Public Documentation Hub

Understand the system without exposing the private runtime.

This documentation explains what Project KAI does, how authority flows through it, which capabilities are operational, and where the product is still being validated. It is a curated public layer—not a dump of internal reports, credentials, or account data.

01 · Overview

Start with the product, its verified status, and the evidence behind public claims.

Project Overview

Available

What Project KAI is, what local-first means here, and where the owner remains in control.

Read →

System Status

Available

A build-time snapshot of real public system counts—not a live uptime dashboard.

View →

Engineering Changelog

Available

Git-sourced engineering history with evidence attached to each public entry.

Read →

02 · Architecture

How authority, policy, specialized intelligence, tools, and memory fit together.

How KAI Works

Available

Observe, understand, plan, govern, act, review, and learn—with the action boundary stated clearly.

Explore →

Memory & Orchestration

Available

The governed path from owner intent to domain action, evidence, and multiple persistent memories.

Explore →

Agent Registry

Available

The 13 declared technical roles with underlying capability and autonomous-agent status separated.

Review →

Command Center

Available

The public system map and status-aware navigation across KAI's current surfaces.

Open →

03 · Capabilities

Operational and in-development domains, without converting roadmap work into product claims.

Verified Capabilities

Available

Conversation, memory, content, trading research, analytics, and governance status in one view.

Explore →

Voice & Conversation

Available

Local spoken responses, English/Hindi routing, fallback behavior, and the neural-voice roadmap.

Explore →

Knowledge Brain

Available

A real local keyword-search capability over the engineering report corpus—not semantic universal memory.

Open →

Desktop Operator

In Development

The approval-first control architecture and the honest boundary around live desktop execution.

Review →

04 · Content Pipeline

The working production system and its mandatory human publishing boundary.

Production Flow

Available

Topic, script, claims review, scenes, visuals, narration, captions, music, assembly, thumbnail, compliance, review, and publishing.

Explore →

Media Network

Available

The real channels, deterministic content router, and public discovery surfaces fed by the production system.

Explore →

Human Review

Available

Why production automation stops before publishing and how that boundary shapes the system.

Review →

05 · Trading Intelligence — Paper Only

Research and broker validation under explicit paper-only and risk-governed constraints.

Paper-First Trading

In Development

Market observation, strategy evaluation, governance, risk, and Alpaca Paper integration status.

Review →

Safety Boundary

In Development

A tightly bounded Alpaca Paper observation baseline is active; AI-directed and live-money execution remain disabled.

Review →

Research Lab

Available

Strategy research, simulation, backtesting, reporting, and the distinction between research and advice.

Explore →

06 · Safety Model

Human control, permission policy, stop controls, audit, privacy, and evidence.

Safety by Design

Available

Human checkpoints, paper-first validation, pause and emergency controls, and auditable operation.

Explore →

Security Architecture

Available

Default-deny capability authorization, revocation, kill switches, rate limits, and fail-closed policy handling.

Read →

Privacy

Available

What the public site collects, what it does not, and how public-safe disclosure is handled.

Read →

07 · Roadmap

Operational, next, and later work kept explicitly separate.

Public Roadmap

Available

Current foundations, next controlled milestones, and longer-term governed expansion.

Explore →

Contact

Available

The current owner contact path while dedicated project channels remain unresolved.

Contact →