imagencn — AI Image Generation with Chinese Text Excellence

中文文档

imagencn — Image Generation, Cloud-Native: one CLI, every image cloud. The project started with China-friendly clouds and now covers international providers as well.

A Claude Code / OpenClaw skill for AI image generation using Alibaba Cloud Bailian, ByteDance Volcano Ark, Tencent Hunyuan, Zhipu BigModel, StepFun, Google Gemini, Grok (xAI), OpenAI, and Black Forest Labs (FLUX) APIs.

📋 Model Reference — browse all 44 models with pricing, resolution, and feature comparison.

Pipeline

Features

• Alibaba Cloud Bailian (DashScope): Qwen-Image 2.0, Edit, Wan Series, Z-Image — 21 models
• ByteDance Volcano Ark: Doubao-Seedream series (5.0/4.5/4.0) — 3 models, up to 4K
• Tencent Hunyuan: Hunyuan Image 3.0 — flagship, complex Chinese composition
• Zhipu / BigModel: CogView-4, GLM-Image — 3 models, native Chinese text in images
• StepFun / 阶跃星辰: Step-2X, Step-Image-Edit-2 — 2 models, ultra-cheap volume gen
• Google Gemini (international): Gemini 3 Pro Image / 3.1 Flash Image — 4 models, 512/1K/2K/4K plus aspect ratios
• Grok / xAI (international): Grok Imagine — 3 models, aspect-ratio + resolution presets up to 4K
• OpenAI (international): GPT Image 1/1.5/2 — 4 models, arbitrary sizes up to 4K (gpt-image-2)
• Black Forest Labs / FLUX (international): FLUX.2 Pro / Max — 3 models, async API, flexible ratios + seed control
• Multiple size presets: 1:1, 16:9, 9:16, 4:3, 3:4, plus 1K/2K/3K/4K
• Cross-platform: Windows, macOS, Linux support
• Multiple API regions: China (default), Singapore, Virginia (DashScope)

Installation

Claude Code Marketplace (recommended)

/plugin install imagencn@365-skills

Or tell your coding agent:

help me to install https://github.com/Agents365-ai/imagenCN.git

Manual

# Global install
git clone https://github.com/Agents365-ai/imagenCN.git ~/.claude/skills/imagencn

# Project-specific
git clone https://github.com/Agents365-ai/imagenCN.git .claude/skills/imagencn

OpenClaw

git clone https://github.com/Agents365-ai/imagenCN.git skills/imagencn

SkillsMP

Search imagencn on skillsmp.com for one-click install.

Requirements

System Requirements

• Python 3.8+
• pip

Install Dependencies

pip install dashscope requests

API Keys

# Alibaba Cloud Bailian (DashScope)
export DASHSCOPE_API_KEY="your_api_key"
# Get key: https://bailian.console.aliyun.com/

# ByteDance Volcano Ark (optional)
export ARK_API_KEY="your_api_key"
# Get key: https://console.volcengine.com/ark/region:ark+cn-beijing/apikey

# Tencent Hunyuan (optional)
export HUNYUAN_API_KEY="your_api_key"
# Get key: https://console.cloud.tencent.com/tokenhub/apikey

# Zhipu / BigModel (optional)
export ZHIPUAI_API_KEY="your_api_key"
# Get key: https://bigmodel.cn

# StepFun / 阶跃星辰 (optional)
export STEP_API_KEY="your_api_key"
# Get key: https://platform.stepfun.com/interface-key

# Google Gemini (optional, international)
export GEMINI_API_KEY="your_api_key"
# Get key: https://aistudio.google.com/

# Grok / xAI (optional, international)
export XAI_API_KEY="your_api_key"
# Get key: https://console.x.ai/

# OpenAI (optional, international)
export OPENAI_API_KEY="your_api_key"
# Get key: https://platform.openai.com/api-keys

Config File (Optional)

Create ~/.imagencn.json for personal defaults:

{"platform": "ark", "model": "doubao-seedream-5-0-260128", "size": "2K"}

Project-level .imagencn.json overrides user-level. CLI args override both.

Upgrading from imagenCN (≤1.1.x)? The skill was renamed to lowercase imagencn — rename your config file too: mv ~/.imagenCN.json ~/.imagencn.json (and the install dir ~/.claude/skills/imagenCN → ~/.claude/skills/imagencn).

