Official JavaScript/TypeScript client for the LiblibAI API. This SDK provides a convenient way to interact with LiblibAI's image generation services. 更多模型参数信息请参考 LiblibAI API Documentation。
# Using npm
npm install liblibai
# Using yarn
yarn add liblibai
# Using pnpm
pnpm add liblibai// JavaScript
const { LiblibAI } = require('liblibai');
// Create a client instance
const client = new LiblibAI({
apiKey: 'your-api-key',
apiSecret: 'your-api-secret',
});
// Generate an image
client.text2img({
prompt: 'a beautiful sunset',
width: 1024,
height: 1024,
})
.then(result => console.log(result))
.catch(error => console.error(error));// TypeScript
import { LiblibAI } from 'liblibai';
// Create a client instance
const client = new LiblibAI({
apiKey: 'your-api-key',
apiSecret: 'your-api-secret',
});
// Generate an image
async function generateImage() {
try {
const result = await client.text2img({
prompt: 'a beautiful sunset',
width: 1024,
height: 1024,
});
console.log(result);
} catch (error) {
console.error(error);
}
}
generateImage();Create a new LiblibAI client with your API credentials:
const client = new LiblibAI({
apiKey: 'your-api-key',
apiSecret: 'your-api-secret',
baseURL: 'https://openapi.liblibai.cloud', // Optional, defaults to https://openapi.liblibai.cloud
});Generate an image from a text prompt:
const result = await client.text2img({
"templateUuid": "e10adc3949ba59abbe56e057f20f883e",
"generateParams": {
"checkPointId": "0ea388c7eb854be3ba3c6f65aac6bfd3", // 底模 modelVersionUUID
"prompt": "Asian portrait,A young woman wearing a green baseball cap,covering one eye with her hand", // 选填
"negativePrompt": "ng_deepnegative_v1_75t,(badhandv4:1.2),EasyNegative,(worst quality:2),", //选填
"sampler": 15, // 采样方法
"steps": 20, // 采样步数
"cfgScale": 7, // 提示词引导系数
"width": 768, // 宽
"height": 1024, // 高
"imgCount": 1, // 图片数量
"randnSource": 0, // 随机种子生成器 0 cpu，1 Gpu
"seed": 2228967414, // 随机种子值，-1表示随机
"restoreFaces": 0, // 面部修复，0关闭，1开启
// Lora添加，最多5个
"additionalNetwork": [
{
"modelId": "31360f2f031b4ff6b589412a52713fcf", //LoRA的模型版本versionuuid
"weight": 0.3 // LoRA权重
},
{
"modelId": "365e700254dd40bbb90d5e78c152ec7f", //LoRA的模型版本uuid
"weight": 0.6 // LoRA权重
}
],
// 高分辨率修复
"hiResFixInfo": {
"hiresSteps": 20, // 高分辨率修复的重绘步数
"hiresDenoisingStrength": 0.75, // 高分辨率修复的重绘幅度
"upscaler": 10, // 放大算法模型枚举
"resizedWidth": 1024, // 放大后的宽度
"resizedHeight": 1536 // 放大后的高度
}
}
});Generate a new image based on an existing image and a text prompt:
const result = await client.img2img({
"templateUuid": "9c7d531dc75f476aa833b3d452b8f7ad", // 预设参数模板ID
"generateParams": {
// 基础参数
"checkPointId": "0ea388c7eb854be3ba3c6f65aac6bfd3", //底模
"prompt": "1 girl wear sunglasses", //正向提示词
"negativePrompt": //负向提示词
"clipSkip": 2, // Clip跳过层
"sampler": 15, //采样方法
"steps": 20, // 采样步数
"cfgScale": 7, // 提示词引导系数
"randnSource": 0, // 随机种子来源，0表示CPU，1表示GPU
"seed": -1, // 随机种子值，-1表示随机
"imgCount": 1, // 1到4
"restoreFaces": 0, // 面部修复，0关闭，1开启
// 图像相关参数
"sourceImage": "https://liblibai-online.liblib.cloud/img/081e9f07d9bd4c2ba090efde163518f9/7c1cc38e-522c-43fe-aca9-07d5420d743e.png",
"resizeMode": 0, // 缩放模式， 0 拉伸，1 裁剪，2 填充
"resizedWidth": 1024, // 图像缩放后的宽度
"resizedHeight": 1536, // 图像缩放后的高度
"mode": 4, // 0图生图，4局部重绘
"denoisingStrength": 0.75, // 重绘幅度
// Lora添加，最多5个
"additionalNetwork": [
{
"modelId": "31360f2f031b4ff6b589412a52713fcf", //LoRA的模型版本uuid
