Open source and service

The code is open. What we derive from it is a service.

The code that wraps public endpoints is open source. What we derive from the public web, the curated sources, the index and the search over it, and running all of it for you, is a service.

Open source

  • Every API wrapper: the client, the command-line tool, its Claude Code skills, and its MCP server with the container image. Licensed under the AGPL-3.0-or-later: run it yourself for free; whoever runs a modified version as a network service publishes the changes.
  • A commercial license, in parallel, for building the wrappers into a proprietary product without the AGPL’s obligations.
  • The classifiers: the models that tell a Drucksache, a notice or a speech from a listing, a calendar or a login page. In the making

Offering both licenses needs a single copyright holder, so we don’t accept external code contributions. Bug reports and AGPL forks are welcome; each repository’s LICENSING.md and CONTRIBUTING.md have the details.

Service

  • Source dataset: the curated list of around 12,000 sources. How much of it will be open is still being decided. In the making
  • Search: the search index over the maintained corpus, as an API and on the web. In the making
  • Hosted MCP servers: the wrappers and the search tool, run and kept current for you. Finding and fetching from one endpoint, without running containers or following API changes yourself. In the making

The data belongs to its publishers

The wrappers fetch data; they don’t own it. Every upstream API has its own terms, which each tool documents in its DATA_LICENSE.md. We provide the tool, not the data.