CNOS — Container Native OS

Dashboard

System at a glance

📦 Containers

⚙ Scheduler

running

23 completed · 1 failed

💻 CPU

--%

Load --

📌 Memory

--%

💾 Storage

--%

⏳ CNOS Uptime

OS: --

📶 WiFi

Checking...

🔋 Battery

📦 Container Overview

Name	Image	Type	State	Uptime	Actions

📈 Resource Monitor

Name	CPU	Mem Usage	Mem %	Net I/O	Block I/O	PIDs

Containers

Name	Image	Type	State	Priority	Resources	Uptime	Restarts	Actions

Deploy Container

Name *

Image *

Command

Namespace

Type

Priority

Ports

Environment

Build Context

CPU Shares

Memory (MB)

Network

Auto-restart Auto-upscale

🚀 Quick Deploy Templates

🌐

Nginx

Web server · nginx:alpine · :80

⚡

Redis

In-memory cache · redis:alpine · :6379

🗃

PostgreSQL

SQL database · postgres:15 · :5432

💚

Node.js 20

JavaScript runtime · node:20 · :3000

🐍

Python 3.12

Python runtime · python:3.12 · :8000

💾

MySQL 8

SQL database · mysql:8 · :3306

System Information

💻 Hardware

Hostname	--
OS	--
Kernel	--
Architecture	--
CPU	--
Cores	--
Temp	--

📈 Resources

CPU--%

Memory--%

Disk--%

Load: -- · Uptime: --

🔴 GPU / CUDA

GPU

CUDA

VRAM

Toolkit

Driver: -- Devices: 0 Temp: -- Utilization: --

🔋 Battery

Status

Charge

Power

AC Power

WiFi

Status

SSID

Signal

IP Address

MAC: -- Gateway: -- DNS: --

📶 Available Networks

📶

Click Scan Networks to discover WiFi

🔗 Connect to Network

SSID

Password

USB Devices

Connected Devices

🔌

💾 USB Storage

💾

🔄 Mount / Unmount

Device

Mount Point

Docker Images

Loaded Images

🗃

💾 Export Image

Image

Save To (optional)

File Manager

📁

📦 Package Manager

Install, remove and update system packages

Search & Install

Search Packages

Quick Install

Package Name

Source

Upgradable Packages

Package	Available Version	Source	Details	Action
Click "Check" to scan

Installed Packages

Package	Version	Source	Size	Action
Loading...

Output Log

Ready.

💻 Linux Shell

Interactive bash terminal

bash — cnos

Welcome to CNOS Linux Shell Type commands below. Use ↑↓ for history, Tab for suggestions. ────────────────────────────────────────────────────

cnos:~$

⌨ Enter = run↑↓ = historyCtrl+C = cancelCtrl+L = clearTimeout: 30s per command

Terminal

Command console

cnos $ Welcome to CNOS Terminal Type a command to get started. Examples: cnosctl status cnosctl list docker ps ls -la /etc/cnos/ df -h

Tip: Up/Down for history. Commands run with 10s timeout. Use cnosctl for container management.

Bridge Links

Bridge Name

Container A

Container B

Bridge	Container A	Container B
Loading...

