ai-video-gen

Generate AI video clips from prompts and reference images via HeyGen or fal.ai.

1|Updated Apr 11, 2026
One-click install
npx skills add https://github.com/shige1014-dev/backup-OpenMontage --skill ai-video-gen-shige1014-dev
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ai-video-gen
Source: https://github.com/shige1014-dev/backup-OpenMontage/tree/main/.agents/skills/ai-video-gen
Command: npx skills add https://github.com/shige1014-dev/backup-OpenMontage --skill ai-video-gen-shige1014-dev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Generates high-quality AI video clips from natural-language prompts and reference images, removing the need for manual sourcing, editing, and multi-provider orchestration when producing short-form or cinematic video assets.

Core Features & Use Cases

  • Multi-gateway generation: Supports HeyGen and fal.ai gateways and multiple provider models (VEO, Kling, Sora, Runway, Seedance, MiniMax) to match quality, speed, and cost needs.
  • Automated workflow + polling: Submit generation jobs, poll for completion, and retrieve temporary video URLs and asset IDs for downstream composition or download.
  • Image-to-video support & aspect ratios: Use reference or tail images for guided animation and specify aspect ratios for social or cinematic formats.
  • Use Case: Create a 30–60 second promotional social clip by describing the scene, selecting a provider for cost/quality tradeoffs, then polling until the generated video URL and asset_id are returned.

Quick Start

Generate a 30-second cinematic landscape video of a sunrise over a coastal city with soft cinematic lighting, 16:9 aspect ratio, and realistic camera movement.

Frequently Asked Questions about ai-video-gen

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

FAQPage Schema
How do I generate AI video clips from text prompts and reference images?

To generate AI video clips from text prompts, submit a natural-language generation job to a provider like HeyGen or fal.ai, poll for completion, and retrieve the temporary video URL and asset ID for download or downstream composition.

What is the best way to use image-to-video generation for social media formats?

Image-to-video generation uses a reference image to guide animation, while specifying an aspect ratio tailors the output for social or cinematic formats. You can submit jobs with reference images to providers like VEO, Kling, Sora, Runway, Seedance, or MiniMax.

Do I need API keys to generate cinematic AI videos using HeyGen and fal.ai?

Yes, you need API keys to generate cinematic AI videos. The workflow satisfies multi-gateway authentication requirements by requiring either a HEYGEN_API_KEY or a FAL_KEY to submit generation jobs and poll for completion.

Can I select different AI video generation models like Sora or Runway within one workflow?

Yes, you can select different AI video generation models within one workflow. The provider selection feature supports multiple models including VEO, Kling, Sora, Runway, Seedance, and MiniMax to match your specific quality, speed, and cost tradeoff needs.

How does polling work when submitting a text-to-video generation job?

Polling works by continuously checking the status of a submitted text-to-video generation job until completion. Once finished, polling retrieves the temporary video URL and asset ID, allowing you to download the generated video or use it for downstream composition.