Open Cognition Commons · Foundation

Public infrastructure for AI governance knowledge.

OCC builds and operates open, public-interest AI infrastructure — knowledge bases, search APIs, and governance tooling — grounded in public-domain sources and run transparently as a foundation.

About the foundation → Browse knowledge → Governance model → API status →
Foundation

Mission

AI governance frameworks — the EU AI Act, NIST AI RMF, GDPR, UNESCO Recommendation — are public documents. They exist for everyone. But in practice, access to these frameworks is unequal: large organisations have legal teams and consultants; smaller ones often do not.

Open Cognition Commons exists to close that gap. We build open infrastructure that makes authoritative AI governance knowledge searchable, accessible, and useful — for developers, educators, auditors, and organisations of any size.

Public domain only

Every document in the knowledge base is verified public-domain. No copyright risk, no licensing complexity.

EU-hosted, privacy-first

All infrastructure runs on EU servers. No user tracking, no analytics, no data collection.

Open infrastructure

Ingestion pipeline, chunking logic, and search API are open-source and reproducible. No black boxes.

Foundation model

Governed as a foundation — analogous to Apache, Linux Foundation, or Mozilla — not as a commercial product.

Projects

What OCC builds

The foundation operates infrastructure projects in the public interest. Each project is independently useful and openly documented.

Active

Knowledge Infrastructure

A BM25 search and RAG API over 199 chunks from 10 public-domain AI governance documents. Available as a free, open API.

Explore →
Planned

Governance Schema Library

Open schemas, templates, and audit tools for AI governance — machine-readable, versioned, free to use.

Planned

Public AI Tooling

Open-source CLI and MCP servers for AI governance workflows. Integrable into existing development environments.

Project — Knowledge Infrastructure

AI Governance Knowledge Base

BM25 full-text search and RAG generation over public-domain AI governance documents. Free to use, no authentication required.

Indexed sources
ID Document Jurisdiction Type Chunks
 
Search — loading…
prohibited AI practices high-risk AI systems GPAI model obligations data minimisation human oversight conformity assessment
Enter a query to search across EU AI Act, NIST AI RMF, GDPR, and more.
Generate — RAG via loading…

Retrieves relevant chunks, then generates structured content via an open LLM. Cite the primary sources linked below each result.

Governance

How OCC is governed

OCC follows a two-phase governance model designed for long-term independence. All foundational decisions are documented and publicly accessible.

Phase 1 — Current

Build phase

  • Technical operations carried by Intelego GmbH
  • Decisions by OCC Tenant Lead (Zolo, Intelego)
  • All foundational decisions documented
  • No proprietary dependencies — Phase 2 compatible
  • Competent authority: LfDI Baden-Württemberg
Phase 2 — Planned

Independent Foundation

  • Independent legal entity (gGmbH / e.V. / Stiftung)
  • Own governance board
  • Open membership
  • All decisions public within 48h
  • Trigger: legal, organisational, technical foundations complete
OCC is modelled on public software foundations (Apache, Linux Foundation, Mozilla Foundation) — but for AI governance standards, open knowledge infrastructure, and public-domain AI resources. Intelego GmbH serves as technical carrier during the build phase, not as owner. The goal is full independence.
API

Open API

No authentication required. CORS enabled. Free to use.

GET /api/occ/search BM25 search over all chunks. Returns ranked results with snippets and source metadata.
qQuery string (required)
topMax results — default 3, max 10
sourceFilter by source ID, e.g. SRC-001
POST /api/occ/rag Retrieve top-K chunks via BM25, then generate structured content via LLM.
topicTopic or question (required)
formatexplanation | quiz | summary | glossary
languagede | en
top_kChunks to retrieve — 1–10, default 5
GET /api/occ/sources List all indexed sources with chunk counts and metadata.
GET /api/occ/health Service health. Returns chunk count, RAG availability, and LLM model.
Example
# Search
curl "https://opencognitioncommons.org/api/occ/search?q=high-risk+AI+systems&top=3"

# Generate
curl -X POST "https://opencognitioncommons.org/api/occ/rag" \
  -H "Content-Type: application/json" \
  -d '{"topic":"prohibited AI practices","format":"summary","language":"en"}'

# Health
curl "https://opencognitioncommons.org/api/occ/health"