What is the MCP Server?

MCP (Model Context Protocol) Server is a secure URL fetching tool designed for AI agents and applications. It allows AI models to access and interact with external data sources and tools safely.

What are the key features of the MCP Server?

Key features include secure URL fetching, domain resolution, private/internal IP check, redirect handling, custom HTTP methods and headers, an allowlist for trusted domains/IPs, and restriction to HTTP and HTTPS schemes.

How does the MCP Server ensure security?

The MCP Server ensures security by preventing access to private/internal IPs, using an allowlist for trusted domains/IPs, restricting to HTTP and HTTPS schemes, and setting SNI to match the hostname for HTTPS connections.

What is the purpose of the allowlist?

The allowlist is a list of trusted domains and IPs that the MCP Server is permitted to access. This prevents access to potentially malicious or untrusted resources.

How does the MCP Server handle redirects?

The MCP Server handles redirects up to a certain limit (typically 3) to ensure that AI agents can follow links to the correct destination without getting stuck in redirect loops.

Can I use custom HTTP methods with the MCP Server?

Yes, the MCP Server supports custom HTTP methods and headers, allowing for flexible interaction with web servers and APIs.

What is SNI, and why is it important?

SNI (Server Name Indication) is a feature that allows the MCP Server to present the correct SSL certificate for HTTPS connections. This prevents man-in-the-middle attacks and protects against data interception.

What is FastMCP, and how does it relate to the MCP Server?

FastMCP is a high-performance library for secure URL fetching that the MCP Server is built upon. It provides the underlying infrastructure for secure communication and efficient data transfer.

How do I install the MCP Server in my UBOS environment?

You can install the MCP Server from the UBOS Asset Marketplace. Simply locate the asset and click the install button to add it to your UBOS environment.

Can I use the MCP Server with AI agents developed outside of the UBOS platform?

While the MCP Server is designed to integrate seamlessly with the UBOS platform, it can potentially be used with external AI agents, provided they can interface with the server's API.

What kind of logging and monitoring is available with the MCP Server?

The MCP Server provides logging of requests and security-related events, which can be monitored to identify potential issues or threats.

Is there a limit to the size of the content that can be fetched by the MCP Server?

While there isn't a hard-coded limit, extremely large content may impact performance. It's recommended to test and optimize for your specific use cases.

What kind of support is available for the MCP Server?

UBOS provides documentation, tutorials, and community support to help you get started with the MCP Server. You can also contact UBOS support for assistance with any issues or questions.

How do I update the MCP Server to the latest version?

You can update the MCP Server through the UBOS Asset Marketplace. The platform will notify you when updates are available.

Secure Fetch – Overview | MCP Marketplace

UBOS Asset Marketplace: Secure URL Fetching with MCP Server

In the ever-evolving landscape of AI and machine learning, the need for secure and reliable data access is paramount. UBOS, a full-stack AI Agent Development Platform, recognizes this need and offers a comprehensive solution through its Asset Marketplace. One of the key assets available is the MCP (Model Context Protocol) Server, designed to facilitate secure URL fetching for AI agents and applications. This document provides an in-depth overview of the MCP Server asset, its use cases, key features, and integration with the UBOS platform.

What is MCP Server?

MCP, or Model Context Protocol, is an open protocol that standardizes how applications provide context to Large Language Models (LLMs). The MCP Server acts as a bridge, allowing AI models to access and interact with external data sources and tools securely. In the context of URL fetching, the MCP Server ensures that AI agents can retrieve data from the web without compromising security or exposing internal resources.

The UBOS Asset Marketplace provides a curated collection of tools and components that enhance the capabilities of AI agents. The MCP Server is a critical asset within this marketplace, offering a secure and efficient way to fetch URLs, validate domains, and prevent unauthorized access.

Use Cases

The MCP Server asset addresses a wide range of use cases across various industries. Here are some notable examples:

AI-Powered Web Scraping: AI agents can use the MCP Server to securely scrape data from websites for training datasets, market research, or competitive analysis. The server ensures that only trusted domains are accessed, mitigating the risk of malicious code or data breaches.
Content Moderation: AI models can leverage the MCP Server to fetch content from URLs for analysis and moderation. This is particularly useful in identifying and filtering out harmful or inappropriate content on social media platforms or online forums.
News Aggregation: News aggregators can use the MCP Server to fetch articles from various news sources securely. The server helps to ensure that the aggregator only retrieves content from reputable sources, maintaining the integrity of the information provided to users.
Threat Intelligence: Security firms can use the MCP Server to fetch threat intelligence feeds from trusted sources. The server ensures that the intelligence data is accessed securely and that internal systems are protected from potential threats.
Financial Analysis: Financial analysts can use AI agents to fetch data from financial websites and APIs securely. The MCP Server helps to ensure that sensitive financial data is accessed safely and that compliance requirements are met.
E-commerce: AI agents for e-commerce can fetch product data, reviews, and pricing information from competitor websites, enabling dynamic pricing and market analysis. The secure URL fetching ensures data integrity and prevents security vulnerabilities.
Healthcare: AI agents in healthcare can securely fetch medical research papers, clinical trial data, and patient information (with proper authorization and compliance) to aid in diagnosis, treatment planning, and personalized medicine.

