Kaidera OS documentation

What Kaidera OS is

Kaidera OS is the local Kaidera control plane for AI worker teams. It runs a console, connects to Cortex-backed project memory, manages provider and model configuration, tracks run state, and dispatches work against project files you control.

What it gives you

A fresh install gives you a browser-based local console, a Cortex memory backend, an app database for settings and run state, a project registry, worker configuration, provider setup, live run streaming, and update plumbing.

• Run AI workers against a local project workspace.
• Keep decisions, handoffs, evidence, project state, and memory attached to the right project.
• Configure model providers without hard-coding keys into the product.
• Approve or hold work before autonomy is allowed to run.
• Update the console without replacing the project workspace or Cortex data.

How to use these docs

Start with Getting started if you are installing for the first time. Use First project when you are bringing a repository or workspace online. Use Workers, Cortex memory, Providers, and Architecture when you need deeper operating detail.

What is intentionally not baked in

Kaidera OS is project-agnostic. It does not ship with a customer project, a fixed worker team, or a hidden model key. Project identity, workspace root, roster, provider access, and autonomy controls are configured by the operator.

Current product boundary

The harness owns runtime startup, project registration, provider settings, Cortex API integration, worker orchestration, run-state streaming, package verification, and updates. Customer project content stays outside the core and is loaded through project data, Cortex memory, or extension modules.

Recommended path

Install Kaidera OS, open the console, confirm Cortex and the app database are healthy, add one project, add a lead worker, configure one model provider, run in propose mode first, then expand the roster and autonomy settings after review.