GitHub - algojogacor/mesin-cuan

██████████████████████████████████████████████████████████
█                                                        █
█   ███╗   ███╗███████╗███████╗██╗███╗   ██╗            █
█   ████╗ ████║██╔════╝██╔════╝██║████╗  ██║            █
█   ██╔████╔██║█████╗  ███████╗██║██╔██╗ ██║            █
█   ██║╚██╔╝██║██╔══╝  ╚════██║██║██║╚██╗██║            █
█   ██║ ╚═╝ ██║███████╗███████║██║██║ ╚████║            █
█                                                        █
█          C U A N   V I R A L   A R C H I T E C T      █
█                                                       █
█                                                        █
██████████████████████████████████████████████████████████

Automate the Fame. Master the Algorithm.

📖 🇮🇩 Baca dalam Bahasa Indonesia

The Vision

"This isn't a bot. It's an AI Video Architect that works around the clock — discovering trends, writing scripts, rendering cinematic footage, and uploading to YouTube, fully automated."

Mesin Cuan Viral Architect is a 24/7 AI-powered YouTube content production system. It's not merely a script; it's an autonomous video factory that understands the algorithm, detects trends before they peak, and executes every frame with cinematic precision.

Imagine having a complete production team — trend researcher, scriptwriter, voice actor, video editor, and upload manager — working without rest, without salary, without complaints. That's Mesin Cuan.

🤔 Why Mesin Cuan?

	Manual Editing	Hiring Editors	Other AI Tools	Mesin Cuan
Cost	Your time (8h/video)	$500-2000/month	$20-100/month	Free (self-hosted)
Speed	1 video/day	2-3 videos/week	1-2 videos/day	12+ videos/day
Quality	Skill-dependent	Variable	Generic AI feel	Dual AI cross-scored
Data Privacy	✅	❌ Shared with editor	❌ Uploaded to cloud	✅ 100% local
24/7 Operation	❌	❌	❌	✅ Autonomous

The secret sauce: Qwen + Ollama generate scripts in parallel, then cross-score each other to eliminate AI slop before it reaches production.

🏗️ Architecture

 Campaign    →  Topic Engine  →  Research Engine
 Schedule        (trends)          (web context)
                                        ↓
    ┌───────────────────────────────────┴───────────────────────────────┐
    │                    DUAL PARALLEL AI GENERATION                     │
    │  ┌──────────┐     cross-score     ┌──────────┐                    │
    │  │   QWEN   │ ←───────────────→ │  OLLAMA  │                    │
    │  │ self-host │                    │  local   │                    │
    │  └────┬─────┘                    └────┬─────┘                    │
    │       └──────────┬───────────────────┘                          │
    │                  ↓                                              │
    │         Script Quality Scorer                                   │
    │    (hook · pacing · curiosity · retention)                       │
    └──────────────────────┬──────────────────────────────────────────┘
                           ↓
    Hook → TTS → Footage → FFmpeg → Vision QC → Thumbnail → GDrive → ☁️ Uploader
    Engine   (F5)  (Pexels)  Render    (NVIDIA)   Intelligence  Queue      (Koyeb)

Auto-upload flow: Mesin Cuan renders videos locally → pushes to Google Drive queue → mesin-cuan-uploader (deployed on Koyeb/Railway) pulls from GDrive and uploads to YouTube with paced scheduling. This two-step architecture prevents YouTube from flagging bulk uploads from a single IP.

⚙️ 12-Engine Architecture

📡 Viral Loop Engine

Real-time trend intelligence combining YouTube Data API v3, Google Trends, and YouTube Search Suggestions — refreshed every 6 hours.

🧠 Dual Parallel AI Scripting

Qwen + Ollama generate scripts in parallel, then cross-score each other (anti-sycophancy). Highest-scoring script proceeds to production. Fallback chain: Ollama → Qwen → Groq.

🎯 Script Quality Scorer

Multi-dimensional scoring: hook strength, curiosity gap, pacing, emotional impact, retention prediction — scripts below threshold are auto-rewritten.

🔗 Hook Engine

Auto-injects opening hooks that maintain viewer retention in the first 3 seconds — driven by per-channel retention analytics.

📚 Memory Engine

Tracks previously produced topics — prevents content repetition and ensures variety across videos.

🔍 Research Engine

Automated web research before script generation — ensures factual accuracy and up-to-date context.

📺 Series Engine

Auto-generates multi-part content (Part 1, 2, 3...) with cliffhanger logic and cross-episode callbacks.

