Node Stable Diffusion Web UI Client

A Node.js client for using Automatic1111's Stable Diffusion Web UI.

Current Features

This project is a work in progress. Current features include:

• txt2img generation
• img2img generation
• pngInfo support

Extensions

• ControlNet support both txt2img and img2img

Getting Started

Requirements

• Node.js >= 18
• Automatic1111's Stable Diffusion Web UI
• ControlNet support requires sd-webui-controlnet

Enable Stable Diffusion WebUI's API

In order to use the API, you first need to enable it. You can do so by appending --api to the command line arguments when running the webui.

For Linux: [webui-folder]/webui-user.sh

export COMMANDLINE_ARGS="--api"

For Windows: [webui-folder]/webui-user.bat

set COMMANDLINE_ARGS=--api

Installation

Add node-sd-webui to your project.

Using npm:

npm i -S node-sd-webui

Using yarn:

yarn add node-sd-webui

Using pnpm:

pnpm add node-sd-webui

Usage

txt2img

The txt2img function allows you to generate an image using the txt2img functionality of the Stable Diffusion WebUI.

Basic Example:

import sdwebui from 'node-sd-webui'
import { writeFileSync } from 'fs'

sdwebui()
  .txt2img({
    prompt: 'A photo of a mushroom',
  })
  .then(({ images }) => writeFileSync('path/to/image.png', images[0], 'base64'))
  .catch((err) => console.error(err))

Another example with a few more options:

import sdwebui, { SamplingMethod } from 'node-sd-webui'
import { writeFileSync } from 'fs'

const client = sdwebui()

try {
  const { images } = await client.txt2img({
    prompt: 'A photo of a mushroom, red cap, white spots',
    negativePrompt: 'blurry, cartoon, drawing, illustration',
    samplingMethod: SamplingMethod.Euler_A,
    width: 768,
    height: 512,
    steps: 20,
    batchSize: 5,
  })

  images.forEach((image, i) =>
    writeFileSync(`path/to/image-${i}.png`, images[i], 'base64')
  )
} catch (err) {
  console.error(err)
}