Diavgeia Mcp Server

Diavgeia MCP Server

A Model Context Protocol (MCP) server that provides access to Greek government transparency data from the Diavgeia platform. This server enables AI assistants like Claude to search and retrieve Greek government decisions, making public administration data more accessible.

🌟 Features

• Search Government Decisions: Query decisions by keywords, date ranges, organization, and type

MCP demo 📹

• Retrieve Decision Details: Get comprehensive information about specific decisions using their ADA (unique identifier)

MCP demo📹

• Date-aware Queries: Natural language date parsing (e.g., "last year", "January 2024")
• Organization Search: Find decisions by ministry or government organization

🚀 Quick Start

Local Development

1. Clone the repository

``bash git clone https://github.com/michalisKout/diavgeia-mcp-server.git cd diavgeia-mcp-server ``

1. Install dependencies

``bash pnpm install ``

3a. Start development server ``bash pnpm dev ` Your MCP server will be available at http://localhost:8787/sse`

4a. Use mcp inspector More info: https://modelcontextprotocol.io/docs/tools/inspector

3b. Start development server ``bash pnpm build:local `` Run locally using stdio transport protocol:

``bash node /Users/michaliskoutridis/dev/projects/ai-agents/diavgeia-mcp-server/dist/local.js ``

See [mcp-config-schema.json](./mcp-config-schema.json) for the full configuration schema.

🛠 Available Tools

The Diavgeia MCP Server provides two main tools for accessing Greek government transparency data:

1. search-decisions - Search for government decisions with various filters
2. get-decision - Retrieve detailed information about a specific decision

For detailed documentation on all available tools, parameters, examples, and best practices, see [docs/TOOLS.md](docs/TOOLS.md).

Quick Examples

Search for education decisions in 2024:

{
  "q": "εκπαίδευση",
  "from_date": "2024-01-01",
  "to_date": "2024-12-31"
}

Get details for a specific decision:

{
  "ada": "ΨΧ465Κ8Ω-123"
}

📚 Documentation

• [Tools Reference](docs/TOOLS.md) - Complete tools documentation
• [Security Guide](SECURITY.md) - Security policy and vulnerability reporting
• [Contributing Guidelines](CONTRIBUTING.md) - How to contribute

🏗 Project Structure

src/
├── index.ts                 # Main MCP server entry point
├── api/
│   └── diavgeia.ts         # Diavgeia API client
├── modules/
│   ├── decision/           # Decision retrieval tool
│   │   ├── constants.ts
│   │   ├── schema.ts
│   │   └── tool.ts
│   └── search/             # Decision search tool
│       ├── constants.ts
│       ├── schema.ts
│       └── tool.ts
├── types/
│   └── diavgeia.ts         # TypeScript type definitions
├── utils/                  # Utility functions
└── prompts/
    └── template.ts         # System prompts

🔧 Development

Prerequisites

• Node.js 23+
• pnpm package manager

🤝 Contributing

We welcome contributions! Please see our [Contributing Guidelines](CONTRIBUTING.md) for details.

Quick Contribution Steps

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes
4. Run tests and linting (pnpm lint:fix && pnpm type-check)
5. Commit your changes following conventional commits
6. Push to the branch (git push origin feature/amazing-feature)
7. Open a Pull Request

Development Setup

# Install dependencies
pnpm install

# This will automatically set up git hooks via lefthook
# The hooks will:
# - Lint and format code on pre-commit
# - Validate commit messages
# - Run checks before push

Commit Message Convention

This project uses Conventional Commits. All commit messages are validated automatically.

Examples:

git commit -m "feat: add new search filter"
git commit -m "fix: resolve date parsing bug"
git commit -m "docs: update api documentation"

See [Commit Conventions](.github/COMMIT_CONVENTION.md) for detailed guidelines.

Restricted: Deploying

Contributors CANNOT deploy because:

1. They don't have CLOUDFLARE_API_TOKEN
2. They aren't logged in as the owner
3. They don't have permissions to do so

Setting Up Your Own Deployment

If someone wants to deploy their OWN instance:

1. Create Cloudflare Account

• Sign up at https://dash.cloudflare.com/sign-up
• Free tier is available

2. Get Your Account ID

npx wrangler login
npx wrangler whoami

3. Update wrangler.jsonc (Optional)

If deploying to their own account, they should update:

{
  "name": "their-own-server-name",  // Change this
  // account_id will be read from their login
}

4. Deploy

npx wrangler login  // Authenticate
pnpm deploy         // Deploy to THEIR account

5a. Cloudflare AI Playground

1. Go to Cloudflare AI Playground
2. Enter your deployed MCP server URL: your-deployment.workers.dev/sse
3. Start using the Diavgeia tools directly in the playground!

5b. Claude Desktop Integration

To connect this MCP server to Claude Desktop, add the following to your Claude Desktop configuration:

{
  "mcpServers": {
    "diavgeia": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://your-deployment.workers.dev/sse"
      ]
    }
  }
}

�📝 License

This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.

Running Checks Locally

Before opening a PR, you can run checks locally:

# Type check
pnpm run check

# Lint
pnpm run lint

# Build
pnpm run build:local

🙏 Tools and APIs used

• Diavgeia Platform - Greek Government Transparency Initiative
• Model Context Protocol - Standard for AI tool integration
• Cloudflare Workers - Serverless deployment platform
• Smithery - Smithery mcp servers hosting

📬 Support

• 🐛 Bug Reports: Open an issue
• 💡 Feature Requests: Start a discussion

---

Made with ❤️ for transparency in Greek governance

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: michalisKout
• Source: michalisKout/diavgeia-mcp-server
• License: MIT

Install and usage instructions live in the source repository linked above.