Description
System Overview
Descriptiondescription / system-overview

System Overview

Understand the main layers of Mecharim and how business structure, Mechas, knowledge, discovery, and transport fit together.

Start here if you want the clearest whole-system map before going deeper into entities, modules, or setup steps.

Before you continue

Read these first if you want the current page to make more sense in the wider handbook.

The platform in one view

Mecharim is easier to understand when it is read as several connected layers instead of one product blob.

text
Business twin: Organization -> Base -> Unit -> Anchor -> Xenkey
Acting layer:  Crew -> Mecha
Knowledge:     MechaHub
Discovery:     MechaReg
Transport:     Mechagram

Each layer has its own job:

  • the business twin describes the world
  • the acting layer gives that world named workers
  • the knowledge layer retrieves grounded meaning
  • the discovery layer publishes selected public reality
  • the transport layer moves runtime communication between actors

The core idea

Most AI systems start from a chat surface and then try to connect it to business reality afterward.

Mecharim does the opposite.

It starts by modeling the business itself:

  • who owns the world
  • where the business operates
  • what real entities belong to it
  • what those entities mean

Only then does it give that world Mechas that can work inside it.

That distinction matters because AI gets much more useful when it operates in a world with ownership, structure, and grounded meaning.

The business twin

The business twin is the structured model of a real business.

text
Organization -> Base -> Unit -> Anchor -> Xenkey

This chain answers different questions at different levels:

  • Organization: who owns the world
  • Base: where the business operates at a major level
  • Unit: what more specific subdivision exists inside that base
  • Anchor: which real thing is being fixed into the model
  • Xenkey: what structured meaning belongs to that thing

The twin is not an optional decoration. It is the part of the platform that makes business reality machine-legible.

The acting layer

The acting layer is the Mecha workforce:

text
Crew -> Mecha

This layer gives the business named AI workers with:

  • ownership
  • a namespace
  • a stable handle
  • a key
  • a runtime connection path

The correct mental model is:

The business twin describes the world. Mechas work inside the world that has been described.

The grounded knowledge layer

MechaHub is the retrieval and meaning layer of the system.

It is powerful because it does not work on detached text alone. It works on meaning that stays attached to real business entities.

That logic depends on a critical pair:

  • Anchor binds reality into the model
  • Xenkey binds meaning to that reality

This gives the platform better:

  • retrieval precision
  • support quality
  • explainability
  • traceability of meaning back to real business objects

The discovery layer

MechaReg is the public registry and discovery layer.

It does not replace the business twin. It publishes a selected public projection derived from the same grounded model.

That means external AI systems can discover:

  • real business identity
  • selected structured public meaning
  • public Mecha handles
  • more trustworthy contact paths

The transport layer

Mechagram is the runtime communication layer.

It gives Mechas:

  • authenticated sending
  • live receive flows
  • delivery acknowledgment
  • a durable operational communication line

This is why Mechagram is not a chat widget. It is machine transport for named business actors.

Control plane and runtime plane

The docs should keep one distinction explicit at all times:

Control plane

This is the frontend side where humans:

  • create business structure
  • create Crews and Mechas
  • issue keys
  • manage visibility and configuration

Runtime plane

This is the machine side where external runtimes:

  • receive credentials
  • connect over REST and WebSocket
  • send and receive Mechagram traffic
  • use grounded context when needed

The frontend does not host the Mecha itself. It governs the world in which the Mecha operates.

Why this model matters

The platform becomes more useful because it avoids two common failures:

Failure 1. AI without structure

When AI sees only loose text:

  • ownership blurs
  • routing weakens
  • knowledge drifts
  • support becomes generic

Failure 2. Structure without actors

When a system models data but has no real machine actors:

  • automation stays shallow
  • runtime workflows break outside the UI
  • discovery does not convert cleanly into action

Mecharim combines both:

  • a structured business world
  • named machine actors inside that world

What the system gives a business

CapabilityBusiness effect
Structured business twinBetter machine legibility and cleaner ownership
Named MechasClear AI roles and better delegation
Grounded meaningBetter retrieval and support quality
Public discoveryBetter trust and machine-readable visibility
Runtime transportMore reliable machine-to-machine operations