HappyImage

English | 中文

HappyImage is a local AI visual content studio for creators and teams. Turn an idea, article, URL, or draft into image cards, infographics, comics, cover images, slide decks, and platform-ready publishing copy.

It brings the full baoyu-skills content workflow into three interfaces:

Interface	Best for	Start with
Desktop App	A focused local workspace with native app behavior	bun run dev:gui
Web UI	Visual generation, previews, settings, and project history	bun run dev:web
CLI	Automation, scripting, diagnostics, and repeatable jobs	bun run cli -- doctor

---

Highlights

Visual workflow, not guesswork Create from a topic, article, URL, or local file. HappyImage plans the content, proposes a visual direction, generates prompts, renders images, and keeps the full project history together. Outputs: analysis.md, outline.md, prompts/*.md, copy.md, and final images.	Style controls that are easy to scan Pick styles visually instead of memorizing option names. The gallery covers card, infographic, diagram, cover, comic, slide, and article illustration workflows. Examples: fresh, warm, bold, minimal, retro, notion, chalkboard, screen-print, sketch-notes.
Platform-aware publishing Generate copy and drafts for Xiaohongshu, WeChat, Weibo, and X. Each platform uses its own limits, hashtag rules, title format, and image constraints. Safety: HappyImage fills forms for review, but it never clicks "Publish" for you.	One engine, three ways to work Use the desktop app for daily creation, the Web UI for visual workflows, or the CLI for automation. They share the same local runtime, settings, skills, and output folders. Local-first: API keys, Chrome profile, generated projects, and settings stay on your machine.

---

What You Can Create

Skill	Use it for	Example
baoyu-image-cards	Xiaohongshu-style image cards and carousels	/baoyu-image-cards article.md --style fresh --layout balanced
baoyu-infographic	Structured visual explainers and data/story infographics	/baoyu-infographic content.md --layout pyramid --style technical-schematic
baoyu-diagram	Flowcharts, sequence diagrams, ER diagrams, architecture diagrams, timelines	/baoyu-diagram architecture.md --type flowchart
baoyu-cover-image	Article, blog, podcast, and newsletter covers	/baoyu-cover-image article.md --palette corporate-tech
baoyu-slide-deck	Presentation decks with consistent visual direction	/baoyu-slide-deck talk.md --aspect 16:9
baoyu-comic	Multi-panel comics with character sheets and dialogue	/baoyu-comic story.md
baoyu-article-illustrator	Header images, section visuals, diagrams, and editorial illustrations	/baoyu-article-illustrator article.md

---

Core Features

AI generation pipeline

HappyImage turns source material into a reproducible project folder:

• analysis.md - content analysis and style recommendations
• outline.md - page-by-page storyboard or structure
• prompts/*.md - complete image prompts with frontmatter
• copy.md - editable publishing copy
• NN-image.png - rendered image outputs

The pipeline supports batch generation, dependency-aware ordering, parallel backend calls, retries, and persistent project state.

Multi-platform publishing

Platform	Images	Title	Body	Hashtags	Method
Xiaohongshu	Up to 18	20 chars	1000 chars	Up to 10, separate lines	Chrome CDP
WeChat	Unlimited	64 chars	20000 chars	None	API or Chrome
Weibo	Up to 18	Inline 【】	2000 chars	Inline #topic#	Chrome CDP
X	Up to 4	Inline	280 chars	Inline #tag	Chrome CDP

Publishing skills automate draft creation and form filling only. You always review and confirm manually.

Session chat

Generations are tied to persistent chat sessions with SSE streaming. A session tracks messages, plans, artifacts, images, output files, and project history so you can close the browser and continue later.

Watermark and brand protection

Configure per-skill watermarks through EXTEND.md. You can customize text, position, and opacity for generated images.

---

AI Backends

Skill	Description
baoyu-imagine	Primary image generation backend. Supports OpenAI, Azure OpenAI, Google Gemini, OpenRouter, DashScope, Replicate, Z.AI, MiniMax, Jimeng, and Seedream.
baoyu-danger-gemini-web	Gemini Web image generation through browser cookies. Useful for experimentation, but web-interface behavior can change.

