video

Create marketing videos using AI generation models, AI avatars, and programmatic video frameworks.

Updated Jul 14, 2026
One-click install
npx skills add https://github.com/usmangurowa/sifter --skill video-usmangurowa
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: video
Source: https://github.com/usmangurowa/sifter/tree/main/.agents/skills/video
Command: npx skills add https://github.com/usmangurowa/sifter --skill video-usmangurowa

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Producing marketing videos traditionally requires filming, editing skills, and expensive tooling. This Skill guides you through creating product demos, explainers, social clips, and ads using AI video generation models, AI avatars, and code-based video frameworks — no camera or editing expertise required. ## Core Features & Use Cases - Programmatic Video: Build templated, data-driven videos with Hyperframes (HTML/CSS) or Remotion (React) for batch rendering and repeatable content. - AI Video Generation: Generate original footage from text prompts using Veo 3, Sora 2, Runway, Kling, Seedance, and other models, with a structured prompting formula (subject + action + camera + style + mood). - AI Avatars: Produce talking-head videos without filming using HeyGen (with MCP server for agent integration) or Synthesia, including custom digital-twin avatars. - Editing & Repurposing: Turn long-form content into short clips with Descript, Opus Clip, and CapCut, and reverse-engineer viral edit styles into reusable beat-sheet specs. - Use Case: You need 50 personalized sales outreach videos. The Skill recommends a pipeline combining HeyGen's MCP server for avatar narration with Hyperframes for branded overlays, driven by a script template with per-company variables. ## Quick Start Ask the assistant to help you create a 60-second product explainer video and it will walk you through choosing the right production approach, tools, and workflow.

Frequently Asked Questions about video

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a product demo video without filming?

Screen record your product flow, then add programmatic titles and callouts with Hyperframes or Remotion. Use AI-generated B-roll from Veo or Runway for establishing shots, and record a voiceover or use a HeyGen AI avatar for narration. Never use AI generation for product UI screens since models hallucinate interfaces.

Hyperframes vs Remotion: which should I use for programmatic video?

Choose Hyperframes for agent-driven workflows since it uses plain HTML/CSS with no React DSL, has minimal learning curve, and is Apache 2.0 licensed. Choose Remotion if you already use React, need advanced animations like Spring and interpolate, or require large-scale batch rendering via AWS Lambda.

Which AI video generation model should I use?

Use Veo 3 or Sora 2 for highest quality with synced audio, Kling or Seedance for batch volume at lower cost, Hailuo for character consistency across shots, and Hunyuan Video or Wan 2 for self-hosted open-weight generation. Runway works well for image-to-video and storyboard workflows.

How do I turn a webinar into short social media clips?

Clean up the recording in Descript using transcript-based editing, then run it through Opus Clip to auto-extract the best moments with virality scores. Polish clips in CapCut with captions and platform styling, exporting at 9:16 for TikTok, Reels, and Shorts.

Can AI video models render readable text in videos?

No, AI video models cannot reliably render readable text or logos. Generate the visual footage with AI, then add text overlays programmatically using Hyperframes, Remotion, or CapCut in post-production.

When should I use an AI avatar instead of filming myself?

Use AI avatars like HeyGen for recurring content, multilingual versions, and personalized outreach at scale. Film yourself instead when authentic founder content matters, and use screen recording rather than avatars for product UI walkthroughs.