Start with universal AI context: the other words describe its parts. The questions people ask first are on the home page.
Agent
An agent is knowledge plus a toolbox, set on a recurring stream of work: the weekly overview, inbox triage, meeting preparation, the signals you want to see. It only gets the actions you switched on for it. It answers by name, in your own AI tool, and leaves a trail of what it did and what it cost.
AI agents that cannot do what you did not allow
AI memory
AI memory is what an AI tool keeps about you between chats: your preferences, your work and what you told it before. It belongs to one person, inside one tool. A colleague's tool does not share it, and another tool does not read it. For a team, the shared version is one knowledge base that every AI tool reads.
AI memory for a whole team
AI second brain
An AI second brain is one person's notes and knowledge, kept where an AI tool can read them, so the AI answers from what that person knows and can help keep the notes in order. It is a second brain with an AI as its reader. It serves one person. For a team, the shared version is a company brain, with roles and a history of every change.
What is an AI second brain?
Company brain
A company brain is the shared knowledge of an organisation, kept where every person and every AI tool can reach it. Where a personal second brain organises one person's notes, a company brain holds who decides what, what was decided and how the work runs, with roles and the history of every change. Dienox calls this universal AI context.
What is a company brain?
Connection
A connection links Dienox to a system where your information lives: mail, calendar, documents, numbers, dashboards. Each one is a toolbox with explicit actions, and every action has a permission level. A connection works under the role of whoever connected it, and it belongs to one organisation.
Connector
A connector is how an AI tool reaches something outside itself. To connect Dienox, you add it in your AI tool as a connector, sign in and pick your organisation. It takes about a minute per person, and there is nothing to install. From then on, the tool reads your knowledge base when a question needs it.
Connect your AI tools to your company knowledge
Context engineering
Context engineering is the work of giving an AI model the right information for a task: the facts, the instructions and the tools it needs, and nothing it does not. For one person, it often means pasting background into a chat. For a team, it means keeping that background in one place that every AI tool reads, which is what universal AI context does.
Your AI starts every conversation from zero
Context layer
A context layer sits between an AI model and an organisation, so the model knows what it is working with. The words name two things. In analytics, a context layer describes data: what a table holds and how a metric is calculated. For the organisation as a whole, it holds people, decisions, products and processes. Dienox calls the second kind universal AI context.
What is universal AI context?
Custom connector
A custom connector is a connector you add to an AI tool yourself, instead of picking it from the tool's own directory. Claude uses the name for connectors added through remote MCP, and on its Team and Enterprise plans an owner adds one before members connect. In any AI tool that speaks MCP, a custom connector is how you add a source the tool does not list.
Connect your AI tools to your company knowledge
Edition
Dienox comes in two editions of the same platform. Self-serve runs in the Dienox cloud, hosted in the Netherlands, and is ready the same day. Enterprise runs in your own environment, in your Azure or your AWS, with a local model where nothing may leave. Your knowledge, your data and your access then never leave it.
Human-in-the-loop
Human-in-the-loop means a person approves an action before an AI agent takes it. It helps where an action matters. It is not enough on its own: someone who approves forty things a day stops reading them. In Dienox the first limit is what an agent can do at all, and approval is kept for irreversible actions, where it is called sign-off.
AI agents that cannot do what you did not allow
Knowledge base
In Dienox, the knowledge base is one living record of how an organisation works: who you are, what you sell, how you decide, how you sound and where everything lives. Every AI tool you connect reads it, and writes back what was decided. It keeps the history of every change, in open files you can take with you.
Knowledge. Tools. Agents.
Dienox has three parts: Knowledge (one knowledge base of how the organisation works), Tools (connections to its systems, with every action set to read, write or irreversible) and Agents (recurring work, with irreversible actions waiting for a person). An organisation starts in that order: first its knowledge, then its connections, then its first agent.
LLM wiki
An LLM wiki is a set of linked pages that an AI model writes and keeps up to date from your sources, so knowledge is worked out once and read many times. Andrej Karpathy described the pattern in April 2026, as a way to build a personal knowledge base. For a company, the same wiki needs roles, a history of every change and every AI tool reading it.
An LLM wiki for your whole company
MCP (Model Context Protocol)
MCP, the Model Context Protocol, is the open standard for connecting AI tools to data and systems. Anthropic introduced it in November 2024, and other AI providers have adopted it since. Any AI tool that speaks MCP can connect to Dienox. The specification is published at modelcontextprotocol.io.
MCP client
An MCP client is an AI tool that can reach other systems through the Model Context Protocol. Claude, ChatGPT, Copilot, Cursor, Windsurf and Gemini CLI are examples. Dienox works with any MCP client, not only the ones named here: if a tool speaks the protocol, it can read your knowledge base.
Permission level
Every action on every connection has one of three permission levels. Read looks at something without changing it. Write creates or changes something, such as a report or a draft reply. Irreversible cannot be taken back, such as mailing the board or making a payment. An irreversible action waits for a person with the right role, or is switched off.
AI agents that cannot do what you did not allow
RAG (retrieval-augmented generation)
RAG, retrieval-augmented generation, is a way to let an AI model answer from your own documents. The documents are cut into passages and kept in a search index. At every question, the passages that seem to fit are looked up and handed to the model. Nothing is worked out in advance, which suits very large or fast-changing collections.
An LLM wiki for your whole company
Role
A role decides what a person may do in an organisation on Dienox: administrator, editor or reader. Roles decide who may change what in the knowledge base, and an irreversible action waits for a person with the right role. People have a role per organisation, so the same person can be an editor in one and a reader in another.
Second brain
A second brain is a personal system for keeping notes, ideas and references outside your head, often in a note tool. The name comes from Tiago Forte's Building a Second Brain. When an AI tool reads those notes, people call it an AI second brain. It serves one person. A colleague cannot read it, and the AI tools of the team do not either. For an organisation, the shared version is a company brain.
What is an AI second brain?
Sign-off
Sign-off is a person approving an irreversible action before it happens. The action waits in Dienox for someone with the right role, and until someone signs off, it does not happen. An approval counts once: it covers that one action, not the next one like it. The next time, the agent asks again.
AI agents that cannot do what you did not allow
Single source of truth
A single source of truth is one place that holds the current version of a fact, so nobody has to ask which copy is right. For AI, it means every AI tool reads the same source instead of what each person pasted in. Correct something there once, and it is corrected for everyone who asks afterwards, in any tool.
What is universal AI context?
Toolbox
A toolbox is the set of actions one agent may take on a connection. For the numbers, that could be: read revenue and orders, build a report, make a payment. You switch each action on or off per agent. An action that is off is not in the agent's toolbox, so it cannot be called, however the agent is asked.
AI agents that cannot do what you did not allow
Trail
The trail is the record of every action: what was read and what was changed, by which person or agent, for whom and at what cost. Every change to a knowledge base is kept in its history, with who made it and when, and can be undone.
Universal AI context
Universal AI context is the knowledge of an organisation, kept in one place and read by every AI tool its people use, so every answer starts from the same facts. Claude, ChatGPT, Copilot and your own agents all know who decides what, what was decided, what you sell and how the work runs. Dienox is the platform that provides it.
What is universal AI context?