Skip to content

ADR-0009: Licensing

  • Status: Proposed
  • Date: 2026-10-25
  • Deciders: @konethegreat

Context

The project is public and open source, but "open" must not extend to other people's cultural heritage. We need a licensing approach that keeps the code broadly usable, the documentation shareable, and the knowledge under community control.

Options considered

For code: MIT (simplest), Apache-2.0 (patent grant, explicit contribution terms), AGPL-3.0 (strong copyleft against closed SaaS forks), GPL-3.0.

For docs: CC BY 4.0, CC BY-SA 4.0.

For knowledge: blanket CC licence (inappropriate), per-record terms.

Decision

  • Code: Apache-2.0. Maximises adoption (including by SA universities, government, NGOs) and gives contributors patent protection. Open point: an AGPL core would deter closed commercial forks that could exploit the data pipeline — raise as a Discussion before the first release.
  • Documentation: CC BY 4.0.
  • Community knowledge: no blanket licence. Governed per record by the consent's redistribution field and the contributor/community agreement. Public-tier data may be offered under a community-chosen licence (e.g. CC BY-NC-SA) plus TK Labels.
  • Open reference data (taxonomy, safety flags we curate): CC0 or CC BY, subject to upstream licences.
  • Contributions: DCO sign-off for code; a separate, plain-language Knowledge Contribution Agreement (to be drafted with counsel) for knowledge.
  • Trademark: the name "Reservoir" and logo to be governed so forks can't impersonate the community project.

Consequences

  • ✅ Clear separation between what is free to reuse and what is not.
  • ❌ Requires legal drafting (agreement, terms of use, API terms prohibiting bioprospecting uses).

Validation

Legal review in Phase 0; revisit code licence before first release.