Image Generation

Generate AI images from text descriptions. Choose between a higher-quality pro model and a faster, cheaper flash model, pick a resolution and aspect ratio, and optionally pass reference images for style guidance. Images are returned as local files or displayed inline.

Trigger

Invoke this skill with /image-gen, or use any of these phrases:

Requires ListenHub Skills to be installed — see Getting Started.

Quick Example

Generate an image: cyberpunk city at night, 16:9, 2K

The AI collects your preferences one question at a time, summarizes them for confirmation, and then generates the image.

Parameters

Choosing a Model

Both models are Google Gemini image models. Pick based on quality versus speed and which aspect ratios you need.

The four extreme ratios — 1:4 (narrow portrait), 4:1 (wide landscape), 1:8 (extreme portrait), and 8:1 (panoramic) — are available only on the flash model. The eight standard ratios work on both models.

Writing Good Prompts

A good prompt covers these elements:

1. Subject — what is in the image
2. Style — art style or visual treatment
3. Composition — how elements are arranged
4. Lighting/Mood — atmosphere and time of day
5. Quality — detail level and rendering quality

Examples

Basic:

a cat sitting on a windowsill

Better:

a fluffy orange tabby cat sitting on a sunny windowsill, warm afternoon light, cozy interior, highly detailed, photorealistic

Style Keywords

Prompt Enrichment

The AI passes your prompt through unchanged by default. It only offers to enrich the prompt when:

• The prompt is very short (a few words), and
• You haven't asked for verbatim generation.

When you accept enrichment, the AI adds style, lighting, and composition detail, then shows you the expanded prompt before submitting. For example, "cyberpunk" becomes "neon lights, futuristic, dystopian, rain-slicked streets."

The AI never rewrites a prompt that is already long, detailed, or structured — it treats you as experienced. It also leaves the prompt exactly as written when you say something like "use this prompt exactly."

Reference Images

Reference images guide the AI on style, not content. Your prompt still controls what appears in the image. You can supply up to 5 references, mixing local files and URLs in one request.

Supported formats: jpg, png, webp, gif. Max 10 MB per file.

Using Local Files

The skill accepts local file paths directly. Provide the path (for example ./sketch.png) when the AI asks about references — it uploads the file for you. No image hosting needed.

Using Image URLs

1. Upload your reference to an image hosting service (imgbb.com, sm.ms, postimages.org)
2. Copy the direct image URL (ending in .jpg, .png, .webp, or .gif)
3. Provide the URL when the AI asks about references

You can mix local files and URLs in the same request — for example one local sketch plus one hosted photo.

Using Base64 Inline Data (API)

When calling the Image Generation API directly, you can also provide reference images as base64-encoded data via the inlineData field — no image hosting required. This is useful for programmatic workflows where you already have the image in memory.

Output

Output behavior follows the outputMode set during config:

• inline (default) — the image is displayed directly in the conversation
• download — saved to .listenhub/image-gen/YYYY-MM-DD-{id}/ in the current project
• both — displayed inline and saved locally

The output mode can be changed at any time by saying "reconfigure" when the AI shows your current config.

API Reference

See the Image Generation API reference for endpoint details, request parameters, the provider/model matrix, and code examples.