P20 / PROJECT

OpenDataMCP

Danish language resources exposed as local Model Context Protocol tools for word verification, word games, spelling assistance and grammatical lookup.

Conceptual illustration of a local database connected across a bridge to an abstract AI network and blank dictionary tiles.

Connect language data to an assistant

OpenDataMCP distributes Model Context Protocol servers that make structured Danish language resources available to compatible AI assistants and agent environments. Instead of asking a language model to recall every word form, a client can use a dedicated tool backed by a local language database.

At the time of review, the repository lists one active server: COR, based on Det Centrale Ordregister and the Danish Language Council. The documented release includes server code, a pre-indexed SQLite database, tests and client setup guidance.

The COR toolset

  • Verify Danish words against the register's indexed word forms.
  • Search language data for word-game tasks such as patterns, anagrams, Wordle and Scrabble.
  • Support spelling-related lookup and grammatical analysis through explicit tools.
  • Use a dataset described by the repository as containing more than 386,000 word forms.

A local, structured connection

The server communicates through the documented JSON-RPC/stdio setup. A compatible MCP host starts the Python entry point and exchanges tool requests and responses with it, while SQLite supplies the indexed data locally.

This is a different access pattern from downloading a CSV list into a script. The interface exposes language operations as tools, making them available inside an assistant workflow without turning the repository into a general-purpose public HTTP service.

Packaging and verification

The repository organizes each server in its own directory. The COR directory includes the implementation, test suite and server-specific README, while a release manifest records file integrity information.

The quick-start documentation describes cloning the repository, installing the MCP dependency and configuring a compatible client to invoke the local entry point. Exact configuration locations and paths depend on the selected client and installation, so the repository guide is the reference for setup.

How it relates to OpenDataSets

OpenDataSets provides static processed files for direct reuse, while OpenDataMCP packages a tool-oriented route into related Danish language resources. They serve different consumers: ordinary data-processing programs on one side, compatible assistant toolchains on the other.

The COR distribution is identified as CC0 1.0 in the repository documentation, with attribution to the original language-data source. This description concerns the published server package; it does not imply that a hosted MCP endpoint has been deployed. The featured image is conceptual artwork rather than a client screenshot.

Explore the repository

Read the source code and project documentation for the latest implementation details and setup instructions. This description is based on the public repository reviewed on October 29, 2026.

Project at a glance

License
CC0 1.0, as documented for the COR MCP distribution
Requirements / platforms
Python, the documented mcp[cli] dependency and a compatible MCP client. The COR server uses a local SQLite language database.
Technologies
JSON-RPCModel Context ProtocolPythonSQLitestdio
Project categories
AI toolsLanguage resourcesOpen data