aduib -router

Aduib MCP Router is a multi-MCP aggregator router that supports multiple MCP protocols, designed to simplify the management and usage of MCP servers. It acts as a central hub that routes requests to various MCP servers and tools, enabling you to orchestrate different MCP sources through a single interface. The router can be configured via a local or remote MCP configuration file (specified by MCP_CONFIG_PATH) and is designed to work with uvx, a lightweight launcher for MCP components. The built-in Playground provides a visual, in-browser interface to inspect loaded tools, view health status, trigger calls, and export execution history, making it easy to validate routing chains before going live. Use the provided configuration to start the router and expose the playground for quick experimentation and diagnostics.

Prerequisites:

• Python 3.x and the uv tool (a Python-based MCP runner) installed on your system.
• Access to the network as needed to fetch any required MCP resources.

Installation steps:

1. Install uv (Python package) globally if not already installed:

pip install uv

2. Ensure uv is available in your PATH. If you prefer a package-based approach, you can also install via your OS package manager (where available):

• macOS: brew install uv
• Windows: follow the uv installation instructions from the official docs or use pip to install the Python package and ensure Scripts is in PATH

3. Prepare your MCP router configuration file and place it at the path specified by MCP_CONFIG_PATH in your environment (default in readme is ./config.json). Example placement:

./config.json

4. Start the router using the configured command from your MCP config:

uvx aduib-mcp-router

5. Verify that the router is running and accessible, and open the Playground at http://localhost:<port>/playground to inspect available tools and health status.

Tips and notes:

• Ensure MCP_CONFIG_PATH points to a valid JSON configuration (or a URL if remote retrieval is supported).
• The router uses a caching layer with configurable TTLs; adjust MCP_CACHE_TTL and related settings as needed for your environment.
• Health checks can be tuned via MCP_HEALTH_CHECK_INTERVAL, MCP_HEALTH_CHECK_TIMEOUT, MCP_MAX_CONSECUTIVE_FAILURES, MCP_AUTO_RECONNECT, and MCP_RECONNECT_DELAY.
• For debugging, the Playground provides an interactive way to trigger MCP calls, inspect responses, and export histories. Check the route and tool descriptions to understand input schemas.
• If you encounter connection timeouts (MCP_CONNECTION_TIMEOUT) or frequent retries, review network connectivity, target endpoints, and server-side rate limits.
• Since this router aggregates multiple MCP sources, ensure proper security headers and access controls are in place for any external endpoints (e.g., the Authorization header shown in the sample json configuration).
• The configuration path can be local or remote; if you use a remote URL, ensure authentication and network access are properly configured.