Optional Environment Variables

# Set default model per platform
export DASHSCOPE_MODEL="wan2.7-image-pro"       # DashScope default
export ARK_MODEL="doubao-seedream-5-0-260128"   # Volcano Ark default
export HUNYUAN_MODEL="hy-image-v3.0"            # Tencent Hunyuan default
export ZHIPUAI_MODEL="cogview-4"                # Zhipu default
export STEP_MODEL="step-2x-large"               # StepFun default
export GEMINI_MODEL="gemini-3-pro-image-preview" # Google Gemini default
export XAI_MODEL="grok-imagine-image-quality"    # Grok / xAI default
export OPENAI_MODEL="gpt-image-1"                # OpenAI default
export BFL_MODEL="flux-2-pro-preview"            # FLUX default

# Set API endpoint (DashScope only, default: cn)
export DASHSCOPE_API_BASE="cn"  # or "sg", "us", or full URL

Quick Start

Natural Language (Claude Code)

Just tell Claude what you want:

Generate an image of a cute orange cat
Create a poster with text "Happy New Year" in Chinese
Make a photorealistic 4K mountain sunset photo using wan2.7-image-pro
Generate a 16:9 landscape wallpaper

Command Line

# Basic usage (default model: qwen-image-2.0-pro, native 2K)
python scripts/generate_image.py "A cute cat" output.png

# Photorealistic 4K with Wan2.7 (DashScope)
python scripts/generate_image.py --model wan2.7-image-pro --size 4K "Mountain sunset" photo.png

# Volcano Ark (ByteDance) — requires ARK_API_KEY
python scripts/generate_image.py --platform ark "Editorial portrait, Vogue style" portrait.png

# Tencent Hunyuan — requires HUNYUAN_API_KEY
python scripts/generate_image.py --platform hunyuan "Astronaut on the moon, cinematic" scifi.png

# Google Gemini (international) — requires GEMINI_API_KEY
python scripts/generate_image.py --platform gemini "Japanese garden, morning light" garden.png

# Grok / xAI (international) — requires XAI_API_KEY
python scripts/generate_image.py --platform grok --size 16:9 "Cyberpunk city street at night" city.png

# OpenAI (international) — requires OPENAI_API_KEY
python scripts/generate_image.py --platform openai --quality high "Minimalist ceramic teapot product shot" teapot.png

# Black Forest Labs / FLUX (international) — requires BFL_API_KEY
python scripts/generate_image.py --platform bfl --size 16:9 "Volcanic coastline at dusk" coast.png

# Edit an existing image (DashScope, requires --image)
python scripts/generate_image.py --model qwen-image-edit-max --image input.png "Change the background to a beach" edited.png

# With negative prompt (DashScope)
python scripts/generate_image.py --negative "blurry" "High quality portrait" portrait.png

# List all 9 platforms' models
python scripts/generate_image.py --list-models

Models

Model	Best For
qwen-image-2.0-pro	Default, latest flagship, native 2K, strongest typography and detail
qwen-image-2.0-pro-2026-06-22	Latest snapshot (Jun 2026), generation + editing fusion
qwen-image-2.0	Standard 2.0 tier, native 2K
qwen-image-max	Previous-gen flagship
qwen-image-max-2025-12-30	qwen-image-max snapshot, improved realism
qwen-image-plus	Distilled accelerated version
qwen-image-plus-2026-01-09	qwen-image-plus snapshot (Jan 2026)
qwen-image-edit-max	Flagship image editing (requires --image)
qwen-image-edit-max-2026-01-16	Latest editing snapshot (Jan 2026)
qwen-image-edit-plus	Fast, lower-cost image editing
qwen-image	Base model
wan2.7-image-pro	Latest photorealistic, up to 4K output
wan2.7-image	Wan 2.7 standard, up to 2K
wan2.6-t2i	Wan 2.6, flexible sizing
wan2.5-t2i-preview	High quality art
wan2.2-t2i-flash	Fast generation
wan2.2-t2i-plus	Professional tier
wanx2.1-t2i-turbo	Fast execution
wanx2.1-t2i-plus	Professional tier
wanx2.0-t2i-turbo	Earlier generation
z-image-turbo	Lightweight, fast & low-cost; portraits and product images
doubao-seedream-5-0-260128	ByteDance latest, up to 3K, PNG/JPEG, best text rendering
doubao-seedream-4-5-251128	ByteDance Seedream 4.5, up to 4K
doubao-seedream-4-0-250828	ByteDance Seedream 4.0, budget-friendly 4K
hy-image-v3.0	Tencent Hunyuan flagship, strong Chinese composition understanding
cogview-4	Zhipu CogView-4, native Chinese text in images
cogview-4-250304	CogView-4 fixed snapshot (Mar 2025), reproducible results
glm-image	Zhipu GLM-Image flagship, up to 2048×2048
step-2x-large	StepFun high quality, 0.1 RMB/image
step-image-edit-2	StepFun ultra-cheap, 0.02 RMB/image, negative prompt support
gemini-3-pro-image-preview	Google Gemini flagship (international), 512/1K/2K plus aspect ratios
gemini-3-pro-image	Google Gemini stable flagship, 1K/2K/4K
gemini-3.1-flash-image	Google Gemini fast generalist (Nano Banana 2), 512/1K/2K/4K
gemini-3.1-flash-lite-image	Google Gemini fastest/cheapest, 1K only
grok-imagine-image-quality	Grok high-quality image model, up to 4K
grok-imagine-image	Grok standard image model
grok-2-image	Grok legacy JPG model
gpt-image-1	OpenAI GPT Image 1, 1024/1536 sizes
gpt-image-1-mini	OpenAI GPT Image 1 Mini, fast & cheap
gpt-image-1.5	OpenAI GPT Image 1.5, improved quality
gpt-image-2	OpenAI flagship, arbitrary sizes up to 4K
flux-2-pro-preview	FLUX.2 Pro latest rolling (international), async API, prompt upsampling
flux-2-pro	FLUX.2 Pro pinned snapshot, reproducible results
flux-2-max	FLUX.2 highest quality, search-grounding

