Tip
Click here to Preview the Dashboard UI Layout (Note: This is a static mockup. To run the full agent, please follow the local installation guide below.)
An automated, end-to-end framework for synthesizing, generating, and publishing highly-retaining educational videos fully autonomously directly to YouTube. Engineered by Malik Abualzait, this faceless channel automation system perfectly handles custom topics driven by Universal Language Models (OpenAI, Gemini, Ollama) and Glassmorphic UI Configurations.
🔥 Production Proven: This exact pipeline is currently actively utilized in production by Malik Abualzait to power and auto-publish content for the Togetherbudget YouTube channel alongside several other fully automated faceless channels!
Whether you want to build an AI YouTube Automation channel, scale a faceless content strategy, or dive into Python programmatic video generation, this repository is the ultimate boilerplate!
The system combines state-of-the-art textual synthesis, hyper-dynamic visually engaging graphic compositions, professional voice generation, FFmpeg sub-mixing, and automatic YouTube OAuth uploading right from a slick browser-based interface!
- Universal LLM Integration: Dynamic script structures using the universally supported OpenAI
$v1/chatformat (compatible with GPT-4, Gemini Pro, and Local Ollama!). - Auto Audio Ducking: FFmpeg sidechain compression automatically "ducks" background music under human-like Coqui/Bark generated vocals for cinematic audio quality!
- Automated Visual Compositing: Integrates native Pexels API image and video B-roll parsing with smart-transparency PNG infographics, stitched dynamically via Python!
- Zero-Touch YouTube Publisher: Built directly into the pipeline to securely upload rendered
.mp4masterclasses with SEO-optimized titles, descriptions, and tags.
Everything revolves around the beautiful Glassmorphic Control Panel. You no longer need to edit JSON files or dive into terminal commands repeatedly to manage your fully automated faceless channel!
From the dashboard you can:
- Fire custom specific keywords right into the generator pipe.
- View previously built content schemas natively in the built-in archive gallery folder!
- Control specific LLM, URL APIs, and Google Credentials seamlessly via forms.
- Clone the repo
git clone https://github.com/mabualzait/youtube-published.git cd youtube-published - Install system-level dependencies
You require
ffmpegto composite the multi-channel audios and overlays!# On macOS brew install ffmpeg - Install Python environments
The application fundamentally requires python 3.9+. We strongly suggest spinning off a fresh virtual environment.
python3 -m venv .venv source .venv/bin/activate pip install -r requirements.txt
Launch the graphical dashboard! We've utilized FastAPI as the engine routing core.
python dashboard/app.pyThe dashboard will instantly boot onto
http://localhost:8080.
Inside the Configurations panel of your web dashboard, you can define exactly what API providers power the generation!
1. Connecting AI Brains:
Select either Ollama, OpenAI, or Gemini inside the Quick Connect Dropdown templates! Paste your API Key natively inside the dashboard context. No need to touch .env vars!
2. Pexels Media Auth: Sign up at Pexels API natively to obtain a free key and throw it directly into the Visuals panel to fetch high-res looping background mp4s!
3. Authenticating YouTube:
- Acquire your OAuth 2.0 Client tokens directly from Google Cloud Platform Console for the YouTube API v3.
- Drag the
.jsoncontents into your dashboard and clickSave Configurations. - Run
python auth_youtube_agent.pyonce manually in your local terminal. It will cleanly pop open your browser and secure atoken_youtube.jsonbinding it to the machine forever!
Developer & Architect: Malik Abualzait
Focused on architecting the next era of autonomous AI agents and intelligent production pipelines.