hirokidaichi/ergon

Overview

This skill is a CLI tool for fast AI media creation using Imagen 4, Veo 3.1, and Gemini TTS. It lets you generate and edit images, produce videos with synchronized audio, and create narrated voice tracks with character direction. The interface is command-line focused, scriptable, and designed for quick iteration and automation.

How this skill works

Run commands like image gen, image edit, video gen, and narration gen to invoke the appropriate model pipeline. Image generation and editing use Imagen 4 with style and aspect options; Veo 3.1 produces videos with integrated audio from prompts or input images; Gemini TTS renders narration with selectable voices, characters, and acting directions. All commands support dry-run and JSON output for integration into build scripts or creative workflows.

When to use it

• Create hero images, product photos, icons, or artwork from text prompts.
• Quickly edit or restyle existing images (background swaps, object removal, style transfer).
• Produce short videos with synchronized audio or animate a static visual asset.
• Generate voiceovers or character narration with precise tone, speed, and performance direction.
• Automate media tasks in CI scripts or content pipelines using JSON output and dry-run previews.

Best practices

• Start with a clear descriptive prompt and include sound cues when you want audio in videos.
• Pick a style (-t) and aspect ratio (-a) that match your delivery platform (e.g., 9:16 for mobile vertical video).
• Use dry-run to preview settings before consuming API credits or producing large outputs.
• Combine generation and edit steps iteratively: generate a base image, then refine via image edit for precise results.
• Specify character and direction for narration to get expressive, on-brand voice performances.

Example use cases

• Generate a 1:1 anime mascot for branding: npx ergon image gen "cute cat mascot" -t anime -a 1:1.
• Edit a product photo to a clean white background: npx ergon image edit product.jpg "change background to white and add soft shadow".
• Create an 8-second promotional video with music and voice: npx ergon video gen "host introduces new feature, upbeat music, friendly tone" -d 8
• Animate a logo into a short reveal with whoosh sound: npx ergon video gen "logo reveals with whoosh" -i logo.png -a 16:9
• Produce a guided meditation narration with a warm female voice: npx ergon narration gen "Breathe in..." -c "calm guide" -v Aoede -d "soft, slow"