ContentOS

AI-powered YouTube content production system with channel memory.

ContentOS is a CLI tool that helps you produce YouTube content with AI assistance. Each channel has its own "brain" that remembers what works, what doesn't, and evolves over time.

Designed for both humans AND AI agents.

Features

• Channel Brain - Persistent knowledge per channel (identity, themes, learnings)
• Context Surfing - AI can explore system progressively (contentos index)
• Kit Production - Generate scripts, prompts, and assets
• Research Agents - Scout (market research) and Analyst (audience insights)
• Health Checks - System diagnostics

---

Installation

Prerequisites

• Python 3.10+
• Ollama (optional, for AI features)
• YouTube API credentials (optional, for analytics)

Quick Install

# Clone the repository
git clone https://github.com/yourusername/contentos.git
cd contentos

# Install dependencies
pip install -r requirements.txt

# Run guided setup
python contentos.py setup

---

Using with AI Agents (Agentic IDEs)

ContentOS is designed to be operated by AI assistants in agentic IDEs like:

• Cursor (with Agent mode)
• Windsurf
• Cline/Claude Dev
• Antigravity
• Any terminal-based AI

Step 1: Onboard the AI

When starting a new session, tell your AI:

"Run python contentos.py boot to understand this system"

Or for progressive exploration:

"Run python contentos.py index to see the system structure"

Step 2: AI Explores Context

The AI can navigate like browsing a website:

python contentos.py index                    # Root sitemap
python contentos.py index brain              # Channel knowledge
python contentos.py index brain.themes.loop  # Specific theme
python contentos.py index kits.018           # Specific kit

Step 3: AI Creates Content

python contentos.py kit create "video_name" --theme loop

Step 4: AI Updates Brain

After research or analysis:

python contentos.py scout --keyword "topic"  # Adds to learnings
python contentos.py scan comments            # Adds audience insights

Workflow File (Optional)

Create .agent/workflows/contentos.md for your IDE:

---
description: ContentOS - YouTube production workflow
---

1. Run `python contentos.py boot` to understand the system
2. Check health: `python contentos.py health`
3. View brain: `python contentos.py brain show`
4. Create kit: `python contentos.py kit create "<name>" --theme loop`
5. After creating, read the prompt.txt and script.txt in the kit folder

---

Ollama Integration (AI Features)

ContentOS uses Ollama for local LLM features (scout, scan, suggestions).

How It Works

1. ContentOS connects to Ollama at http://localhost:11434
2. Auto-selects the best available model from priority list:
   • deepseek-v3.1:671b-cloud (best reasoning)
   • qwen3-coder:480b-cloud (good backup)
   • llama3.2:1b (local fallback)
   • mistral (old faithful)
3. Falls back gracefully if Ollama isn't running

Setup Ollama

# 1. Install Ollama
# Download from https://ollama.ai/

# 2. Pull a model
ollama pull deepseek-r1:8b

# 3. Enable in ContentOS
python contentos.py config enable llm_swarm

# 4. Test connection
python contentos.py test-crew

What Uses Ollama

Command	Uses LLM For
kit enrich	Extracting DNA ingredients from prompts
scout --keyword	Analyzing competitor videos
scan comments	Map-reduce sentiment analysis
strategy suggest	Content recommendations

Ollama Auto-Start

ContentOS v1.1+ automatically tries to start Ollama (ollama serve) when you run AI commands.

• Disable this: python contentos.py config disable ollama_autostart

---

Quick Start (Human Usage)

# Check active channel
python contentos.py channel status

# View channel knowledge
python contentos.py brain show

# Create a production kit
python contentos.py kit create "my_video" --theme loop

# Run health check
python contentos.py health

---

Architecture

contentos/
├── contentos.py          # CLI entry point
├── core/                 # Engine code
│   ├── brain.py          # Channel knowledge system
│   ├── llm.py            # Ollama integration
│   └── ...
├── commands/             # CLI commands
├── channels/             # Your channel data
│   └── {name}/
│       ├── brain/        # AI knowledge (state, themes, learnings)
│       ├── production/   # Active kits
│       └── analytics/    # YouTube data
├── protocols/            # Your prompt engineering knowledge
└── .contentos/           # System config

---

Commands Reference

Command	Description
setup	Guided first-time setup
boot	AI onboarding (dump all context)
index [path]	AI context surfing
channel list/use/create	Manage channels
brain show/set-theme/init	Manage channel knowledge
kit create/list/publish	Content production
kit enrich	Extract ingredients from prompt
kit link	Link YouTube videos/shorts
scout --keyword "x"	Market research (needs Ollama)
scan comments	Audience analysis (needs Ollama)
health	System diagnostics
config enable/disable	Feature flags

---

License

MIT