REST API Reference

Method	Endpoint	Description
GET	/api/v1/status	System status overview
GET	/api/v1/health	Health check
GET	/api/v1/containers	List all containers
POST	/api/v1/containers	Create a new container
GET	/api/v1/containers/{name}	Get container details
DEL	/api/v1/containers/{name}	Delete a container
POST	/api/v1/containers/{name}/start	Start container
POST	/api/v1/containers/{name}/stop	Stop container
POST	/api/v1/containers/{name}/restart	Restart container
GET	/api/v1/containers/{name}/logs	Container logs
GET	/api/v1/containers/{name}/stats	Container metrics
POST	/api/v1/containers/{name}/hotswap	Hot-swap container
POST	/api/v1/containers/{name}/archive	Archive container
GET	/api/v1/containers/stats	All container metrics
GET	/api/v1/system/stats	System resource metrics
GET	/api/v1/system/info	Full system information
GET	/api/v1/system/wifi	WiFi connection status
GET	/api/v1/system/wifi/scan	Scan WiFi networks
POST	/api/v1/system/wifi/connect	Connect to WiFi
POST	/api/v1/system/wifi/disconnect	Disconnect WiFi
GET	/api/v1/system/battery	Battery status
GET	/api/v1/system/usb	List USB devices
POST	/api/v1/system/usb/mount	Mount USB storage
POST	/api/v1/system/usb/unmount	Unmount USB storage
GET	/api/v1/files/list?path=/	List directory
GET	/api/v1/files/read?path=...	Read file
GET	/api/v1/files/download?path=...	Download file
POST	/api/v1/files/upload	Upload file (multipart)
POST	/api/v1/files/write	Write/create text file
POST	/api/v1/files/mkdir	Create directory
POST	/api/v1/files/rename	Rename/move file
POST	/api/v1/files/delete	Delete file/directory
POST	/api/v1/cli	Run shell command
POST	/api/v1/shell	Execute bash command (interactive shell)
Package Manager
GET	/api/v1/packages	List installed packages
GET	/api/v1/packages/search?q=...	Search available packages
POST	/api/v1/packages/install	Install a package (apt/snap)
POST	/api/v1/packages/remove	Remove a package
POST	/api/v1/packages/update	Update package lists (apt update)
GET	/api/v1/packages/upgradable	List upgradable packages
POST	/api/v1/packages/upgrade	Upgrade package(s)
GET	/api/v1/scheduler/status	Scheduler status
GET	/api/v1/bridges/{name}/links	List bridge links
POST	/api/v1/bridges/{name}/links	Assign bridge link
DEL	/api/v1/bridges/{name}/links	Remove bridge link
AI Engine
GET	/api/v1/ai/status	AI system status (Ollama, models, auto-learn)
POST	/api/v1/ai/chat	Chat with AI (routed inference)
POST	/api/v1/ai/feedback	Submit feedback for auto-learning
GET	/api/v1/ai/models	List all AI models
POST	/api/v1/ai/models/pull	Pull model from Ollama registry
GET	/api/v1/ai/models/{id}	Get model details
DEL	/api/v1/ai/models/{id}	Delete model
GET	/api/v1/ai/routes	List model routes
POST	/api/v1/ai/routes	Create/update route
GET	/api/v1/ai/train/jobs	List training jobs
POST	/api/v1/ai/train/jobs	Create training/fine-tune job
POST	/api/v1/ai/train/cancel	Cancel training job
POST	/api/v1/ai/train/build-dataset	Build dataset from interactions
GET	/api/v1/ai/datasets	List datasets
POST	/api/v1/ai/datasets	Create dataset
POST	/api/v1/ai/datasets/upload	Upload dataset file
POST	/api/v1/ai/datasets/append	Append rows to dataset
GET	/api/v1/ai/pipelines	List AI pipelines
POST	/api/v1/ai/pipelines	Create pipeline
POST	/api/v1/ai/pipelines/run	Execute pipeline
GET	/api/v1/ai/eval	List evaluation jobs
POST	/api/v1/ai/eval	Create model evaluation
AI Agents (OpenClaw)
GET	/api/v1/ai/agents	List all agents
POST	/api/v1/ai/agents	Create new agent
GET	/api/v1/ai/agents/{id}	Get agent details
DEL	/api/v1/ai/agents/{id}	Delete agent
POST	/api/v1/ai/agents/{id}/run	Execute agent with input
GET	/api/v1/ai/tools	List available OpenClaw tools
POST	/api/v1/ai/tools	Create a tool
POST	/api/v1/ai/tools/{id}/run	Execute a tool with JSON input
GET	/api/v1/ai/skills	List available OpenClaw skills
POST	/api/v1/ai/skills	Create a skill
POST	/api/v1/ai/skills/{id}/run	Execute a skill prompt
GET	/api/v1/ai/chat/sessions	List persisted AI chat sessions
POST	/api/v1/ai/chat/sessions	Create a persisted chat session
POST	/api/v1/ai/chat/sessions/{id}/messages	Send a message within a session
POST	/api/v1/ai/workflow	Run multi-step agent workflow
GET	/api/v1/ai/agents/history	Agent execution history

🔬 System Diagnostics

Live health probes — all 12 subsystems

Subsystem Probes

Subsystem	Status	Message	Details
Click "Run Now" to probe all subsystems

⚠ Active Issues

Runtime

Goroutines
—

Heap
—

Duration
—

🤖

CNOS AI Chat

● Connecting...

🤖

How can I help you today?

Ask me anything — I can help with code, analysis, system administration, and more. Select a model and routing above to customize responses.

Enter to send · Shift+Enter for new line · Powered by Ollama

🧟

CNOS Agent◆UI V3

NEURAL LINK

0 msgs

● Initializing neural link...

Ln 1, Col 1plainUTF-8

Settings

Project Folder

No project folder selected.

Files

No project folder set.

