skills.rest
quanganh208
quanganh208COMMUNITY
Explore all skills by this author →
GitHub

ck:ai-multimodal

Process images, audio, video, and documents into transcripts, OCR, and generated media.

1|Updated Mar 6, 2026
Search Tags:#multimodal#transcription#gemini#image-generation#tts#minimax#video-generation
One-click install
npx skills add https://github.com/quanganh208/cookmate --skill ck-ai-multimodal-quanganh208
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ck:ai-multimodal
Source: https://github.com/quanganh208/cookmate/tree/main/.opencode/skills/ai-multimodal
Command: npx skills add https://github.com/quanganh208/cookmate --skill ck-ai-multimodal-quanganh208

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Enable teams and creators to automatically analyze, transcribe, and convert multimodal media (images, audio, video, PDFs) into structured outputs and generated assets, removing manual processing and format friction.

Core Features & Use Cases

  • Multimodal Analysis: Vision understanding, OCR, object detection, VQA, and scene/timeline extraction from images and videos.
  • Transcription & Audio: Long-form and chunked transcription, speaker labeling, and TTS generation using MiniMax.
  • Media Generation: Image, video, speech, and music generation via Gemini (Imagen/Veo/Nano Banana) and MiniMax (Hailuo, TTS, music).
  • Preflight & Automation: Media optimization (ffmpeg/Pillow), file uploads (inline vs File API), key rotation, batch processing, and organized outputs saved to docs/assets.
  • Use Case: Convert recordings and media assets into searchable transcripts, captioned videos, and generated marketing images or background music for production pipelines.

Quick Start

Run the setup checker to verify API keys and dependencies, then run the batch processor with your file path and chosen task to analyze or generate media.

Frequently Asked Questions about ck:ai-multimodal

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

FAQPage Schema
How do I transcribe long-form audio and extract structured transcripts with speaker labels?▼

To transcribe long-form audio into structured transcripts with speaker labels, use the batch processor with MiniMax endpoints. It supports chunked transcription and inline files under 20MB or File API uploads, saving outputs to project docs.

Can I generate images, video, and audio from text using Gemini and MiniMax?▼

Yes, you can generate images, video, speech, and music from text via Gemini (Imagen/Veo/Nano Banana) and MiniMax (Hailuo/TTS) endpoints. The processor handles API key rotation and saves generated media assets directly to project folders.

What is the best way to extract OCR and scene descriptions from video and PDF documents?▼

The best way to extract OCR, vision Q&A, and scene descriptions from video and PDFs is multimodal analysis. It processes images and videos to extract timeline data and structured text, removing manual format friction from content pipelines.

Does this multimodal processing workflow require ffmpeg and Pillow for media preflight?▼

Yes, ffmpeg and Pillow are required for media preflight. They optimize media files before processing, ensuring inline uploads under 20MB or File API transfers run smoothly against the Gemini and MiniMax model endpoints.

How do I handle large video files exceeding the 20MB inline upload limit for transcription?▼

For large video files exceeding the 20MB inline limit, use the File API upload method. The skill's preflight automation optimizes media with ffmpeg before routing files to the File API for Gemini or MiniMax transcription.

Why does batch processing fail when API keys are not configured correctly?▼

Batch processing fails without correctly configured API keys because the workflow requires authentication to access Gemini and MiniMax endpoints. Run the setup checker first to validate dependencies and rotate API keys to prevent processing interruptions.

Related Skills

Explore More Design & Creative Skills→
nguyenvanlinh1902nguyenvanlinh1902

ai-multimodal

Extract insights from images, audio, and video using Gemini multimodal APIs.

Community
Advanced
sunLeeesunLeee

ref-image-ocr

Extract text from images via OCR and save it as Markdown.

Community
Intermediate
rafaelcallejarafaelcalleja
2

media-processing

Process multimedia files with FFmpeg and ImageMagick for conversion and encoding.

Community
Advanced
jangtrinhjangtrinh

ck:ai-multimodal

Process images, audio, video, and documents with Gemini and MiniMax APIs.

Community
Advanced