How narendramp15/yt-video works
This repository implements an automated, trend-driven YouTube video pipeline. The main workflow discovers and ranks topics, generates long- and short-form scripts and SEO metadata, creates voice and visual assets, aligns narration to scenes, renders videos, reviews them, and optionally uploads them. Checkpoints support resuming runs; TTS and YouTube Analytics memories feed lessons back into later runs. The README documents capabilities beyond the sampled excerpts; those links are cited to it, while unsampled implementation details are not assumed.
Main entry points: long_video.py Pipeline workflow · yt_analytics_memory_service.py Analytics memory · topic_service.py Trend discovery · main.py CLI
Components 23
Pipeline control
- CLI
main.py - Pipeline workflow
workflows/long_video.py
Topic and content
- Trend discovery
services/topic_service.py - Script generation
services/script_service.py - Title and thumbnail packaging
services/packaging_service.py - Script review
services/review_service.py - X trends integration
services/x_service.py
Media production
- Voice synthesis
services/tts_service.py - TTS style memory
services/tts_memory_service.py - Scene image generation
services/image_service.py - Image relevance check
services/clip_service.py - Transcript and scene sync
services/sync_service.py - Video rendering
services/video_service.py - Chart generation
services/chart_service.py
Publishing and learning
- Analytics memory
services/yt_analytics_memory_service.py - YouTube publishing
services/youtube_service.py - Analytics collection
services/yt_analytics_memory_service.py
Runtime support
- Step checkpoints
services/checkpoint_service.py - API cost tracking
services/cost_service.py - Configuration
config.py - Logging
logger.py
Other components
- Channel operator
- Trend sources
Connections 27
- Channel operator → CLI: runs commands
README.md - CLI → Pipeline workflow: starts pipeline
main.py - Pipeline workflow → Step checkpoints: resumes steps
workflows/long_video.py - Pipeline workflow → Trend discovery: fetches and ranks
workflows/long_video.py - Trend discovery → Trend sources: collects signals
README.md - Pipeline workflow → Analytics memory: uses learned signals
workflows/long_video.py - Analytics memory → Trend discovery: biases ranking
README.md - Pipeline workflow → Script generation: generates scripts
workflows/long_video.py - Analytics memory → Script generation: provides performance notes
README.md - Pipeline workflow → Title and thumbnail packaging: selects package
workflows/long_video.py - Pipeline workflow → Script review: reviews script
workflows/long_video.py - Pipeline workflow → Voice synthesis: generates narration
workflows/long_video.py - TTS style memory → Voice synthesis: supplies corrections
services/tts_service.py - Pipeline workflow → TTS style memory: updates voice memory
README.md - Pipeline workflow → Scene image generation: generates scene images
workflows/long_video.py - Scene image generation → Image relevance check: checks relevance
services/image_service.py - Pipeline workflow → Transcript and scene sync: aligns narration
workflows/long_video.py - Pipeline workflow → Video rendering: renders video
workflows/long_video.py - Pipeline workflow → YouTube publishing: uploads video
workflows/long_video.py - Pipeline workflow → Analytics collection: fetches analytics
README.md - YouTube publishing → Analytics collection: records video metadata
README.md - Analytics collection → Analytics memory: stores performance
README.md - Pipeline workflow → X trends integration: optionally fetches trends
workflows/long_video.py - Pipeline workflow → Chart generation: uses chart assets
- Pipeline workflow → API cost tracking: tracks run costs
workflows/long_video.py - Pipeline workflow → Configuration: reads settings
config.py - Pipeline workflow → Logging: logs progress
logger.py