Runtime

Backend

Agent

Model

Iterations

Streaming

Context

BackendAuto

AgentSystem

ModelDefault

Iters10

StreamOn

Context0 msgs

Chat Controls

Modes

Quick Actions

Pinned Prompts

Recent Runs

Agent Memory

Key-value store persisted for this session. Agent can save/recall facts automatically.

🎨 Design Intelligence Skills

Skills are auto-injected into agent context when relevant keywords are detected. Data from ui-ux-pro-max (67 styles, 161 palettes, 57 fonts) and awesome-cursorrules (178 coding rules).

Goal

→

Plan

→

Act

→

Observe

→

Reflect

→

Evaluate

🧟

CNOS Neural Agent

Agent-first workspace active. Give a goal and I will plan, edit, run, and verify.

Use modes above for Code, Autonomous, Reasoning, and Debug workflows.

Click the mic to speak

✎ No file open Auto-context

Output

👁 Review Agent Changes

⚙ Advanced Model Parameter Tuning

Select Model to Tune

Model

Apply Preset

🎓 Training & Fine-tuning

Auto-Learning Feedback Loop

🧪 Model Distillation

Create compact 3-bit TurboQuant models

🔗 Open Full Studio

📈 Status

Idle

No active distillation jobs

⚡ Quick Start

Create a compact student model from a teacher model using knowledge distillation + 3-bit quantization.

🧠 Models

No .cnosmodel files found

📚 Student Templates

💬

Tiny (0.5B)

Ultra-compact for edge

~300MB quantized

🧠

Small (1.5B)

Balanced performance

~900MB quantized

⭐

Medium (3B) ⭐

Recommended

~1.8GB quantized

💡 About TurboQuant 3-bit Quantization

TurboQuant uses 3-bit quantization to compress models by ~10x while maintaining quality:

8 values → 3 bytes packing (3 bits per weight)
Block-wise quantization with scale/zero-point per 32 elements
Custom .cnosmodel format optimized for CNOS inference
KV Cache + RoPE for efficient transformer inference

📊 AI Data Hub

Datasets

Pipelines

Model Routes

Model Evaluations

🧙 AI Agents

OpenClaw Agent Orchestration

Ollama

...

Trainer

...

OpenClaw

...

Registered Agents

Workflow Runner

Execution History

No history yet

🛠 AI Tools

Create, run, and manage OpenClaw tools

Create Tool

Type

Builtin

Method

💬 Tool Execution Result

Run a tool to inspect its output.

Registered Tools

✨ AI Skills

Create reusable prompts and executable skills

Create Skill

Execute against model when run

💬 Skill Execution Result

Run a skill to inspect rendered prompts and model output.

Registered Skills

💬 AI Sessions

Persisted chat sessions backed by OpenClaw agents or direct models

Create Session

Selected Session

Select a session to inspect details.

Sessions

Conversation

No session selected

💬

Session Chat

Create or select a session to begin chatting.

Advanced inputs (tool/skill JSON)

🧠 Intelligence — Core Algorithms

8 algorithms keeping the AI engine alive, fast, and self-improving

📦 Cache

Hit rate: --

📈 Patterns

Events: --

💡 Prefetch

Transitions: --

❤ Pulse

Uptime: --

📦 Response Cache (LRU + TTL)

Avoids redundant LLM inference by caching recent responses. Uses SHA-256 cache keys, LRU eviction, and configurable TTL.

📈 Pattern Engine

Detects recurring patterns, bursts, and periodic behaviors using frequency counting, burst detection, and coefficient of variation analysis.

💡 Prefetch Engine (Markov Prediction)

Learns query sequences via Markov chain transitions and pre-generates answers for predicted next queries. Triggers at ≥ 50% confidence.

🔗 Knowledge Graph (BFS Reasoning)

Graph-based reasoning with typed nodes and weighted edges. Uses BFS traversal, shortest path finding, and semantic search over concepts.

⚙ Adaptive Optimizer (EMA Tuning)

Auto-tunes temperature and max_tokens per task type using Exponential Moving Average (α=0.2). Classifies queries into: analysis, code, training, system, improvement, chat.

⚠ Anomaly Detector (Z-Score)

Real-time anomaly detection using windowed z-scores. Thresholds: info ≥ 2.0σ, warning ≥ 2.5σ, critical ≥ 3.5σ.

⏳ Decay Scheduler (Ebbinghaus Curve)

Implements forgetting curves: retention = e^{(-t/halfLife)}. Knowledge decays over time unless used. Half-life: 24h. Prune threshold: 5% retention. Verified entries are immune.

