image-service

Generate, edit, and analyze multimodal images via prompts and references.

266|71|Updated Jan 26, 2026

Search Tags:#workflow #editing #image #multimodal #generate #long-image

Ready to run

npx skills add https://github.com/zrt-ai-lab/opencode-skills --skill image-service

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: image-service
Source: https://github.com/zrt-ai-lab/opencode-skills/tree/main/image-service
Command: npx skills add https://github.com/zrt-ai-lab/opencode-skills --skill image-service

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires httpx, Pillow, numpy, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill unifies multimodal image generation, editing, and analysis into a single workflow, reducing tool fragmentation and accelerating visual content production.

Core Features & Use Cases

文生图: 根据中文提示生成高质量图片，覆盖信息图、图解、产品图等多种风格。
图生图/图生文: 在现有图片上执行编辑、描述、OCR 与内容分析，支持长图拼接等扩展场景。
长图拼接与信息图调研: 将多张图片拼接成微信长图，或生成信息图风格的可视化材料，便于营销与报告输出。
模板与自动化工作流: 集成模板与脚本，快速产出统一风格的视觉内容。具体示例：为电商详情页生成统一风格的长图，或为文档配图。

Quick Start

使用 text_to_image.py 为信息图风格生成一张示例图片，尺寸为16:9。

Frequently Asked Questions about image-service

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

FAQPage Schema

How do I generate images from Chinese prompts with configurable sizes?▼

Generate images from Chinese prompts using the text-to-image workflow, which supports configurable sizes and aspect ratios like 16:9. This feature produces high-quality visuals such as infographics and product images directly from descriptive text prompts.

Can I stitch multiple images into a long scrolling image for marketing?▼

Yes, you can stitch multiple images into a long scrolling image for marketing. The workflow supports long-image stitching, allowing you to combine several pictures into a single WeChat-style long image or unified e-commerce detail page visual.

Does the multimodal workflow support image analysis and OCR tasks?▼

Yes, the multimodal workflow supports image analysis and OCR tasks. You can perform image-to-text operations to extract text via OCR, generate image descriptions, and analyze content from existing images using integrated vision models.

What is the best way to edit existing images using references?▼

The best way to edit existing images using references is the image-to-image workflow. It applies edits based on reference images and prompts, allowing you to modify existing pictures and maintain a consistent style across marketing assets.

Do I need Python and Pillow to run this image generation workflow?▼

Yes, you need Python with Pillow, httpx, and numpy installed to run this image generation workflow. These dependencies handle HTTP requests, image processing, and numerical operations required for the multimodal generation and editing scripts.