🌟 Neon Visuals

Cinematic FFmpeg renderer: glowing neon text, glassmorphism panels, cinematic letterbox + vignette + color grading.

🔊 Smart SFX Mixer

Niche-aware sound effects: Horror → heartbeat/thunder, Psychology → mind-tone/focus hum, Motivation → crowd cheer/stadium echo.

🖼️ Thumbnail Intelligence

AI-driven thumbnail text & style selection based on niche and topic — CTR optimization.

📊 OAuth2 Analytics

YouTube Analytics API v2 integration — per-channel retention dashboard, drop-off analysis, content optimization.

⚡ Pipeline Estimator

ETA prediction for multi-channel batch renders — accurate estimation before execution.

🛠️ Tech Stack

Layer	Technology	Role
Language	Python 3.11+	Pipeline orchestration
AI Primary	Qwen self-hosted + Ollama	Dual parallel script generation + cross-scoring
AI Quality	Script Quality Scorer	Multi-dimensional validation
AI Vision QC	NVIDIA API / Ollama Vision	Automated quality control
AI Fallback	Groq	Last-resort script generation
Text-to-Speech	F5-TTS · Edge TTS · Coqui	50+ multilingual voices
Rendering	FFmpeg 7.x	Video & audio composition
Footage	Pexels API · Pixabay API	B-roll footage + clip cache
SFX	Freesound API	Automated sound effects
Upload	YouTube Data API v3 (OAuth2)	Scheduled uploads
Analytics	YouTube Data API v3	Retention & insights
Trending	YouTube API · Google Trends	Real-time trend detection
Research	Perplexity API / Ollama	Accurate context before scripting
Storage	Google Drive API v3	Upload queue
Notifications	Telegram Bot API	Real-time alerts
Scheduler	APScheduler	Campaign automation

🚀 Installation & Setup

Step 1 — Clone & Environment

git clone https://github.com/algojogacor/mesin-cuan.git
cd mesin-cuan

python -m venv venv
source venv/bin/activate          # Linux / Mac
# venv\Scripts\activate           # Windows

pip install -r requirements.txt

Step 2 — Configure `.env`

Copy .env.example to .env and fill in your API keys:

cp .env.example .env

Step 3 — Asset Bootstrap & Auth

# Download SFX packs and initialize asset folders
python tools/setup_assets.py

# One-time OAuth2 YouTube authorization (opens browser)
python setup_auth.py

# Verify all connections and dependencies
python check_requirements.py

💻 Usage

# Run all channels according to campaign schedule
python main.py

# 🔑 Interactive API Key Setup Wizard (no manual .env editing)
python main.py --setup

# Run a specific channel only
python main.py --channel ch_id_horror

# Preview campaign schedule without rendering
python main.py --preview

# Full pipeline test without uploading to GDrive
python main.py --dry-run

# Skip QC Vision for faster iteration
python main.py --skip-qc

# Refresh per-channel retention analytics
python main.py --analytics

# Generate script only (no render)
python main.py --script-only

# Review hook quality without full render
python main.py --review-hook

# Specify profile: shorts or long_form
python main.py --profile long_form

# Enable debug logging
python main.py --debug

🔑 API Key Reference

🔴 Required — Pipeline won't work without these

Key	Provider	Purpose	Get It Here
`YOUTUBE_API_KEY`	YouTube Data API v3	Trending detection & topic discovery	Google Cloud Console
`PEXELS_API_KEY`	Pexels	B-roll footage & images	pexels.com/api
`GOOGLE_CLIENT_ID`	Google OAuth2	YouTube upload + Google Drive	Google Cloud Console
`GOOGLE_CLIENT_SECRET`	Google OAuth2	Pair for GOOGLE_CLIENT_ID	Google Cloud Console
`GOOGLE_DRIVE_FOLDER_ID`	Google Drive	Upload queue folder	Right-click folder → Share → Copy link
`TELEGRAM_BOT_TOKEN`	Telegram Bot API	Real-time notifications	@BotFather
`TELEGRAM_CHAT_ID`	Telegram	Notification target	@userinfobot

🟠 AI Provider — At least 1 required, recommend all for fallback

Key	Provider	Purpose	Get It Here
`GEMINI_API_KEY`	Google Gemini	Recommended — script generation & QC Vision	aistudio.google.com/apikey
`QWEN_API_KEY`	Qwen DashScope	Dual parallel AI script generation	dashscope.console.aliyun.com
`GROQ_API_KEY`	Groq	Fast fallback LLM	console.groq.com/keys

⚠️ Without an AI Provider, script generation won't run. Fill in at least 1.

