An interlocking suite of open-source AI systems and research infrastructure for self-determining communities.
The HBoK is a comprehensive personal knowledge system — a structured, privacy-sovereign data representation of the whole human being. It holds each person's physical health data, psychological state (grounded in Self-Determination Theory), emotional and mental wellbeing, and spiritual/religious life.
Unlike commercial health platforms, the HBoK is 100% user-owned. There is no central server. No data leaves your device without your explicit consent. The standard is open, documented, and can be implemented by any community or individual.
Biometrics, health conditions, medications, fitness, nutrition
Autonomy, competence, relatedness — Self-Determination Theory scaffolding
Mood tracking, mental health history, therapeutic context
Values, beliefs, practices — defined entirely by the individual
All scores are advisory only. The OVNN never makes decisions for you.
The OVNN is an advisory scoring system that evaluates decisions, products, lifestyles, and systems across seven dimensions grounded in SDT and ecological economics. Scores range from -1.0 to +1.0. The system queries LanceDB for verified scientific facts and is trained through a human-verified feedback loop.
The OVNN never coerces. It never takes actions on your behalf. It provides information to support your own autonomous decision-making — reflecting our commitment to human self-determination above all else.
CommunityLLM is a standalone, politically neutral large language model that any community organization can deploy and own. It is trained on 35,000 expert-curated examples covering community organizing, mutual aid, governance, and peer support — generated and curated through Kestra-orchestrated pipelines.
This is a separate standalone project from OLS — designed so that worker cooperatives, tenant unions, community land trusts, and other organizations can run their own AI assistant without depending on commercial providers.
35,000 expert-curated examples across community organizing, mutual aid, governance, and peer support domains.
Runs locally. No cloud dependency. Any organization can host it on their own hardware using Ollama or vLLM.
Politically neutral by design. Useful for any community — progressive, conservative, or non-aligned. The community sets the values.
Only those actually affected by a decision participate in making it. Scales participation appropriately.
Decisions pass through graduated consent thresholds, not simple majority rule.
Designed for communities up to 15,000–20,000 people with federated council structures for larger scales.
A formal computational model for direct participatory democracy at scale. DirectDemocracyLLM operationalizes the theoretical frameworks of participatory governance into software — handling proposal writing, deliberation support, affectedness filtering, and consent-based decision processes.
Designed for communities of up to 15,000–20,000 people. The system augments human deliberation — it does not replace it. Every decision is made by the people it affects.
One World CSC provides accessible peer recovery support for mental health and substance use — via chat, phone, and video. Human-centered. Free to access. Built on the principle that communities heal communities.
OWCSC also serves as the fiscal sponsor for Optimal Living Systems, enabling OLS to operate with nonprofit status while building toward independent legal recognition.
Chat, phone, and video support. No insurance required. No cost to participants.
Mental health and substance use recovery. Peer-led, trauma-informed, and community-grounded.
Peer support — people with lived experience supporting others. Not clinical. Complementary to professional care.
1 Fine-tuned Model — Qwen 2.5 14B, specialized in research methodology
4 Domain RAG Collections — Sociology, Psychology, Environmentalism, Neuroscience in LanceDB
BGE-M3 Embeddings — Multilingual, multi-granularity vector representations
Kestra Orchestration — Automated research collection and processing pipelines
The Open Science Research Lab produces the scientific foundations for all AiH AI systems. We maintain four active domain research collections and operate a manifest-driven, multi-agent research collection pipeline for continuous knowledge acquisition.
All research is conducted under open science principles — transparent methodology, open datasets, reproducible pipelines, and public archiving through OSF and Zenodo. The ELSST social science thesaurus (OWL/SKOS) is integrated for semantic interoperability.
We welcome researchers, engineers, organizers, and domain experts. All skill levels. All backgrounds.
Get Involved