UBOS Asset Marketplace: Unleashing the Potential of MCP Servers with Furikake

In the rapidly evolving landscape of AI, the ability to seamlessly integrate and manage AI models with diverse data sources and tools is paramount. UBOS, a full-stack AI Agent Development Platform, recognizes this critical need and provides a robust ecosystem for building, orchestrating, and deploying AI Agents. At the heart of this ecosystem lies the Model Context Protocol (MCP), an open standard that revolutionizes how applications provide context to Large Language Models (LLMs).

The UBOS Asset Marketplace is designed to empower developers and businesses by offering a curated collection of tools and resources that streamline the AI Agent development process. Among these invaluable assets, Furikake (or furi) emerges as a game-changer for MCP server management.

What is an MCP Server and Why Does It Matter?

Before diving into the specifics of Furikake, it’s essential to understand the role of an MCP server. An MCP server acts as a bridge, facilitating communication between AI models and external data sources, APIs, and other tools. It standardizes how applications provide context to LLMs, ensuring that AI models have access to the information they need to make informed decisions and perform complex tasks.

Think of it this way: LLMs are like brilliant students who need access to textbooks, research papers, and lab equipment to truly excel. The MCP server provides that access, connecting the LLM to the vast world of data and tools it needs to thrive.

Furikake: Your Local CLI & API for MCP Management

Furikake (furi) is a user-friendly, local Command-Line Interface (CLI) and Application Programming Interface (API) designed to simplify the management of MCP servers. Built with Bun and Typescript, Furikake offers a seamless experience for downloading, installing, running, and managing MCPs. It’s particularly well-suited for developers working with Typescript/Javascript-based MCPs.

Key Features and Benefits of Furikake:

• Effortless MCP Downloading: Furikake allows you to easily download MCPs directly from GitHub repositories. This eliminates the need for manual downloads and configurations, saving you valuable time and effort.

• Fully Featured CLI: The intuitive CLI provides a comprehensive set of commands for managing MCPs, complete with visually appealing loading indicators. You can add, delete, list, start, stop, restart, and check the status of MCPs with ease.

• HTTP API Routes: Furikake exposes HTTP API routes for interacting with your MCPs. This allows you to programmatically control and monitor your MCPs from other applications or services.

• Process State Management: Furikake leverages PM2 for robust process state management, ensuring that your MCPs are running reliably and can be easily restarted in case of failures.

• Log Management: Furikake provides access to logs for each MCP process, making it easy to diagnose issues and monitor performance.

• NPM Integration: Furikake utilizes npm to configure and run MCPs, leveraging the vast ecosystem of Node.js packages.

• Bun & Typescript Powered: Built with Bun and Typescript, Furikake offers excellent performance, type safety, and a modern development experience.

• Seamless Integration with UBOS: Furikake integrates seamlessly with the UBOS platform, allowing you to leverage its powerful features for AI Agent development and orchestration.

Use Cases for Furikake in the UBOS Ecosystem

Furikake empowers UBOS users in a variety of scenarios, including:

• Rapid Prototyping of AI Agents: Furikake simplifies the process of setting up and managing MCPs, allowing developers to quickly prototype and test AI Agents with different data sources and tools.

• Building Custom AI Agents: Furikake enables developers to build custom AI Agents that leverage specific data sources and tools relevant to their business needs. By connecting LLMs to these custom MCPs, you can create highly specialized AI Agents that solve unique problems.

• Orchestrating Multi-Agent Systems: Furikake facilitates the orchestration of multi-agent systems by providing a centralized interface for managing multiple MCPs. This allows you to build complex AI systems that leverage the strengths of different AI Agents.

• Connecting to Enterprise Data: Furikake makes it easy to connect AI Agents to enterprise data sources, such as databases, CRMs, and ERP systems. This allows you to build AI Agents that can access and process real-time business data.

• Integrating with Third-Party APIs: Furikake simplifies the integration of AI Agents with third-party APIs, such as weather services, social media platforms, and payment gateways. This allows you to build AI Agents that can access and leverage external services.

Getting Started with Furikake

To start using Furikake, you’ll need to have Bun installed on your macOS or Linux system. You can install Bun by running the following command:

bash curl -fsSL https://bun.sh/install | bash

Once Bun is installed, you can use Furikake to manage MCPs as follows:

• Add an MCP:

  bash furi add <author/repo>

  Example: furi add smithery-ai/mcp-fetch

• Delete an MCP:

  bash furi delete <author/repo>

  Example: furi delete smithery-ai/mcp-fetch

• List installed MCPs:

  bash furi list

• Start an MCP:

  bash furi start <author/repo> -e

  -e <env> is optional and depends on the MCP server. Use furi env <author/repo> to view environment variables.

• Stop an MCP:

  bash furi stop <author/repo>

• Restart an MCP:

  bash furi restart <author/repo>

• Check the status of an MCP:

  bash furi status <author/repo>

• List all running MCPs:

  bash furi status all

• Call a running MCP:

  bash furi call <author/repo>

HTTP API Usage

Furikake also provides an HTTP API for interacting with your MCPs. To enable the HTTP proxy, run:

bash furi http on

To disable the proxy, run:

bash furi http off

Available Routes:

• /api/status - Get a list of all running MCPs.
• /api/tools - Get a list of all available tools for all online MCPs.
• /api/<mcp-name>/status - Get the status of a specific MCP.
• /api/<mcp-name>/restart - Restart a specific MCP.
• /api/<mcp-name>/logs - Get the logs for a specific MCP.
• /api/<mcp-name>/tools - Get a list of available tools for a specific MCP.

<mcp-name> can be either author/repo or alias.

UBOS: Your Partner in AI Agent Development

UBOS is committed to providing developers with the tools and resources they need to build and deploy cutting-edge AI Agents. Furikake is just one example of the many valuable assets available in the UBOS Asset Marketplace.

By leveraging the UBOS platform and tools like Furikake, you can:

• Accelerate your AI Agent development: UBOS provides a comprehensive set of tools and services that streamline the development process, from data ingestion to model deployment.

• Build more intelligent and context-aware AI Agents: UBOS enables you to connect AI Agents to diverse data sources and tools, allowing them to make more informed decisions.

• Orchestrate complex multi-agent systems: UBOS provides a robust platform for orchestrating multi-agent systems, allowing you to build AI systems that can solve complex problems.

• Scale your AI deployments: UBOS is designed to scale to meet the demands of enterprise-level AI deployments.

In conclusion, Furikake is an indispensable tool for anyone working with MCP servers in the UBOS ecosystem. Its ease of use, comprehensive features, and seamless integration with UBOS make it the perfect choice for managing and orchestrating AI Agents. Embrace Furikake and unlock the full potential of your AI Agent development efforts.

We encourage you to explore the UBOS Asset Marketplace and discover the many other valuable resources available to you. Together, we can build the future of AI Agents.