Size Presets

Qwen-Image 2.0 (native 2K):

• 1:1 → 2048×2048 (default)
• 16:9 → 2688×1536
• 9:16 → 1536×2688
• 4:3 → 2304×1728
• 3:4 → 1728×2304
• 1K → 1024×1024
• 2K → 2048×2048

Qwen-Image legacy:

• 1:1 → 1328×1328
• 16:9 → 1664×928
• 9:16 → 928×1664
• 4:3 → 1472×1104
• 3:4 → 1104×1472

Z-Image (pixel area 512×512 to 2048×2048):

• 1:1 → 1024×1024 (default)
• 16:9 → 1280×720
• 9:16 → 720×1280
• 2:3 → 1024×1536
• 3:2 → 1536×1024

Wan Series (Wan2.7 also accepts 1K/2K/4K):

• 1:1 → 1024×1024
• 1:1-large → 1280×1280
• 16:9 → 1280×720
• 9:16 → 720×1280
• 4:3 → 1200×900
• 3:4 → 900×1200
• 2:1 → 1440×720

Volcano Ark (Seedream):

• 1:1 → 2048×2048
• 16:9 → 2848×1600
• 9:16 → 1600×2848
• 4:3 → 2304×1728
• 3:4 → 1728×2304
• 1K / 2K / 3K / 4K (model-dependent)

Tencent Hunyuan (colon-separated):

• 1:1 → 1024:1024
• 16:9 → 1920:1080
• 9:16 → 1080:1920
• 4:3 → 1600:1200
• 3:4 → 1200:1600
• 3:2 → 1920:1280
• 2:3 → 1280:1920

Zhipu (CogView-4 / GLM-Image):

• 1:1 → 1024x1024 (default)
• 16:9 → 1344x768
• 9:16 → 768x1344
• 4:3 → 1152x864
• 3:4 → 864x1152
• 2:1 → 1440x720
• 1:2 → 720x1440

StepFun (Step-2X):

• 1:1 → 1024x1024 (default)
• 1:1-small → 512x512
• 16:9 → 1280x800
• 9:16 → 800x1280

Google Gemini (named sizes + aspect ratios):

• 512 / 1K (default) / 2K / 4K → named output size (4K on Pro / 3.1 Flash; Lite is 1K only)
• 1:1, 16:9, 9:16, 4:3, 3:4 → aspect ratio (no exact pixel sizes)

Grok / xAI (aspect ratio + resolution):

• 1:1, 16:9, 9:16, 4:3, 3:4, 2:1 → sent as aspect_ratio (default: 1:1)
• 1K / 2K / 4K → sent as resolution

OpenAI (GPT Image):

• 1:1 → 1024x1024 (default)
• 16:9 → 1536x1024, 9:16 → 1024x1536
• 4:3 → 1344x1024, 3:4 → 1024x1344
• 1K → 1024x1024, 2K → 2048x2048 (gpt-image-2 only), 4K → 3840x2160 (gpt-image-2 only)

FLUX (Black Forest Labs):

• 1:1 → 1024x1024 (default)
• 16:9 → 1344x768, 9:16 → 768x1344
• 4:3 → 1152x864, 3:4 → 864x1152
• 2:1 → 1440x720, 1:2 → 720x1440
• 1K → 1024x1024, 2K → 2048x2048 (flexible WxH also accepted)

API Endpoints

Region	Alias	URL
China (default)	cn	https://dashscope.aliyuncs.com/api/v1
Singapore	sg	https://dashscope-intl.aliyuncs.com/api/v1
Virginia	us	https://dashscope-us.aliyuncs.com/api/v1

Support

If this project helps you, consider supporting the author:

WeChat Pay

Alipay

Buy Me a Coffee

Give a Reward

Author

Agents365-ai

• Bilibili: https://space.bilibili.com/441831884
• GitHub: https://github.com/Agents365-ai

License

CC BY-NC 4.0 — free for non-commercial use. Commercial use requires permission.