Abfallkalender Deutschland

What Abfallkalender Deutschland does

Abfallkalender Deutschland MCP server retrieves waste collection schedules for German addresses. Depending on the authority, results can include residual waste, organic waste, paper, packaging, bulky waste, and hazardous-waste collection points. The service connects address resolution with the responsible municipal waste authority and then queries that authority’s calendar portal or data source.

The project maintains a registry covering 995 waste authorities from 150 source modules. Coverage is based on the authority list and available provider implementations; it does not guarantee that every address can be resolved automatically or that every portal will respond successfully.

How it works

The normal workflow starts with abfuhrtermine, which accepts an address and returns collection dates. The server uses Nominatim, an OpenStreetMap service, to identify a municipality and district, then searches the provider registry and resolves the arguments required by the selected authority.

Some portals require internal identifiers that are not directly present in an address. Resolver logic handles several of these cases, including Abfall.IO and AbfallPlus authorities, Hamburg, and Berlin. When the server cannot determine a value safely, it returns a follow-up request with possible choices instead of selecting a potentially wrong municipality or address.

Nominatim requests are rate-limited to one per second and cached under ~/.cache/abfall-mcp-server/. Collection responses include the relevant portal address, which can be useful for checking details such as bulky-waste or hazardous-waste requirements.

Setup and configuration

The Abfallkalender Deutschland MCP server runs from a repository checkout rather than as a standalone installed package. Clone the repository with its submodules, enter the project directory, install dependencies with uv sync, and optionally run the test suite with uv run pytest. The checked-in data/providers.json registry allows the server to start immediately. After updating the data-source submodule, rebuild the registry with uv run python scripts/build_registry.py.

For local use, configure an MCP client to run uv --directory /path/to/abfall-mcp-server run abfall-mcp-server over stdio. The server can also be started as an HTTP service with uv run abfall-mcp-server --http --host 127.0.0.1 --port 8000.

A hosted Streamable HTTP endpoint is available at https://abfall-mcp.alpaycelik.dev/mcp, with a health endpoint at /health. For higher-volume HTTP use, the documentation recommends placing a cache in front and configuring a separate Nominatim instance through ABFALL_MCP_NOMINATIM_URL.

Tools and capabilities

• abfuhrtermine: resolve an address and retrieve collection dates.
• finde_traeger: search authorities by place or company name without geocoding.
• traeger_details: inspect the arguments expected by a selected authority.
• abfuhrtermine_fuer_traeger: query dates directly for a known authority.
• abdeckung: report the number of covered authorities and data sources.

The abfall://traeger resource exposes the complete authority list. Tool names and responses are in German.

Limitations and notes

Abfallkalender Deutschland MCP server depends on external municipal portals, Nominatim, and provider-specific behavior. In an August 2026 smoke test of 21 German addresses, 48% returned dates directly, 14% required a follow-up detail, and 38% returned no dates. Failures can result from missing resolver logic, differing portal spellings, outages, rate limits, or malformed responses.

Some known unresolved cases require additional identifiers, including standort in Dresden, idHouseNumber in Leipzig, and streetnr in Stuttgart. HTTP use also creates requests to authority portals and Nominatim, so caching and responsible request volumes matter.