Star on GitHub Clone Repository
MINA × GITHUB // OPEN SOURCE MEDIA OPERATING SYSTEM

The Autonomous
Media Engine.

A low-latency, deterministic media operating system designed for instant AI video editing, interactive presentations, and prediction market broadcasts.

Clone from GitHub Quickstart Guide →
⌨️ STUDIO SWITCHBOARD — INTERACTIVE HARDWARE ACCELERATOR
PRESET: LOW-PROFILE LINEAR // TACTILE LATENCY < 15MS
⌘1 ✂️ Silence
⌘2 🧠 AI Cut
⌘S 🟢 Karaoke
⌘K 🎬 9:16 Crop
⌘M 🎙️ Mix Audio
⌘P 📊 Keynote
⌘T 📈 Markets
⌘D 🚀 Deploy
Command: Silence Excision — Strips low-decibel gaps automatically execution/video_silence_remover.py

Deterministic power. Autonomous by design.

Separating probabilistic LLM reasoning from high-speed deterministic execution code to deliver zero-drift video editing.

Audio Silence Excision
Analyzes raw audio streams chunk-by-chunk in decibels (dB), eliminating dead air and pauses without altering natural speech inflection.
ACTIVE AUDIO STREAM -50.0 dB THRESHOLD
Universal Layout
One-click reformatting between 9:16 vertical shorts (TikTok, Reels) and 16:9 cinematic widescreen with safe zone protection.
9:16
16:9
Multi-Platform
Direct integration with YouTube Data API v3, automated X thread generation, and n8n webhook webhooks.
YouTube v3CONNECTED
X / TwitterCONNECTED
n8n WebhookREADY
Hot-Swappable Pipeline Modules
Each execution module is isolated, deterministic, and callable via REST API or standalone CLI. Plug in Whisper-1, GPT-4o, MoviePy, or native FFmpeg without touching core orchestration.
● Whisper-1
Word-level timestamps
● GPT-4o Editor
False-start retake cutter
● FFmpeg ASS
Native burn-in renderer
● HyperFrames
GSAP motion b-roll
Cyberpunk Karaoke Subtitles
Generates high-contrast SubStation Alpha (.ass) scripts with millisecond word highlights in neon green (#00FF41).
THE INFORMATION SIEVE IS ACTIVE

Engine specs. Precise by specification.

Deterministic runtime parameters ensuring 100% reproducible media renders across any Linux or Windows environment.

Silence Excision Threshold
-50.0 ± 5 dB
Word Timestamp Precision
1.2 ± 0.1 ms
Karaoke Active Highlight
#00FF41 Neon Green
Subtitle Font Profile
Consolas 120pt / 135pt
Render Compositor Engine
MoviePy + FFmpeg H.264
Gateway Latency
< 15 ms
Database Architecture
SQLite 3 ACID + Causal Graph
Frontend Stack
Next.js 15 + React 19 + Turbopack

Deploy in seconds. Run anywhere.

Clone the repository from GitHub and launch the unified media manager locally or in the cloud. 

# 1. Clone repository from GitHub
git clone https://github.com/MuzikayiseKhuzwayo/mina-ai.git
cd mina-ai

# 2. Configure environment
copy .env.example .env
# Set OPENAI_API_KEY=sk-...

# 3. Start Unified Gateway (Port 8080)
python server.py

# 4. Start Next.js Studio Cockpit (Port 3000)
cd dashboard && npm install && npm run dev

Diátaxis Documentation. 100% verified.

Structured according to the four Diátaxis quadrants under the /dubstrata-docs standard.

TUTORIALS
Quickstart Guide
10-minute zero-to-one walkthrough to run the media pipeline and burn your first video short.
Read on GitHub →
HOW-TO GUIDES
Video Pipeline
Step-by-step recipe from raw MP4 camera footage to silence excision, AI cuts, and subtitles.
Read on GitHub →
REFERENCE
API & Schemas
Complete REST, WebSocket, CLI dictionary, and SQLite database schema tables.
Read on GitHub →
EXPLANATION
3-Layer Model
Deep architecture narrative on the Directive-Execution-Memory loop and causal reasoning.
Read on GitHub →

The future of media is autonomous.

Unify your video editing, presentation design, and prediction market broadcasts into a single deterministic operating system.

View on GitHub Documentation Suite