OpenAI Image Gen — OpenClaw Skill

What This Skill Does

The OpenAI Image Gen skill lets your OpenClaw agent generate images using OpenAI's image generation models -- including GPT Image 1, GPT Image 1.5, GPT Image 1 Mini, DALL-E 3, and DALL-E 2. The agent runs a Python script that calls the OpenAI Images API, generates one or more images from text prompts, and creates an HTML gallery page for easy viewing. All generated images are saved locally with a JSON manifest mapping prompts to files.

The skill supports a wide range of generation options. GPT Image models offer four quality levels (low, medium, high, auto), three size options (1024x1024, 1536x1024 landscape, 1024x1536 portrait), transparent backgrounds, and output in PNG, JPEG, or WebP formats. DALL-E 3 adds a style parameter (vivid for hyper-real or natural for a more subtle look) and supports up to 1792x1024 resolution. Batch generation is supported on all models except DALL-E 3 (which is limited to one image per request).

This skill is ideal for content creation workflows -- generating social media graphics, blog post illustrations, product mockups, and creative assets directly from your AI agent. Ask your agent to "generate 4 variations of a hero image for our landing page" and it handles the prompt engineering, API calls, and file management automatically.

Example Prompts

Generate a high-quality landscape image of a serene mountain lake at sunset, 1536x1024, using GPT Image 1

Create 4 variations of a product photo mockup showing a coffee mug on a marble desk with warm studio lighting

Generate a transparent PNG logo concept of a stylized kiwi bird in a minimalist line art style

Use DALL-E 3 in vivid style to create a hyper-detailed illustration of a cyberpunk city street at night with neon signs

Batch-generate 8 social media card backgrounds in a dark, abstract geometric style -- 1024x1024, WebP format

Create a natural-style photo of a rustic Italian restaurant interior with warm candlelight, using DALL-E 3

Generate a set of 4 app icon concepts in a flat design style with transparent backgrounds using GPT Image 1.5

Requirements

Binary dependency: python3

Environment variable: OPENAI_API_KEY

Python: brew install python (macOS) or typically pre-installed on Linux
OpenAI API key: Get one at platform.openai.com/api-keys
API access: Your OpenAI account needs access to the Images API (enabled by default on most accounts)

Setup on KiwiClaw

This skill is pre-installed and configured on all KiwiClaw plans. Python and the generation script are available on every tenant machine. On the Standard plan ($39/mo), image generation uses your managed LLM credits. On the BYOK plan ($15/mo), add your OpenAI API key in the KiwiClaw dashboard settings. No additional setup needed.

Setup Self-Hosted

Ensure Python 3 is installed: python3 --version
Get an OpenAI API key from platform.openai.com
Set the environment variable: export OPENAI_API_KEY="sk-your-key"
The skill will be automatically available once python3 and the API key are configured

Related Skills

Discord -- generate images and share them directly in Discord channels
Slack -- create and post generated images to Slack
Notion -- generate images for Notion page content and covers
Obsidian -- create illustrations for your knowledge base notes

FAQ

Which image models does this skill support?

The skill supports GPT Image 1 (gpt-image-1), GPT Image 1 Mini (gpt-image-1-mini), GPT Image 1.5 (gpt-image-1.5), DALL-E 3, and DALL-E 2. Each model has different capabilities for size, quality, and output format. GPT Image models support transparent backgrounds and WebP output.

Can the skill generate multiple images at once?

Yes. GPT Image models and DALL-E 2 support batch generation of multiple images per prompt. DALL-E 3 is limited to 1 image per request. The skill automatically generates an HTML gallery page for viewing all generated images.

Does the skill require an OpenAI API key?

Yes. You need an OpenAI API key with access to the Images API. On KiwiClaw Standard plan, image generation uses your managed LLM credits. On the BYOK plan, it uses your own OpenAI API key.

What image formats and sizes are supported?

GPT Image models support 1024x1024, 1536x1024 (landscape), and 1024x1536 (portrait) sizes, with PNG, JPEG, and WebP output formats and transparent backgrounds. DALL-E 3 supports up to 1792x1024. DALL-E 2 supports up to 1024x1024. Quality options range from low to high depending on the model.