io.github.rocnubie/z-image-mcp

Official

by rocnubie · JavaScript

Z-Image MCP

Z-Image is an online interface for the Z-Image generation model with fast inference and prompt

Generates images from text prompts using the Z-Image model with fast inference.

README.md6 KBView on GitHub

Z-Image MCP Server

Z-Image - Z-Image AI Image Generator

A Model Context Protocol server that exposes the canonical Z-Image knowledge surface — image generation workflows and styles, pricing, FAQ, official links — to MCP-compatible AI clients such as Claude Desktop, Cursor, Windsurf, and Continue. Read-only, no API keys, no quota, ~50 ms cold start.

Official website: https://z-image.club

🎨 About Z-Image

Z-Image (z-image.club) is a free online AI image generation and editing platform built around a 6-billion-parameter diffusion model. It lets users produce photorealistic images from text prompts, modify existing photos through guided editing, swap faces, remove backgrounds, and even convert still images into short videos — all through a browser-based interface that requires no local installation. For users who prefer running models locally, the weights are available on Hugging Face and ModelScope, with native ComfyUI workflow support and GGUF/FP8 quantized variants that fit on consumer GPUs with under 16 GB of VRAM.

Key Features

Multiple generation modes — Z-Image Turbo delivers fast results in 8 steps, while Z-Image-Edit focuses on precise image modification; both are accessible from the same dashboard.
Face swap and portrait tools — Replace faces in target images while maintaining lighting and composition consistency; dedicated controls for hairstyle changes and portrait stylization.
Image-to-image and image-to-video conversion — Transform existing photos using text-guided prompts or animate them into short cinematic clips.
Photo editing suite — Background removal and replacement, text removal, quality upscaling, and a general photo enhancer are bundled without requiring a separate tool.
Bilingual text rendering — Accurate placement of Chinese and English text inside generated visuals, a capability that many comparable models handle poorly.
Inspiration gallery and prompt enhancer — A curated gallery of trend prompts with detailed examples, paired with a structured-reasoning prompt enhancer that refines vague inputs into precise generation instructions.

Use Cases

Social media content — Generating styled portraits, editorial-look photos, or branded graphics for posts without needing a photographer or designer.
E-commerce product photography — Producing clean, professional product shots against custom backgrounds at a fraction of a traditional shoot's cost.
Rapid design prototyping — Using image-to-image transformation to iterate on visual concepts quickly before committing to final artwork.
Local model experimentation — Downloading quantized weights and integrating them into ComfyUI pipelines for custom, offline workflows.
Marketing asset creation — Generating diverse visual variations for campaigns — different styles, backgrounds, or model looks — in minutes rather than days.

Who Is It For

Z-Image serves a broad range of users who need visual output without deep technical expertise or large budgets. Content creators and social media managers will find the web demo approachable, with one-click tools covering most common tasks. Graphic designers and art directors can use it for fast concept iteration, while e-commerce teams can cut product photography costs by generating studio-quality shots on demand. Developers and ML practitioners are catered to through open weights, ComfyUI compatibility, and multiple quantization options, making it practical to embed the model into existing pipelines. The combination of a free hosted interface and downloadable model weights means both beginners and experienced practitioners can find a workflow that fits their setup.

Tools

`list_styles`

Return the canonical list of image-generation styles or presets the site exposes. (Z-Image)

Input: no parameters. Returns: text/markdown.

`get_pricing`

Return the canonical pricing entry point for Z-Image.

Input: no parameters. Returns: text/markdown.

`get_official_links`

Return the canonical list of official links for Z-Image (website, support, docs when available).

Input: no parameters. Returns: text/markdown.

Resources

site://z-image/styles — Supported image-generation styles and presets.
site://z-image/pricing — Canonical pricing entry point.
site://z-image/faq — Short FAQ generated from public site metadata.
site://z-image/links — Canonical URLs to share with users.

Prompts

`tell_me_about_z_image`

Summarize what the site is, who it's for, and how it works. — Z-Image

`try_image_style_z_image`

Recommend a starting image-generation style for a stated goal. — Z-Image

Installation

Install via Smithery

bash

npx -y @smithery/cli install z-image-mcp --client claude

(Replace claude with cursor, windsurf, or continue for those clients.)

Install from source

bash

git clone https://github.com/rocnubie/z-image-mcp.git
cd z-image-mcp
pnpm install

Then add to your MCP client config (claude_desktop_config.json for Claude Desktop, mcp.json for Cursor / Windsurf / Continue):

json

{
  "mcpServers": {
    "z-image-mcp": {
      "command": "node",
      "args": [
        "/absolute/path/to/z-image-mcp/src/index.mjs"
      ]
    }
  }
}

Debug with MCP Inspector

bash

npx @modelcontextprotocol/inspector node src/index.mjs

Official Links

Development

bash

pnpm install
pnpm start                 # run the server over stdio

License

MIT