Welcome

The Sirens promise convenience. The giants own the gates. The sea is faster if you let someone else steer. OdyssAI is for those who would rather take the helm.

OdyssAI builds CoeOS: frontier-grade AI at open-source prices, where confidential work stays where you decided. Underneath sits OdyssAI-X, our own inference engine for Apple Silicon, for everything that runs on your premises.

This page tells you what the pieces are and which one you need.

The pieces

Piece	What it is
CoeOS	The product. A client, where you work, and the CoeOS box, the router that sends each request to the model proven best at that skill. Two ways to work: the web app, and CoeOS Desktop, the agentic harness on your machine (macOS, Windows, Linux). Run for you as a managed service, one dedicated instance per client.
OdyssAI-X	The engine. Distributed MLX inference on Apple Silicon: single node, multi-node over RDMA (Thunderbolt 5, 120 Gb/s), and replica mode with continuous batching. OpenAI- and Anthropic-compatible API. It serves the Hybrid and Local levels of CoeOS.
TMB	The benchmark of The Monocle Bear: 42 tests, 17 axes, every model scored axis by axis. The routing tables of CoeOS come from it.

CoeOS web app  ·  CoeOS Desktop
              ↓
        the CoeOS box            one address, one axis per request, one class per request
              ↓
  cloud providers  ·  OdyssAI-X on your Apple Silicon

Three levels

	Open	Hybrid	Local
Confidential goes	your choice: no distinction, or EU providers under DPA + ZDR	to a model hosted on your premises	stays on site
The rest goes	to the cloud, best model per axis	to the cloud, best model per axis	stays on site

Everything is a setting: providers, contracts, what runs at home. Details: What CoeOS is.

Which one do you need

You want	Reach for	Read
A team AI at a fraction of the price of closed models, with confidential work kept where you decide	CoeOS, as a managed instance	What CoeOS is
An agent with full access to your machine, routed by CoeOS	CoeOS Desktop	CoeOS Desktop
Frontier open-weight models served on your own Macs	OdyssAI-X	Install the engine

Product documentation

Using CoeOS day to day (first steps, agents, confidentiality, memory) is documented on docs.coeos.io. This guide covers how the pieces fit together, the engine, and the benchmark.

Read first

• Getting started — three ways in.
• What CoeOS is — the levels, the two harnesses, how a request flows.
• OdyssAI-X overview — the engine’s modes and its API.
• TMB Scoreboard — the benchmark behind the routing table.