Jobs mit Impact
About the Project
Jobs mit Impact brings together jobs in the German nonprofit sector from over 100 sources. AI supports the extraction of job data. An MCP server lets AI assistants search for jobs and retrieve listing details.
What is MCP?
The Model Context Protocol (MCP) connects AI applications to data sources and tools. Here, the server provides job search and listing details; its eight tools read data without changing job listings.
8 MCP tools at a glance
The MCP server exposes 8 read-only tools that let AI assistants query the full dataset:
search_jobs Search job listings
list_jobs All jobs with filters (sector, location, org type)
get_job_details Full details for a specific job listing
get_jobs_by_location Filter jobs by city or region
get_jobs_by_organization All positions at a specific organisation
list_sources All 100+ sources and their metadata
get_source_details Details for a specific scraping source
get_stats Statistics on jobs, sources and organisations
Set up an MCP connection
The server endpoint is https://mcp.jobsmitimpact.de/mcp. A configuration entry can look like this; setup depends on the MCP client:
{
"mcpServers": {
"jobcore": {
"url": "https://mcp.jobsmitimpact.de/mcp"
}
}
}How It Works
- Autonomous scraping from 100+ nonprofit job sources
- Four extraction stages: JSON-LD, platform-specific rules, heuristics and language models
- MCP server: 20 requests/minute, fully read-only
My Contribution
I founded Jobs mit Impact and built the platform: job collection and processing, the React frontend and the MCP server. Extraction uses structured data, platform-specific rules, heuristics and language models.