Skip to content

Latest commit

 

History

History
49 lines (35 loc) · 1.94 KB

File metadata and controls

49 lines (35 loc) · 1.94 KB

AI Agent Instructions — LiveStreamScheduler

Project Overview

YouTube/Facebook live stream scheduler with sermon post-processing pipeline. Schedules broadcasts via Google Calendar + YouTube API, extracts sermon clips, generates metadata, and creates Shorts.

Quick Start

pip install -r requirements.txt
python fb_scheduler.py --config fb_config.local.json   # CLI scheduler
python fb_trigger_server.py                            # Web UI (port 8080)

Architecture

Component Purpose
fb_scheduler.py Main scheduler — matches calendar events to YouTube broadcasts
fb_trigger_server.py SSE-streaming web dashboard for manual triggers
Sermons/ Sermon discovery, transcripts, boundaries, and downloads
Shorts/ Extracted short clips (landscape + portrait)
.github/skills/ Domain knowledge for sermon-metadata and short-maker workflows

Key Conventions

  • Config: Copy fb_config.example.jsonfb_config.local.json. Never commit credentials.
  • Python: PEP 8, Python 3.9+, zoneinfo for timezones
  • Git: master branch, SSH remote, one branch per feature
  • Skills: YAML frontmatter + Markdown in .github/skills/<name>/SKILL.md
  • Sermon data: Sermons/sermons.json is the index of processed sermons

Tools & Dependencies

  • yt-dlp (CLI, not imported) for video/audio downloads
  • google-api-python-client for YouTube Data API
  • ffmpeg for segment extraction
  • FaceFollow (/home/kemmie/tools/FaceFollow) for portrait crop

Terminal Notes

  • rm, git branch -D, curl blocked by policy
  • WSL path: /mnt/c/Users/michael.kemmerer/Desktop/LiveStreamScheduler/
  • Redirect output: > /mnt/c/Users/michael.kemmerer/Desktop/out.txt 2>&1

Related Skills