KAI OS
OperationalThe overall modular operating system -- agents, memory, orchestration, and security, developed as a separate real codebase.
See system status →Command Center
KAI Command Center
Explore the systems, agents, intelligence layers and experiments behind Project KAI. Generated August 11, 2026 from verified project data -- this is a system map, not a live feed.
System map available -- click any node to explore that layer.
System Status
What's actually running.
Status derived from verified project data, generated at build time -- not live telemetry. Not everything shown below is operational yet.
Content Engine
OperationalThe real YouTube production pipeline -- 12 videos produced, uploaded, and YouTube-verified to date.
See production status →Security Layer
OperationalDefault-deny capability authorization. 3 of 11 declared capabilities are currently active.
See security architecture →Memory Layer
OperationalAppend-only developer memory -- search and write are both real, active, LOW-risk capabilities today. A dedicated explainer page is still being built.
See the capability grants →Desktop Operator
In DevelopmentReal, extensively tested architecture -- not yet wired to observe or control anything live. The package's own module documentation explicitly states it does not yet activate automation.
See the command & control pipeline →Agent Framework
In DevelopmentThe security-gated orchestration layer is real and operational; most of the 13 declared agents themselves are still in development or planned.
See all agents →Website
OperationalThis Astro public site may use optional serverless or community integrations, but it exposes no live KAI runtime telemetry or private system access.
See how it changed over time →Public Status Boundary
Status without exposing the private runtime.
Internal health checks cover local services, credentials, schedules, logs, and account identity. Those details are intentionally not rendered publicly. The evidence-safe build snapshot remains available through System Status.
Public build snapshot
AvailableAggregate agent, security, production, and engineering facts are published without private endpoints or account identifiers.
Open System Status →Operational diagnostics
PrivateService endpoints, credential state, repair commands, scheduler detail, and internal logs remain inside the owner's environment.
YouTube Channels
Public channels. Private authorization state.
Project KAI operates two public media channels. Internal channel identifiers, authentication state, OAuth detail, and upload authorization evidence are not exposed on this public system map.
Media Network
PublicChannel names and public destinations are available from the Media Network hub; private control-plane data is not.
Open Media Network →Media Network
The ecosystem beyond production.
Two real YouTube channels and a growing website ecosystem -- honestly labeled by what's actually live versus still in development.
Media Network Hub
OperationalThe front door connecting both real YouTube channels and this website.
Visit the Media Network →Content Routing
OperationalA deterministic, explainable content router -- not an AI classifier.
See the router demo →AI News
In Development5 curated, sourced news items. No live feed -- a News Agent is Planned.
Read AI News →Future Radar
In Development8 tracked signals, every probability a labeled KAI Estimate.
See the Future Radar →Discover
OperationalSourced stories and KAI Documentaries -- curiosity-driven, not fabricated.
Explore Discoveries →Community
In DevelopmentComment UI live across the network -- no backend, nothing persisted yet.
See the ecosystem →Content Pipeline
How an idea becomes a video.
A public-safe view of the real production pipeline -- no credentials, no internal file paths, no implementation details beyond the stage names themselves.
Idea
→Research
→Classify
→Channel Routing
→Script
→Review (human)
→Production
→Video
→Website
→Short / Reel
→Analytics
Channel Routing uses the same deterministic router shown on the Media Network hub.
Comment Assistant -- Design Only
How KAI would help answer comments -- someday.
Not built. Not running. Shown here as the documented design -- see PROJECT_KAI_MEDIA_NETWORK_ARCHITECTURE.md for the full spec. Comments themselves have no backend yet, so there is nothing for this to analyze even if it existed.
Comment
↓KAI Analysis
↓Intent Detection
↓Tone Suggestion
↓Suggested Reply
↓Edit
↓Human Approval (mandatory)
↓Publish
Agent Matrix
The agent ecosystem -- real and conceptual.
Some of these are real, working capabilities. Most are future directions, honestly labeled Planned. Trading remains fully separate from this website by design -- see its card for why.
Conversation Layer
OperationalLocal conversation, multilingual routing, and spoken output with an offline-first fallback path.
See verified capabilities →Content Production
OperationalA linear production system with real media stages and mandatory human review before publishing.
See verified capabilities →Knowledge Brain
OperationalA callable, local keyword-search capability over the engineering report corpus; not a semantic universal memory.
Open Knowledge Brain →Developer Memory
OperationalAppend-only records for decisions, fixes, failures, and lessons, exposed through a bounded agent adapter.
See verified capabilities →Trading Intelligence
In DevelopmentProfessional research, strategy, broker, and risk intelligence under controlled paper-only validation.
Review paper-only trading →Desktop Operator
In DevelopmentPermission and approval architecture for future controlled desktop actions; unrestricted control is not active.
Review Desktop Operator →Quick Actions
Jump straight in.
Explore how KAI is being built -- some capabilities are still under development.
This page displays project data, not live telemetry -- no real-time activity, no simulated users, no fabricated metrics. Generated from the KAI OS repository during the latest website data-sync run. Not live. This page will be updated as capabilities change.