YouTube Shorts Automation Platform (`yt-pipeline-n8n`)

# YouTube Shorts Automation Platform (yt-pipeline-n8n)

AI Summary

Purpose:

• Capture durable knowledge about yt-pipeline-n8n, a personal YouTube Shorts

automation platform MVP under ~/hw/project.

Key points:

• The README describes a Mock MVP for operating multiple Japan-target YouTube

Shorts channels without external APIs.

• The system is split into a Control Plane (Backend API, Dashboard,

PostgreSQL, Redis, n8n, approval/docs/upload flows) and a Worker Plane (pull-based worker for heavy generation/rendering).

• Main stack evidence includes FastAPI, SQLAlchemy, Pydantic v2, Streamlit,

Docker Compose, n8n, Python Worker, Remotion, FFmpeg, TTS/image provider routing, and ai_quality/ prompt quality packs.

• Live channels confirmed by owner (2026-06-12): 11 Japan-target Shorts

channels operated via this platform — 30秒で世界史, 日本の今, ランキング30秒, 30秒マネー, サムライフットボール, 30秒でIT, 1日1成長, 脳の話JP, ブランド研究所, 世界ニュース30秒, 30秒レシピ. Revenue and performance metrics are still Needs confirmation.

Relevant when:

• Generating personal project portfolio or onboarding material.
• Working on YouTube Shorts automation, control/worker plane design, n8n,

prompt quality packs, or media rendering workflows.

Do not read full document unless:

• You need environment keys, runbook steps, provider routing, or channel

profiles.

Linked documents:

• ai/repo-notes/yt-pipeline-n8n.md
• ai/workspace/repos.md
• ai/wiki/projects/index.md

Open Questions

• ~~Which channels are live versus mock/demo~~ — resolved 2026-06-12: 11 live

Japan-target channels (owner-confirmed, see AI Summary roster).

• Whether real YouTube upload is enabled in current operation: Needs

confirmation.

• Production deployment, revenue, watch metrics, and active usage: Needs

confirmation.

• Automated test strategy beyond runbook/manual QA: Needs confirmation.

Details

System purpose

The project aims to automate planning, creating, reviewing, uploading, and documenting Japan-target YouTube Shorts content across multiple channels. The README emphasizes that the MVP can be tested end-to-end with mock providers: channel creation, video job creation, worker processing, approval, mock YouTube upload, Markdown docs, and weekly/monthly reports.

Architecture

Control Plane:

• FastAPI backend API.
• PostgreSQL for operational data.
• Redis for future queue or coordination flows.
• n8n for webhook/cron automation.
• Streamlit dashboard for operations.
• Approval, documentation, and upload coordination.

Worker Plane:

• Python pull worker that contacts the Backend API for jobs.
• Generates or renders heavier video assets only when the worker machine is on.
• Uses local storage by default, with prepared keys for future Nextcloud WebDAV

or Google Cloud Storage backends.

Provider and media pipeline:

• Mock provider is the default low-risk execution path.
• Optional CLI fallback paths reference Gemini CLI, Claude Code, and Codex CLI.
• TTS paths include mock, Google TTS, Gemini TTS, and VOICEVOX options.
• Rendering references FFmpeg and Remotion.
• ai_quality/ contains task-specific rules for topic research, script,

metadata, fact check, image, and revisions.

Safety and redaction

The README includes many environment variable examples and machine-specific paths. Any public or human-facing summary must generalize these details and must not copy credentials, tokens, usernames, app passwords, local secret paths, or private operational values.

Portfolio positioning

Suggested Korean portfolio angle:

• "YouTube Shorts 멀티 채널 자동화 운영 플랫폼"
• Highlight control/worker plane separation, n8n automation, Streamlit

operations dashboard, mock-first MVP, AI provider routing, and quality packs.

• Live channel roster is owner-verified (see AI Summary). Keep upload

automation depth, revenue, and performance metrics as Needs confirmation.