Turn screenshots into a narrated promo video — and publish it to YouTube

Stop paying monthly for a cloud video tool just to ship one launch clip.

The AI Promo Video Kit is the exact, battle-tested workflow for turning a folder of screenshots into a polished, AI-narrated promo video — then pushing it straight to YouTube. Everything renders locally with headless Chrome + FFmpeg: no per-render fees, no watermark, no subscription. The only paid pieces are a few cents of AI voice and image generation.

You get the complete skill/playbook plus ready-to-run script templates. Drop it into Claude Code (or any coding agent), point it at your assets, and it builds the video end to end.

Why this is useful

• Promo/launch videos are expensive and slow — editors, render farms, monthly tools.
• This pipeline is scriptable and repeatable: screenshots in, finished MP4 out.
• It's deterministic — same inputs always produce the same video, so iteration is fast.

Best use cases

• Indie hackers & SaaS founders shipping a launch video on a budget
• Agencies / freelancers who want a repeatable, scriptable video pipeline
• Anyone using a coding agent who wants "screenshots → video" as one instruction

Core features

• HTML → MP4 rendering with HyperFrames (local, no per-render fees)
• AI voiceover synced per scene (fal.ai + ElevenLabs) — voice drives the timing
• AI scene art & YouTube thumbnails (xAI grok) — crisp, on-brand
• Ducked background music so narration always stays clear
• One-step YouTube publishing with title, description, tags, thumbnail, and the AI-content disclosure set correctly

What you get

• The full skill/playbook (hyperframes-video-workflow.md) — the complete end-to-end method, including the hard-won gotchas
• Companion guides for voiceover, AI images, and YouTube publishing
• A step-by-step account-setup guide for every service (with where to get each key)
• Ready-to-run templates: tts.mjs, music.mjs, scenes.json, and env.example
• A friendly quick-start so you ship your first video fast

Who it's for

Builders comfortable running a few terminal commands (or letting a coding agent run them). No video-editing experience required.

Requirements

• macOS or Linux, Node.js 22+, FFmpeg, and Google Chrome (for thumbnails)
• A fal.ai account (AI voice + music; pay-as-you-go, cents per clip)
• Optional: xAI for generated images/thumbnails, Zernio to auto-publish to YouTube, RakuBun to deploy a landing page — all covered in the setup guide

Honest framing: this is a workflow + guide, not a one-click app. You bring your screenshots and a few cents of API credit; it does the rest.