diff --git a/oeps/OEP-0002.md b/oeps/OEP-0002.md
new file mode 100644
index 0000000..a2773be
--- /dev/null
+++ b/oeps/OEP-0002.md
@@ -0,0 +1,83 @@
+---
+oep: 0002
+title: Community & Governance Principles
+type: informational
+status: discussion
+authors: Rory Byrne <rory@rory.bio>
+created: 2025-12-07
+labels: governance
+---
+
+# Abstract
+
+This document defines the community and governance principles that guide how the Open Science Archive (OSA) project is run, who makes decisions, and how contributions are welcomed. For technical design principles, see OEP-0003.
+
+# Motivation
+
+OSA aims to be a neutral, open standard, and that requires explicit principles about how its contributors work together. Technical excellence is not enough, the protocol will not be adopted widely unless the governance process is open and participatory.
+
+# Specification
+
+## Principles
+
+### Open governance
+
+No single entity controls the protocol. Decisions are made transparently through the OEP process. Anyone can propose changes, and rationale is preserved for future maintainers.
+
+### Specification over implementation
+
+The protocol specification is separate from any reference implementation. Compliant software can be written in any language by any team. The spec is the source of truth, not the code. This means that pre-existing data platforms can adopt the OSA progressively, without changing their tech stack.
+
+### Community-owned semantics
+
+Vocabularies, validators, and schemas are maintained by domain experts, not by OSA protocol maintainers. The OSA project only provides mechanisms for creating and disseminating these domain-specific components.
+
+### Low barrier to contribution
+
+Rough ideas are welcome. The OEP process includes an "ideation" state for early-stage proposals that aren't fully formed. You don't need a complete specification to start a conversation.
+
+### Diverse perspectives
+
+OSA is not just a software project. Building infrastructure for scientific data requires input from many disciplines: researchers who understand data workflows, ML practitioners who consume datasets, software engineers who build systems, designers who shape user experiences, librarians who understand metadata, and domain experts who define quality standards. We actively seek contributors from all these backgrounds. Technical decisions have human consequences, and the best protocols emerge from diverse perspectives.
+
+### Pluralism over consensus
+
+Different communities may have different needs. The protocol should accommodate divergent approaches rather than forcing premature standardisation. Two communities can define incompatible vocabularies, that's fine. Alignment happens organically when it's useful, not by mandate.
+
+## Communication Channels
+
+### Zulip
+
+For informal discussion, questions, and early ideas. This is the best place to start if you're new to the project and want to get involved.
+
+[Join the Zulip chat](https://opensciencearchive.zulipchat.com/join/uuetdamcz55a2otndpkytkq5/)
+
+### OEP Pull Requests
+
+For formal proposals to change the protocol. See OEP-0001 for the process. Start here if you have a well-formed idea ready for community review.
+
+[opensciencearchive/oeps](https://github.com/opensciencearchive/oeps)
+
+# Rationale
+
+These principles are informed by the success and failures of other open standards:
+
+- **IETF/W3C**: Strong on open process, but high barrier to contribution
+- **PDB**: Successful domain-specific archive, but tightly coupled to one community
+- **Schema.org**: Good model for community-owned vocabularies
+- **ActivityPub/Fediverse**: Demonstrates federation without central control
+
+OSA aims to combine open governance with low barriers and explicit support for community divergence.
+
+# Backwards Compatibility
+
+N/A - this is an informational document.
+
+# Security & Privacy
+
+This document defines governance principles and has no direct security or privacy implications.
+
+# Open Issues
+
+- How should disputes be resolved when communities disagree?
+- Should there be a formal steering committee, or is the OEP process sufficient?