🟡 Optional — Tool still works, specific features gracefully skip

Key	Provider	Purpose	Impact if Missing
`PIXABAY_API_KEY`	Pixabay	Footage fallback	No fallback footage
`PERPLEXITY_API_KEY`	Perplexity	Web research — accurate context	Uses local Ollama instead
`ANTHROPIC_API_KEY`	Anthropic	Last-resort AI fallback	Only 3-tier fallback chain
`NVIDIA_API_KEY`	NVIDIA NIM	Vision QC alternative (free)	QC via Gemini only
`ASSEMBLYAI_API_KEY`	AssemblyAI	Podcast transcription	No transcript feature
`ELEVENLABS_API_KEY`	ElevenLabs	Premium TTS voices	Fallback: Edge TTS (free)
`FREESOUND_API_KEY`	Freesound	Niche-based auto SFX	No sound effects
`CLOUDFLARE_ACCOUNT_ID`	Cloudflare	Google Trends scraping	Limited trending
`CLOUDFLARE_API_TOKEN`	Cloudflare	Pair for ACCOUNT_ID	—
`COVERR_API_KEY`	Coverr	Free stock video	Alternative footage source
`COVERR_APP_ID`	Coverr	Pair for COVERR_API_KEY	—

💡 Pro tip: Run python main.py --setup for an interactive setup wizard — no manual .env editing needed!

⚡ Quick Start (Minimal Mode)

Only need Ollama + FFmpeg — no API keys required:

# 1. Install Ollama and pull a model
ollama pull llama3.3:latest

# 2. Install FFmpeg
# Ubuntu: sudo apt install ffmpeg
# macOS: brew install ffmpeg
# Windows: https://ffmpeg.org/download.html

# 3. Run the API key setup wizard (or skip — Ollama-only mode)
python main.py --setup

# 4. Run with Ollama only (skip footage, upload, analytics)
python main.py --channel ch_id_horror --skip-qc

📜 License

MIT License — free to use, modify, and distribute for any purpose, including commercial.

See LICENSE for full text.

Built for creators who refuse to be limited by time.

[ 🤖 AI-POWERED ]  [ 🎬 CINEMATIC ]  [ 📡 VIRAL-READY ]  [ 🔄 24/7 AUTO ]

Name		Name	Last commit message	Last commit date
Latest commit History 47 Commits
.github		.github
assets		assets
config		config
docs		docs
engine		engine
templates/prompts		templates/prompts
tests		tests
tools		tools
.env.example		.env.example
.gitignore		.gitignore
CODE_OF_CONDUCT.md		CODE_OF_CONDUCT.md
CONTRIBUTING.md		CONTRIBUTING.md
CONTRIBUTORS.md		CONTRIBUTORS.md
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
README_ID.md		README_ID.md
SECURITY.md		SECURITY.md
check_requirements.py		check_requirements.py
cinematic.bat		cinematic.bat
docker-compose.yml		docker-compose.yml
edit.bat		edit.bat
edit.py		edit.py
edit_prompts.py		edit_prompts.py
main.py		main.py
podcast.bat		podcast.bat
requirements.txt		requirements.txt
scheduler.py		scheduler.py
setup.sh		setup.sh
setup_auth.py		setup_auth.py
setup_voices.py		setup_voices.py
tiktok.bat		tiktok.bat

Folders and files

Latest commit

History

Repository files navigation

Automate the Fame. Master the Algorithm.

The Vision

🤔 Why Mesin Cuan?

🏗️ Architecture

⚙️ 12-Engine Architecture

📡 Viral Loop Engine

🧠 Dual Parallel AI Scripting

🎯 Script Quality Scorer

🔗 Hook Engine

📚 Memory Engine

🔍 Research Engine

📺 Series Engine

🌟 Neon Visuals

🔊 Smart SFX Mixer

🖼️ Thumbnail Intelligence

📊 OAuth2 Analytics

⚡ Pipeline Estimator

🛠️ Tech Stack

🚀 Installation & Setup

Step 1 — Clone & Environment

Step 2 — Configure .env

Step 3 — Asset Bootstrap & Auth

💻 Usage

🔑 API Key Reference

🔴 Required — Pipeline won't work without these

🟠 AI Provider — At least 1 required, recommend all for fallback

🟡 Optional — Tool still works, specific features gracefully skip

⚡ Quick Start (Minimal Mode)

📜 License

About

Resources

License

Code of conduct

Contributing

Security policy

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Step 2 — Configure `.env`

Packages