# Dominir

> Dominir is the local-first agentic OS of company intelligence.
> A living company graph on hardware you own — built for people, agents, and Apple Silicon.
> Nothing is sent to a third-party cloud.

## Writing

- [What Is an Ontology?](https://www.dominir.com/writing/what-is-an-ontology): A plain-language definition — what an ontology is, why it is not a database or a folder of documents, and why agents need one.
- [The Company Ontology: What the Context Layer Looks Like When It's Built Right](https://www.dominir.com/writing/the-company-ontology): a16z is right that agents fail without context. But the context layer being described by most vendors is still too narrow. Here is what a properly structured company knowledge architecture actually requires.
- [Why We're Building Dominir](https://www.dominir.com/writing/why-we-are-building-dominir): Every major software category was made inevitable by a new form of coordination. Agentic AI has created the conditions for a new one. This is what we believe about it, and why we are building what we are building.
- [RAG Is Not a Knowledge Architecture](https://www.dominir.com/writing/rag-is-not-a-knowledge-architecture): Retrieval-augmented generation solves a search problem. Enterprise agents have a knowledge problem. These are not the same thing, and conflating them is why 70% of RAG implementations fail to meet their objectives.
- [The Palantir Problem](https://www.dominir.com/writing/the-palantir-problem): Palantir proved that enterprise ontology is the right architecture. They also proved it costs tens of millions to deploy. The thesis is correct. The delivery model is broken.
- [The Knowledge Drain](https://www.dominir.com/writing/the-knowledge-drain): Every year, $31.5 billion in organizational knowledge evaporates. Not through security breaches or system failures — through normal employee turnover. This is not a people problem. It is an infrastructure problem.

## Documentation

- [Introduction](https://www.dominir.com/docs/introduction): What Dominir is, what a CIM is, and where the boundary sits between the Kernel and the browser.
- [The Ontology (Petioles)](https://www.dominir.com/docs/ontology-petioles): How data is structured in Dominir — triplets, petioles, and Shadow Petioles explained.
- [Stores & Workspaces](https://www.dominir.com/docs/stores-workspaces): How the Kernel partitions data into independent stores, and what each store is for.
- [Record IDs](https://www.dominir.com/docs/record-ids): How Dominir identifies records — the ID format, what each segment means, and how provenance is tracked.
- [Human Oversight](https://www.dominir.com/docs/tool-policy-engine): How Dominir enforces human review of sensitive agent actions — the approval system, what triggers it, and what you control.
- [Clearance & Access Control](https://www.dominir.com/docs/clearance): How Dominir controls who and what can read or write sensitive records — the five-level clearance system.
- [Pulse](https://www.dominir.com/docs/pulse): Dominir's automation layer — scheduled triggers, incoming webhooks, and how agents run unattended.
- [Workflows](https://www.dominir.com/docs/workflows): How to run agents on tasks that require human oversight, multi-step coordination, and persistent sessions.
- [Pulse Triggers](https://www.dominir.com/docs/pulse-triggers): Cron and webhook triggers — ontology schema, Hydra expression context, dispatch modes, run storage, and full API reference.
- [Connecting MCP Servers](https://www.dominir.com/docs/mcp-servers): How to connect external MCP servers to Dominir so agents can use their tools alongside built-in capabilities.
- [Hydra SDK Reference](https://www.dominir.com/docs/hydra-sdk): Python surface to the Dominir Kernel — HydraStore, InstanceFrame, graph traversal, ontology writes, lock semantics.

## Optional

- [Full documentation (single file)](https://www.dominir.com/llms-full.txt)
- [Request Access](https://www.dominir.com/request-access)