# Video — Marketing Skill for AI Agents

The first decision in AI video is approach, not tool: use programmatic frameworks (Hyperframes or Remotion) for templated, data-driven, or batch video; AI generation models (Veo 3, Sora 2, Runway, Kling) for original B-roll and scenes you can't film; AI avatars (HeyGen, Synthesia) for talking-head content without a camera, especially recurring or multilingual videos; and editing tools (Descript, Opus Clip, CapCut) to cut long-form into short clips. Don't use AI generation for product UI or readable on-screen text — models hallucinate both, so use real screen recordings and programmatic overlays instead. For agent-driven work, Hyperframes wins because it's plain HTML/CSS with deterministic rendering (AI generates better HTML than React), while Remotion suits React shops needing complex animation and Lambda-scale batch rendering. Whatever the approach, add captions — around 85% of social video is watched without sound — and export in the right aspect ratio: 9:16 for social, 16:9 for YouTube and websites.

> Maintained by Corey Haines · v2.1.0 · Updated 2026-07-15

Install: `npx skills add coreyhaines31/marketingskills`
Source: https://github.com/coreyhaines31/marketingskills/tree/main/skills/video

## What's inside

**The playbook covers:** Before Starting, Choosing Your Approach, Programmatic Video, AI Video Generation, AI Avatars, Editing & Repurposing Tools, Video Production Workflows, Agent-Native Video Pipeline, Common Mistakes, Task-Specific Questions, Tool Integrations

### Reference library

- [AI Video Prompting Guide](https://github.com/coreyhaines31/marketingskills/blob/main/skills/video/references/ai-video-prompting.md) — How to write effective prompts for AI video generation models (Veo, Runway, Kling, Pika).
- [Reverse-Engineering an Edit (The Beat Sheet)](https://github.com/coreyhaines31/marketingskills/blob/main/skills/video/references/edit-anatomy.md) — A viral short-form video usually isn't winning on the footage — it's winning on the *edit*: the cut rhythm, the caption style, the punch-ins, the on-screen text landing on the exact word, the b-roll cutaways, the sound design. This reference turns a reference edit you admire into a reusable edit spec — a beat sheet you (or an editing tool) can execute against your own footage — without copying a single frame of theirs.

## Key data

- 3. Generate 3-4 variations — same concept, different angles or styles

## Example

**Prompt:** We need a 2-minute product demo video for our SaaS homepage. What's the fastest way to produce it?

**60s product-demo outline**

[0-3s]  Hook: show the end result — the dashboard populating in one click
[3-8s]  "Here's how you go from raw CSV to a live report in under a minute."
[8-45s] Screen recording of the real product flow, with programmatic
         title cards + callouts (Hyperframes overlay, not AI-generated UI)
[45-55s] AI B-roll establishing shot (Veo/Runway) of the team using it
[55-60s] CTA: "Try it free →" with captions burned in
Export: 16:9 for the homepage, captioned throughout.

**AI video generation prompt (Subject + Action + Camera + Style + Mood)**

A close-up shot of hands typing on a laptop keyboard in a cozy coffee shop, shallow depth of field, warm afternoon light through a window, camera holds steady then slowly pulls back, cinematic color grading, 4K.

## FAQ

### Should I use Hyperframes or Remotion for programmatic video?

Choose Hyperframes for agent-driven workflows: it's plain HTML/CSS/JS with a minimal learning curve, deterministic rendering, an Apache 2.0 license, and it's LLM-native (AI models generate cleaner HTML than React components). Choose Remotion if you're already a React shop, need advanced animation (Spring, interpolate), or need large-scale batch rendering via Lambda on AWS — accepting the steeper learning curve and company license for commercial use.

### What's the best way to make weekly videos without being on camera?

Use an AI avatar, and HeyGen specifically — it has the best lip-sync, an official MCP server so agents can generate videos directly, 230+ avatars, and 140+ languages, with unlimited videos on paid tiers. Upload 2-5 minutes of yourself to create a custom digital twin for brand consistency. Avatars shine for recurring, multilingual, and personalized-at-scale content; if authentic founder presence matters more than cadence, film yourself instead.

### How do I turn a long webinar or podcast into short social clips?

Run the repurposing workflow: clean up the recording and remove filler in Descript (transcript-based editing), auto-extract the 5-10 best moments with Opus Clip (which scores virality potential), then add captions, effects, and platform styling in CapCut. Export 9:16 for TikTok, Reels, and Shorts, always caption them since most social video is watched muted, and make sure each clip hooks in the first three seconds.

## Related skills

- **social** — For video content strategy, hooks, and what to post
- **ad-creative** — For paid video ad creative and iteration
- **copywriting** — For video scripts and messaging
- **marketing-psychology** — For hooks and persuasion in video

---
Canonical: https://marketing-skills.com/skills/video