At least one image backend API key is required for normal image generation.

---

Utility Skills

Skill	Description
baoyu-youtube-transcript	Download YouTube transcripts/subtitles and cover images.
baoyu-url-to-markdown	Convert URLs to clean Markdown via Chrome CDP.
baoyu-danger-x-to-markdown	Convert X/Twitter threads to Markdown through a reverse-engineered API.
baoyu-compress-image	Compress and optimize images for publishing.
baoyu-format-markdown	Format Markdown with CJK typography support.
baoyu-markdown-to-html	Convert Markdown to WeChat-compatible HTML.
baoyu-translate	Translate articles in quick, normal, or refined modes.
baoyu-wechat-summary	Summarize WeChat group chats into structured digests.

Deprecated but still functional: baoyu-image-gen (use baoyu-imagine) and baoyu-xhs-images (use baoyu-image-cards).

---

Getting Started

1. Install requirements

• Bun runtime: brew install oven-sh/bun/bun or npm install -g bun
• Node.js 18+
• Google Chrome for CDP-based publishing and browser automation
• An Anthropic-compatible key for planning, captions, and iterative chat
• At least one image generation backend key

2. Clone and build

git clone https://github.com/happy-token/HappyImage.git
cd HappyImage
bun install
bun run build

3. Start an interface

# Desktop app
bun run dev:gui

# Web UI
bun run dev:web
# open http://localhost:3200

# CLI diagnostics
bun run cli -- doctor

For the full root-level command matrix for CLI, Web UI, and GUI/Desktop development, builds, packaging, and release, see Development, Build, and Release.

4. Install as an agent plugin

/plugin marketplace add happy-token/HappyImage
/plugin install baoyu-skills@happyimage-skills

You can also tell your agent: "Please install Skills from github.com/happy-token/HappyImage".

---

Environment Configuration

Add keys to .env in the project or to ~/.baoyu-skills/.env globally:

# Planning, captions, and chat
ANTHROPIC_API_KEY=sk-ant-...

# Image generation - choose at least one
OPENAI_API_KEY=sk-...
GOOGLE_API_KEY=...
DASHSCOPE_API_KEY=sk-...
ARK_API_KEY=...

# WeChat API publishing only. Other platforms use Chrome login.
WECHAT_APP_ID=wx...
WECHAT_APP_SECRET=...

Supported image backends include OpenAI, Azure OpenAI, Google Gemini, OpenRouter, DashScope, Replicate, Z.AI, MiniMax, Jimeng, and Seedream.

---

Customization

All skills support project-level and user-level customization through EXTEND.md:

.baoyu-skills/<skill-name>/EXTEND.md
~/.baoyu-skills/<skill-name>/EXTEND.md

Use it for default styles, palettes, watermarks, Chrome profiles, publishing preferences, and other skill-specific settings. Each skill documents its supported options in its own SKILL.md.

---

Repository Structure

HappyImage/
├── skills/          # Self-contained baoyu-* skills
├── packages/
│   ├── core/        # Shared runtime, settings, orchestration, AI pipeline
│   ├── web-ui/      # React + Vite frontend and Hono API server
│   ├── cli/         # happyimage CLI
│   └── desktop/     # Electron desktop wrapper
├── docs/            # Author-facing reference documentation
├── scripts/         # Repo maintenance, packaging, publishing
└── screenshots/     # Preview assets used by the gallery and docs

---

Notes

• baoyu-danger-gemini-web and baoyu-danger-x-to-markdown depend on unofficial web/API behavior and may break when upstream platforms change.
• Publishing skills automate form filling only. They do not click "Publish"; you retain final control.
• Chrome-based skills share a local browser profile. See docs/chrome-profile.md for platform paths and overrides.

---

Credits

HappyImage is built on top of baoyu-skills by JimLiu. The core AI pipeline, skill definitions, and multi-platform publishing engine are adapted from the baoyu-skills ecosystem.

Maintained by happy-token and contributors.

License

MIT

Star History