCP routing.

On the service level, Higress supports three modes of MCP Server access:

● Built-in MCP Server implemented via Wasm plugins within Higress, suitable for connecting to the existing SaaS API ecosystem to provide MCP capabilities.

● Directly forwarding to externally supported MCP protocol services, suitable for connecting to external MCP capabilities that are already in place.

● Dynamically discovering external MCP Servers through service registries (like Nacos), and the ability to dynamically update the tool definitions of MCP Servers through the Nacos configuration center, suitable for businesses upgrading their traditional business APIs to MCP capabilities. Nacos is set to release an MCP Registry next week to achieve “zero modification” upgrades for existing application interfaces to the MCP protocol.

This flexible architectural design allows companies to choose the most suitable deployment method according to their needs, enjoying Higress' all-in-one hosting service while maintaining the independence of existing MCP Servers.

API Gateway Capabilities and Their Natural Fit with MCP Services

Higress is built on Envoy, inheriting traditional API gateway’s mature capabilities in authentication, authorization, rate limiting, and observability while being optimized for AI scenarios. These capabilities align closely with the needs of MCP services:

● Unified Authentication and Authorization: Higress provides an OAuth2 plugin that meets the new MCP protocol's requirements for authentication and authorization. Developers no longer need to write complex authentication code; simple configurations can implement secure user authentication and permission control.

● Fine-grained Traffic Control: Higress's rate limiting plugin can set different call quotas for different users and tools, preventing resource abuse and service overload.

● Full-link Observability: Higress integrates with open-source observability solutions like Prometheus and OTel, providing complete metric monitoring and distributed tracing capabilities, allowing operations teams to monitor the health and performance of MCP services in real-time.

● Audit Logs: Records all tool invocation behaviors to meet compliance requirements while providing data support for security analysis.

Wasm-based Expansion Capabilities of MCP Server

One of Higress's technological highlights is its support for WebAssembly (Wasm) plugins, which provides powerful expansion capabilities for the MCP Server:

With the Wasm plugin mechanism, developers can:

● Quickly add new Servers: Develop MCP Servers in various languages (currently featuring a Go language MCP Server SDK), compile them into Wasm, and dynamically load them into Higress without restarting the service.

● Security Isolation: Each Wasm plugin runs in a separate sandbox, ensuring that even if a plugin crashes, the stability of the gateway is not affected.

● Dynamic Updates: Thanks to Envoy's friendly handling of long connections and the dynamic update mechanism for Wasm plugins, the logic of MCP Servers can be real-time updated without interrupting traffic.

Advantages of Protocol Unloading

Currently, there are two versions of the MCP protocol, namely version 20241105 and version 20250326. Considering the rapid development of the MCP Server ecosystem, the standardization and optimization of the protocols are struggling to keep up. This poses potential risks for early adopters of the MCP, as maintaining a large number of MCP Servers becomes burdensome when facing future protocol version upgrades.

Advantages of Protocol Unloading

However, at the Higress gateway layer, we can help abstract multiple protocol version details and even implement protocol conversion capabilities such as MCP to REST/Dubbo, among others. Higress can complete the unloading of all versions of the MCP protocol at this gateway layer. Just like when we use Higress as a traditional API gateway, which uniformly unloads client protocols like HTTP1/HTTP2/HTTP3.

Currently, Higress can support both the 20241105 and 20250326 versions of the MCP protocol at a single access point. In terms of protocol transmission, it supports the POST+SSE model, which is currently the most widely adopted among AI tools, as well as the latest Streamable HTTP model.

We anticipate that as the application ecosystem demands more real-time capabilities, full-duplex two-way real-time communication protocols like WebSocket will become more popular and widely adopted. Higress natively supports WebSocket, enabling full-duplex two-way real-time communication, and supports connection persistence and graceful disconnection, enhancing Higress's advantages in MCP session persistence.

Deployment and Operational Advantages

The Higress MCP Server Hosting solution presents significant advantages in deployment and operations:

● Elastic Scalability: Based on Kubernetes's auto-scaling capabilities, ensuring high availability and resource utilization by automatically adjusting the number of instances according to traffic.

● Gray Release: Supports gray releases and A/B testing for the MCP Server, reducing update risks.

● One-click Deployment: Provides Helm Charts to simplify the deployment process and lower operational thresholds.

Using the Higress-hosted MCP Server can easily handle hundreds of thousands of tool invocation requests per second while maintaining millisecond-level response times, meeting the performance demands of enterprise-level applications.

5. Welcome to Participate in Building the Higress MCP Ecosystem

The Higress community is actively promoting the development of the MCP ecosystem and plans to provide a public MCP market on higress.ai around mid-April 2025, making it easier for users to experience the advantages of Remote MCP Servers. For example, this video directly demonstrates app integration on Cursor with an MCP Server deployed on Higress to interface with a database schema to write CRUD code:

Through the open MCP market, we can allow users to experience such capabilities without needing to deploy Higress and MCP Servers themselves.

Open MCP Market

Higress's MCP market will offer:

● Rich MCP Server Ecosystem: Covering common application scenarios such as document processing, data analysis, code assistance, and more.

● One-click Usage: Through a web interface, users can acquire MCP Server access points with a single click and use them in their AI tools.

● Community Sharing Mechanism: Developers can list MCP tools they have developed in the market after review by the open-source community, providing them for user use.

Enterprise-Level MCP Market

In addition to the open-source community version, Alibaba Cloud API Gateway will soon launch its MCP market, providing more comprehensive services aimed at enterprise users:

● Enterprise-Level SLA Guarantee: Promising 99.99% service availability to meet the needs of critical business operations.

● Compliance Certification: Meeting strict regulations in industries like finance and healthcare through multiple security compliance certifications.

● Expert Support: Providing 24/7 technical support and consulting services to help enterprises rapidly resolve any issues.

This allows enterprises to enjoy the SaaS MCP ecosystem while also building their own internal MCP Server ecosystem.

Ecological Co-construction

If you are a SaaS service provider looking to build an MCP Server to allow users to pay for their own APIs in AI scenarios, the Higress community welcomes you to collaborate (add WeChat: zjjxg2018 for connection).

● Technical Support: We provide technical guidance for MCP Server development, or our open-source community can assist you in construction.

● Joint Promotion: Promotion of the partner’s MCP services in Higress official channels.

● Market Listing: Assistance in listing your MCP Server in the official MCP market for Higress hosting.

Through open-source means, the Higress community is dedicated to promoting the vibrant development of the MCP ecosystem, allowing more developers and enterprises to easily build and deploy MCP Servers, unleashing the full potential of AI assistants.

With the push from industry giants like OpenAI and Anthropic, as well as the support of open-source projects like Higress, the MCP protocol is transitioning from a de facto standard to an industry standard. In the future, we have every reason to believe that MCP will become a key bridge connecting AI with the real world, bringing unprecedented intelligent experiences to various industries.