Half-life	24 hours
Prune threshold	5% retention
Loop interval	1 hour
Usage boost	1 + count × 0.1

❤ Pulse Monitor (Heartbeat)

Tracks heartbeat of all engine subsystems. A component is dead if errorCount > 5 and lastBeat > 10 minutes ago.

🔄 Background Loops & Data Flow

Loop	Interval	Responsibilities
analysisLoop	2 min	System analysis, pulse beats, anomaly scoring, pattern events
learningLoop	3 min	Extract knowledge, populate graph from learned entries
patternLoop	5 min	Run pattern detection, build graph edges from patterns
decayLoop	1 hour	Apply forgetting curve, prune stale knowledge, evict cache
pulseLoop	2 min	Health checks (Ollama, Trainer), save knowledge graph

🚀 Ask Pipeline — Query Flow

❶ Pattern Event → ❷ Prefetch Transition → ❸ Prefetch Cache? → ❹ System State → ❺ Graph Enrich → ❻ Optimizer Params → ❼ Cache Check? → ❽ LLM Inference → ❾ Cache + Metrics → Async Prefetch

📑 API Endpoints

Method	Endpoint
GET	/intelligence/algorithms
GET	/intelligence/cache
GET	/intelligence/patterns
GET	/intelligence/pulse
GET	/intelligence/anomalies
GET	/intelligence/graph
GET	/intelligence/status
POST	/intelligence/ask
POST	/intelligence/analyze
POST	/intelligence/suggest
GET	/intelligence/knowledge
GET	/intelligence/insights
GET	/intelligence/plans

>>_ CLI Commands

Command	Description
intel algorithms	Full algorithm dashboard
intel cache	Inference cache statistics
intel patterns	Detected system patterns
intel pulse	Component health & liveness
intel graph	Knowledge graph statistics
intel anomalies	Anomaly detection metrics
intel status	Engine status overview
intel ask <q>	Ask intelligence a question
intel chat	Interactive chat session
intel analyze	Trigger system analysis
intel suggest	Get improvement suggestions
intel knowledge	Show learned knowledge

📊 Mathematical Foundations

LRU Cache

key = SHA256(model|temp|msgs)[:24]
evict = oldest in order list

Z-Score Anomaly

z = |x - μ| / σ
info ≥ 2.0σ  warn ≥ 2.5σ  crit ≥ 3.5σ

Ebbinghaus Decay

retention = e^(-t / halfLife)
boost = 1 + usedCount × 0.1

EMA Optimizer

EMA = old*(1-α) + new*α
α = 0.2

Periodicity (CV)

CV = σ / μ
periodic if CV < 0.3

Prefetch Markov

confidence = best / total
trigger if > 50%

Dashboard

Containers

Deploy Container

System Information

💻 Hardware

📈 Resources

WiFi

USB Devices

Connected Devices

Docker Images

Loaded Images

▶ Run Container

File Manager

📦 Package Manager

💻 Linux Shell

Terminal

Bridge Links

REST API Reference

🔬 System Diagnostics

How can I help you today?

CNOS Neural Agent

⚙ Advanced Model Parameter Tuning

🎓 Training & Fine-tuning

⚙ Advanced Training Configuration

🧪 Model Distillation

📊 AI Data Hub

Datasets

Pipelines

Model Routes

Model Evaluations

🧙 AI Agents

Registered Agents

Workflow Runner

Execution History

🛠 AI Tools

Registered Tools

✨ AI Skills

Registered Skills

💬 AI Sessions

Sessions

Conversation

Session Chat

🧠 Intelligence — Core Algorithms

🧩 Cognitive Core

🌐 Web Intelligence

🔍 Web Search

📄 Fetch Web Page

📡 RSS/Atom Feeds

Add Feed

🔬 AI Research

📦 Training Dataset Discovery

📋 Fetch History

🕷 Start Website Crawl

📈 Active Crawls

📋 Crawl History

⚙ CNOS Kernel

🧙 Agent Drivers

Run Agent Task

🕑 AI Scheduler

🧠 Knowledge Memory

Store Memory

📩 Agent Communication Bus

🔒 Security Policy Engine

💻 Hardware Abstraction Layer

🤖 Model Runtime Manager

⚙ Kernel Syscalls

Execute Syscall

⚡ Autonomous Reasoning Loop

🔄 Observe → Reason → Plan → Execute → Learn

🤖 Autonomous Agents

🔔 System Events

🔁 Cycle History

💬 Intent Engine

🛡 Safety Layer

🧬 Living Digital Organism

🎮 CNOS Control Center — Prefrontal Cortex

🗣 Feedback & Voice

🧠 Learned Preferences

🔗 System Integration — Neural Backbone

🌐 Subsystem Map