datagovma-mcp
MCP server for the Moroccan Open Data portal (data.gov.ma) enabling search and retrieval of datasets, resources, organizations, and groups via CKAN API.
README
datagovma-mcp
<p align="center"> <img src="assets/banner.jpg" alt="datagovma-mcp banner" width="100%" /> </p>
MCP server for the Moroccan Open Data portal (data.gov.ma).
datagovma-mcp exposes curated MCP tools backed by CKAN Action API endpoints so AI clients can discover datasets, resources, organizations, and groups in a consistent typed shape.
What this project provides
- Streamable HTTP MCP server built with FastMCP.
- Typed and validated tool inputs/outputs.
- Centralized CKAN envelope validation (
successchecks, JSON checks, timeout handling). - Test suite covering success and error paths.
- Docker image and Docker Compose setup.
Implemented MCP tools
| MCP tool | CKAN action | Purpose |
|---|---|---|
get_portal_status |
status_show |
Check portal identity and CKAN version. |
search_datasets |
package_search |
Full-text search with paging and optional facets. |
get_dataset |
package_show |
Fetch one dataset by slug or UUID. |
list_datasets |
package_list |
List dataset names with pagination. |
get_resource |
resource_show |
Fetch one resource and download metadata. |
search_resources |
resource_search |
Search resources using strict field:value query format. |
list_organizations |
organization_list |
List organization names with pagination. |
get_organization |
organization_show |
Fetch organization details and optional datasets. |
list_groups |
group_list |
List group names with pagination. |
get_group |
group_show |
Fetch group details and optional datasets. |
get_dataset_facets |
package_search (rows=0) |
Aggregate facet buckets (tags, groups, organization). |
Requirements
- Python
>=3.11 - uv for dependency management
- Docker (optional, for containerized runs)
Quick start (local)
uv sync --frozen --group dev
cp .env.example .env
set -a && source .env && set +a
uv run datagovma-mcp
Default endpoint:
- MCP:
http://127.0.0.1:8000/mcp - Health:
http://127.0.0.1:8000/healthz
Quick health check:
curl -fsS http://127.0.0.1:8000/healthz
Expected response:
200 OKwith bodyok.
Run with Docker
docker compose up --build
The service starts on port 8000 by default and exposes /mcp and /healthz.
Connect from LLM clients
Claude Desktop
After the server is running, edit:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
Then add:
{
"mcpServers": {
"opengovma": {
"command": "npx",
"args": ["mcp-remote", "http://127.0.0.1:8000/mcp"]
}
}
}
Claude Code
Add the server with Claude Code CLI:
claude mcp add --transport http opengovma http://127.0.0.1:8000/mcp
Codex (Desktop app and CLI)
Add the server with Codex CLI:
codex mcp add opengovma --url http://127.0.0.1:8000/mcp
VS Code (GitHub Copilot MCP)
Create .vscode/mcp.json in your workspace (or open your user MCP config via command palette) and add:
{
"servers": {
"opengovma": {
"type": "http",
"url": "http://127.0.0.1:8000/mcp"
}
}
}
Cursor
Add this to .cursor/mcp.json (project) or ~/.cursor/mcp.json (global):
{
"mcpServers": {
"opengovma": {
"url": "http://127.0.0.1:8000/mcp"
}
}
}
Configuration
Environment variables are loaded by your shell or container runtime.
| Variable | Default | Description |
|---|---|---|
MCP_HOST |
127.0.0.1 |
Host interface used by Uvicorn inside the app. |
MCP_BIND_HOST |
0.0.0.0 |
Docker Compose helper mapped to MCP_HOST. |
MCP_PORT |
8000 |
Listening port. |
MCP_WORKERS |
1 |
Number of worker processes (>=1). |
MCP_RELOAD |
false |
Hot reload mode for development (true/false, 1/0, yes/no, on/off). |
MCP_TIMEOUT_KEEP_ALIVE |
5 |
HTTP keep-alive timeout in seconds (>=1). |
MCP_LOG_LEVEL |
INFO |
One of DEBUG, INFO, WARNING, ERROR, CRITICAL. |
MCP_LOG_FORMAT |
auto |
One of auto, plain, rich. |
Notes:
MCP_RELOAD=truerequiresMCP_WORKERS=1.MCP_LOG_FORMAT=richfalls back to plain logs if Rich is unavailable.
Continuous integration
GitHub Actions workflow: .github/workflows/ci.yml
CI runs on pull requests and pushes to main/master and includes:
- dependency review (PR only)
- lint (
ruff) - format check (
ruff format --check) - type check (
mypy) - tests (
pytest) - package build (
uv build)
Contributing
Contributions are welcome. See CONTRIBUTING.md for setup, checks, and PR guidance.
License
Licensed under the MIT License. See LICENSE.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。