ai-multimodal

Analyzes images, audio, and video and generates images and short videos via Gemini API.

Updated Mar 7, 2026

Search Tags:#ocr #image generation #multimodal #transcription #gemini #video generation #image analysis

Ready to run

npx skills add https://github.com/MinhHoangDono/antigravity-kit --skill ai-multimodal-minhhoangdono

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ai-multimodal
Source: https://github.com/MinhHoangDono/antigravity-kit/tree/main/.agent/skills/ai-multimodal
Command: npx skills add https://github.com/MinhHoangDono/antigravity-kit --skill ai-multimodal-minhhoangdono

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires google-genai, python-dotenv, pillow, and includes scripts (resource) and references (resource) components.

What problem does it solve?

It helps you analyze and transform images, audio, video, and documents by extracting meaning from them and producing new visual media.

Core Features & Use Cases

Multimodal analysis: Analyze images and visual content for understanding, transcription context, OCR/document extraction, and media description.
Transcription workflows: Transcribe audio/video by chunking long files and combining results into timed output.
Vision generation: Generate images with Imagen 4 and generate short videos with Veo 3, suitable for ideation, prototyping, and creative iteration.
Use case: Upload a product screenshot and a short demo clip to extract key UI/design details and a readable transcript, then generate a matching promotional image.

Quick Start

Use the ai-multimodal skill to analyze an image and return a clear description by asking: "Describe this file and extract any readable text."

Frequently Asked Questions about ai-multimodal

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

FAQPage Schema

Can I transcribe long audio and video files by chunking them into smaller segments?▼

Yes, you can transcribe long audio and video files by chunking them into smaller segments. The transcription workflow processes each chunk individually and combines the results into a single timed output.

What is the best way to generate short videos for creative prototyping?▼

Yes, you need a Gemini API key to generate images and short videos. The generation workflows use Imagen 4 for image creation and Veo 3 for short video generation within the local environment.

How do I handle media size and format constraints before multimodal analysis?▼

Generating short videos for creative prototyping is best handled using Veo 3 integration. This allows you to ideate and iterate on visual designs directly from your local environment using the provided CLI.

Can I extract UI design details from a product screenshot and demo clip together?▼

You handle media size and format constraints by applying local preprocessing utilities before multimodal analysis. The provided media conversion tools adjust files to meet the required specifications for processing.

How do I extract readable text from an image for OCR and document analysis?▼

Yes, you can extract UI design details from a product screenshot and demo clip together using multimodal analysis. It processes both visual content and audio to return design details and a readable transcript.

Can I transcribe long audio and video files by chunking them into smaller segments?▼

To perform OCR and document analysis, you analyze images to extract readable text and understand visual content. This multimodal approach processes the file directly to return a clear description and any embedded text.