2025-08-12 01:33:30 +00:00
2025-07-23 20:54:39 -07:00
2025-08-04 18:39:29 -07:00
2025-07-23 11:52:09 -07:00
2025-07-23 11:52:09 -07:00
2025-07-23 11:52:09 -07:00
2025-07-23 11:52:09 -07:00
2025-07-23 11:52:09 -07:00
2025-07-23 11:52:09 -07:00
2025-08-12 01:33:30 +00:00
2025-07-23 11:52:09 -07:00
2025-07-23 11:52:09 -07:00
2025-07-23 11:52:09 -07:00

🎥 LCS Pipeline

Automated livestream highlight rendering and publishing for Fortnite content featuring Gramps.

This project powers the backend of Llama Chile Shop, transforming raw livestream clips into polished, uploaded videos — complete with titles, thumbnails, intros/outros, and social metadata.


⚙️ Features

  • Daily folder scan for new stream sessions (20250710) [v0.1.0]
  • 📂 Clip classification (hits/, misses/, montages/, outtakes/, timelapses/) (20250807) [v0.1.2]
  • 🧠 AIgenerated titles and descriptions via OpenAI (20250710) [v0.1.0]
  • 🎬 Autostitched intro + title card + outro (20250723) [v0.1.0]
  • 🖼️ Dynamic thumbnail creation with Fortnite styling (20250725) [v0.1.0]
  • ⬆️ Uploads to YouTube (20250729) and PeerTube (20250807) [v0.1.1 & v0.1.2]
  • 📜 Metadata archive and session history (20250726) [v0.1.0]
  • 🐘 (Planned) Social posts to Mastodon and Bluesky (20250720) [v0.2.0]

🚀 Quick Start

git clone https://llgit.llamachile.tube/gramps/video-pipeline.git
cd video-pipeline
pip install -r requirements.txt
cp .env.example .env  # Fill in your API keys and config
python main.py

Requires Python3.13+ and access to mapped NAS directory (e.g. Z:\2025.08.05\hits\).


📁 Folder Structure

video-pipeline/
├── main.py
├── config.py
├── .env.example
├── modules/
│   ├── render_engine.py
│   ├── title_utils.py
│   ├── thumbnail_utils.py
│   ├── yt_poster.py
│   └── ...
├── assets/         # Branding assets (intros, fonts, logos)
├── logs/           # Sync logs, wiki publish logs, etc.
└── metadata/
    └── history/    # Per-clip metadata archive

📚 Documentation

Full documentation is hosted in the 📖 Gitea Wiki

Recommended pages:


🛠️ Development Mode

  • DEBUG=True in .env disables destructive operations
  • All modules can be run/tested independently
  • Wiki editing is supported via local Markdown and wiki_publish.log

👤 About

Created by Gramps for Llama Chile Shop — a custom content pipeline for old-school gaming chaos.

Maintainer: gramps@llamachile.shop Contributions welcome in the form of bug reports, pull requests, or Fortnite gifts.

Description
Video Pipeline Automation The Video Pipeline project automates the process of extracting, processing, and uploading highlights from livestreams, primarily for Fortnite gameplay. This system integrates Eklipse.gg for AI-based highlight detection, DaVinci Resolve for video editing, and custom Python scripts for rendering and metadata management. The pipeline supports seamless uploads to YouTube and PeerTube, automatic social media announcements, and maintains a detailed history of video metadata for each session. Key Features: Automated highlight extraction from livestreams Dynamic title card generation with Fortnite-themed branding Streamlined YouTube and PeerTube uploads with dynamic descriptions Metadata tracking and session management Social media integration for automatic posts to Mastodon and Bluesky This project is designed to optimize video content creation and management for livestreamers, ensuring efficiency and consistency in processing and sharing content.
Readme GPL-3.0 108 KiB
Languages
Python 100%