Skolverket

Skolverket MCP Server provides access to Skolverket's open APIs via the MCP (Model Context Protocol). It exposes three main data domains: the Syllabus API (curricula, subjects, courses, and knowledge requirements), the School Units register (search and filter schools and units), and the Planned Educations API (vocational and adult education start dates, seats, and study pace). The server also ships with a set of tools, resources, and prompt templates to support AI-assisted querying, comparison, and analysis of educational data. To use it, connect your MCP-enabled AI assistant or chat client to the hosted endpoint at https://skolverket-mcp.onrender.com/mcp, or run a local instance and point your client to the local MCP Transport (stdio or HTTP as configured). Once connected, you can issue natural language prompts that leverage the 41 tools, 4 resources, and 5 prompt templates designed to help with planning, comparisons, and data analysis across the three API domains.

Prerequisites:

• Git
• Node.js (v14+ recommended) and npm

Local installation steps:

1. Clone the repository

git clone https://github.com/KSAklfszf921/skolverket-mcp.git
cd skolverket-mcp

2. Install dependencies and build

npm install
npm run build

3. Run the server locally

# Start the built server (adjust path if needed)
node dist/index.js

4. Configure MCP client (example for Claude/ChatGPT):

• Add the MCP server with the following JSON snippet:

{
  "mcpServers": {
    "skolverket": {
      "command": "node",
      "args": ["/absolut/sökväg/till/skolverket-mcp/dist/index.js"]
    }
  }
}

5. Verify health (if provided by the server): visit https://skolverket-mcp.onrender.com/health or the local health endpoint you run.

Notes and tips:

• The server provides access to three major APIs: Syllabus API, Skolenhetsregistret, and Planned Educations API. Explore the available 41 tools (across 17 Syllabus, 4 School Units, 17 Planned Educations, plus diagnostics and support data), 4 resources, and 5 prompt templates for advanced queries.
• If running locally, ensure your build outputs to dist/index.js as shown in the documentation examples; adjust the path in your mcp_config accordingly.
• Environment variables can be added under env in the mcp_config entry if your deployment requires API keys or specific runtime settings.
• When connecting clients (ChatGPT, Claude, Codex, etc.), follow the client-specific instructions in the README to add the MCP server endpoint and transport method (HTTP, stdio, etc.).
• If you encounter connection issues, verify the server is reachable at the host URL and that the MCP transport method (http/stdio) matches the client configuration.