The Ultimate Tech Stack for YouTube Auto-Generation
How we combined React, Remotion, Make.com, and Google Gemini into a 24/7 autonomous video production pipeline without touching video editing software.
In the era of rapid content consumption, maintaining a daily cadence for YouTube Shorts can be exhausting. To solve this, AI Hack Lab developed a fully autonomous pipeline that conceptualizes, writes, voices, edits, and uploads high-retention vertical videos without any human intervention.
Here is the exact battle-tested architectural stack powering our content machines.
1. The Orchestrator: Make.com
We use Make.com (formerly Integromat) as the central nervous system. While cron jobs work for basic tasks, Make.com provides a visual webhook infrastructure that beautifully handles complex API retries and OAuth token management. Our Make.com scenario triggers every 6 hours, pushing parameters to our local worker nodes.
2. The Brains: Google Gemini AI
GPT-4 is excellent, but Google Gemini Pro offers incredibly fast, highly-structured JSON outputs with a remarkably high adherence to formatting constraints. We mandate that Gemini responds strictly in a pre-defined JSON schema representing our 10-second quiz loop (Theme, 4 voice lines, answer imagery keyword).
3. The Engine: React + Remotion
This is where the true magic happens. We completely eliminated the need for Adobe Premiere or AfterEffects by treating video composition as a web development problem.
Using Remotion, we write standard React components (<div style={{...}}>) which represent frames of a video.
By calculating audio durations dynamically and injecting variable delays, React generates a seamless .mp4 file. It fetches external images on the fly, applies CSS transitions (like zooming and panning), and overlays the TTS tracks.
Because the entire “video editor” is just JavaScript, we can mass-produce variations instantly by simply swapping the JSON props!
4. The Last Mile: YouTube Data API v3
Once the MP4 is built, Node.js securely authenticates with the YouTube Data API using OAuth2, uploading the newly minted video with the automatically generated title, description, and hashtags—immediately bringing organic traffic into the ecosystem.
Conclusion
Automating a YouTube channel is no longer a pipe dream—it’s standard engineering. By combining modern web technologies like React with powerful AI models, the barrier to mass content production has dropped to zero. Welcome to the future of media!
関連記事
- Arena Blueprint launch 前夜 — 9 ヶ月の自動化ログから抽出した 10 モジュール — Module 3「コンテンツ配信」の起源
- JSON 出力失敗を解決する Validator 設計 — 本記事 §3 の Validator 拡張版（10 層 Validator pipeline）
- Claude Skills #1：Session Handoff Builder — Skills Bundle の活用例