MAIOS

MAIOS · component architecture

What is under the hood of MAIOS

Each component enters at a precise point in the path and makes the next step possible.

MAIOS organises the operating environment; MMK keeps intent, responsibility and continuity legible; Client Setup collects context; RepoKernel prepares the Project Kernel that accompanies work inside the project. Skills, meta-competences and Labs provide capability and verification without becoming owners of the work.

From context to work

Find the entry point for AI, bring the project to completion and support everyday management.

MAIOS connects the result you want to the project that delivers it and preserves what is needed to continue the work.

  1. 01Context

    Describe what you want to achieve in your domain.

  2. 02Setup

    Context becomes a reviewable configuration.

  3. 03Assistant

    Clarifies activities, sources, constraints and expected result.

  4. 04RepoKernel

    Validates, plans and prepares the project form.

  5. 05Project Kernel

    State, sources, competences and authority enter the project.

  6. 06Continuity

    The system preserves context and helps manage the work.

The MAIOS combination

Components and function.

Each element performs a recognizable task and passes organised context to the next one.

ComponentFunction in the system
MAIOSConnects the operating environment, components, projects, research and public distribution.
MMKCoordinates intent, owning front, faculties, authority, result and resumption.
MAIOS Client SetupCollects context and prepares the MAIOS Project Starter or a human consulting handoff.
RepoKernelValidates the handoff, builds the plan and stages the Project Kernel.
Project KernelGives the project a portable form with state, sources, competences, boundaries and reentry.
Seed and meta-competencesProvide reusable functions selected according to context and intent.
Labs and MiniAGIDevelop and test components, methods, workflows and new possibilities.
N0 — MMK DirectionCoordinates instances and handoffs when the selected environment provides this function.

What operation requires

AI builds with you. Human competence defines and verifies the result.

Environment

A coding agent that can work on files.

The initial path is designed for ChatGPT/Codex, Claude Code or another environment that can work on the project files. Ordinary browser chat is not equivalent to a project-aware environment.

Operator

Intent, domain knowledge and the ability to review.

You can begin by describing the desired result in natural language, but you must recognise what is correct and seek support when the project exceeds your competence.

Work

A project folder and one first verifiable result.

The system grows from a real context, available sources, clear boundaries and a proof that supports the decision to continue.

If you know what you want to achieve, you can prepare the context and first project.

Start with MAIOS Client Setup