YT Content Factory

What this project is

YT Content Factory is an active build: a local-first AI video production system that turns ideas and source clips into review-ready Shorts/Reels with dashboards, media validation, safety gates, and repeatable editorial workflows.

It is designed for creator operations, not blind autopublishing. The system is strongest when it helps shortlist clips, validate media, prepare review renders, and give a human operator a clear approval path.

Why I built it

Short-form video production gets messy fast when ideas, source clips, validation, renders, and review notes live in different places. I built YT Content Factory to make that workflow repeatable without pretending that AI should own the final editorial decision.

What it proves

YT Content Factory is meant to prove that creator workflows can use automation without losing editorial control. The important layer is not a fake viral machine; it is media validation, review manifests, local renders, and human approval before output is treated as publish-ready.

What is already working

• Mission Control dashboard direction for operator visibility
• Operator UI for managing review and production state
• Local media pipeline for working with source clips safely
• FFmpeg and FFprobe validation for media integrity checks
• Review manifests for documenting what was selected and rendered
• 360 clip shortlisting support
• Shot and reframe selection workflow
• Local review renders for checking output before publication
• Approval gates so AI-assisted output stays human-reviewed
• AI-assisted review with human final approval

How it is designed

YT Content Factory combines local media processing, validation checks, operator dashboards, review manifests, clip shortlisting, reframe decisions, and approval gates. The system is designed to make AI-assisted video production repeatable without pretending that the AI should own final editorial judgment.

The strongest engineering boundary is the review loop: local renders and validation artifacts are evidence for a human decision, not automatic proof that something is publish-ready.

Current boundaries

• The workflow keeps human final approval in the loop before anything is treated as publish-ready.
• Local review renders are not the same thing as final published content.
• The system should not be described as autopublishing or fully autonomous editing.
• Docker should not be described as running the full media workflow end to end unless separately verified.
• Output quality and editorial taste still require human review.

What I am improving next

I am improving shot selection, reframe review, approval gates, media validation, and operator UI polish before making stronger production-scale claims.

Proof/assets coming next

Public-safe proof can include dashboard screenshots, operator UI states, review manifests, validation summaries, local review render screenshots, FFmpeg/FFprobe check output, and redacted shot-selection examples.

Proof should show the workflow and review gates without exposing private media files or unreleased render details.

Proof slots: operator UI screenshot slot, review manifest evidence slot, local render demo coming, and verification-first asset publishing.