home / mcp / educhain mcp server

Educhain MCP Server

Provides Educhain MCP capabilities to generate MCQs, flashcards, and lesson plans via a Gemini-backed MCP backend.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "anudeep-codespace-educhain_mcp_server": {
      "command": "uv",
      "args": [
        "run",
        "mcp",
        "dev",
        "main.py"
      ],
      "env": {
        "GOOGLE_API_KEY": "YOUR_GOOGLE_API_KEY"
      }
    }
  }
}

You can run Educhain’s MCP Server locally to generate learning content such as MCQs, flashcards, and lesson plans by interfacing a client (like Claude Desktop) with a Python-based MCP backend. It runs alongside an LLM (Gemini) and exposes a debugging proxy for development and testing.

How to use

To use the Educhain MCP Server, start it in dev mode and connect a client that supports MCP endpoints. You will obtain a local proxy URL to access the server’s resources and tools, then use the client to request MCQs, flashcards, or lesson plans.

How to install

Prerequisites: Node.js LTS is required for debugging steps that involve the MCP tooling. Python 3.10 is recommended for running the MCP server. Use a package manager that you prefer for Python packages.

# 1) Install Python (ensure python3 and pip are available)
# 2) Install uv (universal) as the MCP runtime
pip install uv
# 3) Initialize an existing project (if needed)
uv init
# 4) Add required MCP packages for Educhain
uv add "educhain" "mcp[cli]"
# 5) Set your Google Gemini API key in the environment file
# Inside .env file
GOOGLE_API_KEY=YOUR_GOOGLE_API_KEY
# 6) Run the MCP server in development mode
uv run mcp dev main.py

Additional sections

Configuration notes: The server exposes a local HTTP proxy for debugging when started in dev mode. You will see a tokenized proxy URL like http://localhost:6274/?PROXY_API_TOKEN=<proxy token> which you paste into a browser to connect. The client can then interact with Educhain’s components.

Environment and logs: Logs are typically written to platform-specific locations. Check the following paths to review logs after launching the server: - Windows: %APPDATA%\\Claude\\logs\\mcp.log - macOS: ~/Library/Logs/Claude/mcp.log

Notes and caveats: - Node.js (LTS) is required for debugging flows. - pyenv is not recommended for this workflow. - The frontend client (Claude Desktop) can connect to the MCP server once it is registered in the client with the appropriate command and arguments.

Config example

{
  "mcpServers": {
    "Educhain - MCP server": {
      "command": "uv",
      "args": [
        "run",
        "mcp",
        "dev",
        "main.py"
      ],
      "env": {
        "GOOGLE_API_KEY": "YOUR_GOOGLE_API_KEY"
      }
    }
  }
}

Available tools

generate_mcq

Generates multiple-choice questions based on a given topic or curriculum outline.

generate_flashcard

Creates flashcards for key concepts to aid quick review and spaced repetition.

generate_lessonplan

Produces structured lesson plans outlining objectives, activities, and assessments.