Key Features

The MCP Server asset comes equipped with a comprehensive set of features designed to ensure secure and efficient URL fetching:

Secure URL Fetching: The core functionality of the MCP Server is to fetch URLs securely. It enforces strict security policies to prevent unauthorized access and protect against potential threats.
Domain Resolution: The server resolves domains to IPs, allowing for granular control over the destinations that AI agents can access. This feature helps to prevent access to malicious or untrusted domains.
Private/Internal IP Check: One of the key security measures implemented by the MCP Server is the prevention of access to private or internal IPs. Unless explicitly allowed, the server blocks requests to internal network addresses, mitigating the risk of internal network breaches.
Redirect Handling: The MCP Server handles redirects (up to 3 by default), ensuring that AI agents can follow links to the correct destination. This feature is essential for navigating complex websites and retrieving data from multiple sources.
Custom HTTP Methods and Headers: The server supports custom HTTP methods and headers, allowing for flexible interaction with web servers and APIs. This feature enables AI agents to perform a wide range of tasks, from simple data retrieval to complex data manipulation.
Allowlist for Trusted Domains/IPs: To provide an extra layer of security, the MCP Server uses an allowlist for trusted domains and IPs. Only domains and IPs on the allowlist are permitted, ensuring that AI agents can only access approved resources.
HTTP and HTTPS Scheme Restriction: The MCP Server only permits HTTP and HTTPS schemes, preventing the use of other potentially insecure protocols. This restriction helps to minimize the attack surface and protect against protocol-based vulnerabilities.
SNI (Server Name Indication): For HTTPS connections, the MCP Server sets the SNI to match the hostname, ensuring that the correct SSL certificate is presented. This feature helps to prevent man-in-the-middle attacks and protect against data interception.
Integration with FastMCP: The MCP Server is built on top of FastMCP, a high-performance library for secure URL fetching. FastMCP provides the underlying infrastructure for secure communication and efficient data transfer.

Getting Started with MCP Server on UBOS

Integrating the MCP Server asset into your UBOS environment is a straightforward process. Here’s a step-by-step guide:

Access the UBOS Asset Marketplace: Log in to your UBOS account and navigate to the Asset Marketplace. You’ll find a variety of pre-built assets designed to enhance your AI agent development.
Locate the MCP Server Asset: Use the search functionality or browse the available categories to find the MCP Server asset. The marketplace provides detailed information about each asset, including its features, use cases, and documentation.
Install the Asset: Once you’ve found the MCP Server asset, click the install button to add it to your UBOS environment. The installation process typically involves configuring the asset with your specific requirements.
Configure the Asset: After installation, you’ll need to configure the MCP Server asset to suit your specific needs. This may involve setting up the allowlist, configuring HTTP methods, and defining other security parameters.
Integrate with Your AI Agents: With the MCP Server asset installed and configured, you can now integrate it with your AI agents. This involves calling the server’s API to fetch URLs securely.

Example Configuration

To help you get started, here’s an example configuration for the MCP Server asset:

Allowlist: Define a list of trusted domains and IPs that your AI agents are allowed to access. For example, example.com, trusted-domain.org.
HTTP Methods: Specify the HTTP methods that are allowed. Common methods include GET, POST, PUT, and DELETE.
Headers: Configure any custom headers that need to be included in the HTTP requests.

Benefits of Using MCP Server with UBOS

Using the MCP Server asset within the UBOS platform offers several key benefits:

Enhanced Security: The MCP Server provides a robust set of security features that protect against unauthorized access and potential threats. This is particularly important for AI agents that handle sensitive data.
Simplified Integration: The UBOS platform simplifies the integration of the MCP Server asset with your AI agents. The platform provides a consistent and intuitive interface for managing and configuring assets.
Increased Efficiency: The MCP Server is built on top of FastMCP, a high-performance library that ensures efficient URL fetching. This helps to minimize latency and improve the overall performance of your AI agents.
Reduced Complexity: By using the MCP Server asset, you can reduce the complexity of your AI agent development. The server handles the low-level details of URL fetching, allowing you to focus on the higher-level logic of your agents.
Cost Savings: The UBOS platform offers a cost-effective way to access and use the MCP Server asset. Instead of building your own secure URL fetching solution, you can leverage the pre-built asset and save time and resources.

Best Practices

To maximize the benefits of the MCP Server asset, consider the following best practices:

Regularly Review the Allowlist: Keep the allowlist up-to-date with the latest trusted domains and IPs. Remove any domains or IPs that are no longer needed.
Monitor Server Logs: Monitor the MCP Server logs for any suspicious activity. This can help you to identify and respond to potential security threats.
Apply Security Patches: Keep the MCP Server software up-to-date with the latest security patches. This will help to protect against known vulnerabilities.
Follow the Principle of Least Privilege: Grant your AI agents only the minimum level of access that they need to perform their tasks. This will help to minimize the impact of any potential security breaches.

Conclusion

The MCP Server asset is a valuable tool for AI agent developers who need to fetch URLs securely. By integrating this asset into your UBOS environment, you can protect your AI agents from unauthorized access and potential threats. With its comprehensive set of features, simplified integration, and increased efficiency, the MCP Server asset is an essential component of any secure AI agent development platform. As the demand for secure and reliable data access continues to grow, the MCP Server asset will play an increasingly important role in the world of AI and machine learning.

UBOS is committed to providing a comprehensive and secure platform for AI agent development. The Asset Marketplace is a key part of this commitment, offering a curated collection of tools and components that enhance the capabilities of AI agents. The MCP Server asset is just one example of the many valuable resources available in the marketplace. By leveraging these resources, you can build powerful and secure AI agents that drive innovation and deliver real business value.