explainer-video-pipeline

name: explainer-video-pipeline description: End-to-end pipeline for creating Finding Peak explainer videos — AI voiceover, auto music, synced timing, and Remotion rendering metadata: tags: video, explainer, elevenlabs, voiceover, tts, remotion, pipeline, automation, finding-peak

When to use

Use this skill when Ryan asks to create an explainer video, short-form video, or run any part of the video production pipeline. The project lives at ~/Projects/explainer-video/.

The Proven Workflow

Every video follows this exact sequence:

1. Develop the premise — Collaborate on the topic, angle, and narrative arc
2. Write the script JSON — Create scenes with voiceover text, templates, and content
3. 🛑 Creative Review — Present the full script to the user for review. Ask for edits, special instructions, or creative ideas before proceeding. Do NOT move to step 4 until the user approves.
4. Run the pipeline — One command generates music + voiceover + timing sync
5. Render — Export to MP4
6. Iterate — Adjust voice speed, music levels, or scene content and re-run

Prerequisites

• Node.js 18+
• ElevenLabs API key in ~/Projects/explainer-video/.env
• Dependencies installed: cd ~/Projects/explainer-video && npm install

Quick Start (Full Pipeline)

cd ~/Projects/explainer-video

# 1. Create the script JSON (Claude writes this collaboratively)
# Save to: ./scripts/input/my-video.script.json

# 2. 🛑 STOP — Present script to user for review & approval
#    Ask: "Here's the script. Any edits, special instructions, or ideas before I run the pipeline?"
#    Wait for approval before continuing.

# 3. Run the full pipeline (music + voiceover + timing sync)
npm run pipeline -- ./scripts/input/my-video.script.json

# 4. Render the video
npx remotion render src/index.ts ExplainerVideo out/my-video.mp4 --props=./video-config.json

# 5. Open for review
open out/my-video.mp4

Brand Constants (Finding Peak)

• Sign-off: Every video ends with "This is the way."
• Logo: Finding Peak watermark at bottom (45% opacity)
• Voice: Ryan Hanley clone (ID: Vs1aGUxp6Uuqr3rdNl9v)
• Theme: Dark background (#333333), yellow accent (#f8f024), white text
• Captions: Word-synced TikTok-style — floating white words, active word in yellow, bottom-third positioning
• Format: 1080x1920 vertical (9:16), 30fps

How to use

New user? Start here:

• rules/onboarding.md - Step-by-step setup: brand config, voice cloning, environment, first video

Then read individual rule files for detailed explanations and guidance:

• rules/pipeline-overview.md - End-to-end workflow, npm commands, file flow
• rules/script-writing.md - Script structure, pacing, voiceover text, voice settings
• rules/scene-mapping.md - Which template for each narrative beat
• rules/voiceover-generation.md - ElevenLabs integration, voice settings, speed tuning
• rules/timing-sync.md - How timestamps map to scene boundaries + word-level timing
• rules/captions.md - Word-synced captions: global overlay architecture, timing formula, styling, troubleshooting
• rules/audio-layering.md - Background music, ducking, volume calibration
• rules/visual-polish.md - Theme, transitions, spring tuning, motion graphics
• rules/rendering.md - Studio preview, render commands, output formats
• rules/config-reference.md - Full VideoConfigSchema documentation
• rules/custom-motion.md - Custom motion graphics: declarative SVG elements, animations, color tokens, recipes
• rules/template-gallery.md - All 20 scene templates with descriptions