"weight": 0.3 // LoRA权重
},
{
"modelId": "365e700254dd40bbb90d5e78c152ec7f", //LoRA的模型版本uuid
"weight": 0.6 // LoRA权重
}
],
// 局部重绘相关参数
"inpaintParam": {
"maskImage": "https://liblibai-online.liblib.cloud/img/081e9f07d9bd4c2ba090efde163518f9/323fc358-618b-4c7d-b431-7d890209e5a5.png", // 蒙版地址
"maskBlur": 4, // 蒙版模糊度
"maskPadding": 32, //蒙版边缘预留像素，也称蒙版扩展量
"maskMode": 0, // 蒙版模式
"inpaintArea": 0, //重绘区域, 0重绘全图，1仅重绘蒙版区域
"inpaintingFill": 1 //蒙版内容的填充模式
},
// controlNet，最多4组
"controlNet": [
{
"unitOrder": 1, // 执行顺序
"sourceImage": "https://liblibai-online.liblib.cloud/img/081e9f07d9bd4c2ba090efde163518f9/7c1cc38e-522c-43fe-aca9-07d5420d743e.png",
"width": 1024, // 参考图宽度
"height": 1536, // 参考图高度
"preprocessor": 3, // 预处理器枚举值
"annotationParameters": { // 预处理器参数， 不同预处理器不同，此处仅为示意
"depthLeres": { // 3 预处理器 对应的参数
"preprocessorResolution": 1024,
"removeNear": 0,
"removeBackground": 0
}
},
"model": "6349e9dae8814084bd9c1585d335c24c", // controlnet的模型
"controlWeight": 1, // 控制权重
"startingControlStep": 0, //开始控制步数
"endingControlStep": 1, // 结束控制步数
"pixelPerfect": 1, // 完美像素
"controlMode": 0, // 控制模式 ，0 均衡，1 更注重提示词，2 更注重controlnet，
"resizeMode": 1, // 缩放模式， 0 拉伸，1 裁剪，2 填充
"maskImage": "" // 蒙版图
}
]
}
});For higher quality generations using the Ultra models:
// Text to Image Ultra
const result = await client.text2imgUltra({
"templateUuid":"5d7e67009b344550bc1aa6ccbfa1d7f4",
"generateParams":{
"prompt":"1 girl,lotus leaf,masterpiece,best quality,finely detail,highres,8k,beautiful and aesthetic,no watermark,",
"aspectRatio":"portrait",
//或者配置imageSize设置具体宽高
"imageSize": {
"width": 768,
"height": 1024
},
"imgCount":1,
"steps": 30, // 采样步数，建议30
//高级设置，可不填写
"controlnet":{
"controlType":"depth",
"controlImage": "https://liblibai-online.liblib.cloud/img/081e9f07d9bd4c2ba090efde163518f9/7c1cc38e-522c-43fe-aca9-07d5420d743e.png",
}
}
});
// Image to Image Ultra
const result = await client.img2imgUltra({
"templateUuid":"5d7e67009b344550bc1aa6ccbfa1d7f4",
"generateParams":{
"prompt":"1 girl,lotus leaf,masterpiece,best quality,finely detail,highres,8k,beautiful and aesthetic,no watermark,",
"aspectRatio":"portrait",
//或者配置imageSize设置具体宽高
"imageSize": {
"width": 768,
"height": 1024
},
"imgCount":1,
"steps": 30, // 采样步数，建议30
//高级设置，可不填写
"controlnet":{
"controlType":"depth",
"controlImage": "https://liblibai-online.liblib.cloud/img/081e9f07d9bd4c2ba090efde163518f9/7c1cc38e-522c-43fe-aca9-07d5420d743e.png",
}
}
});Run custom ComfyUI workflows:
const result = await client.runComfy({
"templateUuid": "4df2efa0f18d46dc9758803e478eb51c",
"generateParams": {
"workflowUuid": "fa2e042e32fa4aabbbacc255b4ab2cca",
"30":
{
"class_type": "LoadImage",
"inputs":
{
"image": "https://liblibai-online.liblib.cloud/img/081e9f07d9bd4c2ba090efde163518f9/5fae2d9099c208487bc97867bece2bf3d904068e307c7bd30c646c9f3059af33.png"
}
},
"31":
{
"class_type": "ImageScale",
"inputs":
{
"width": 2048,
"height": 2048
}
}
}
});The SDK provides methods to upload files that can be used in generation tasks.
Upload a file to use in image generation tasks:
// In Node.js environment using Buffer
const fs = require('fs');
const path = require('path');
// Read a file into a Buffer
const imagePath = path.join(__dirname, 'my_image.jpg');
const fileBuffer = fs.readFileSync(imagePath);
// Upload the file with a custom filename
const uploadedUrl = await client.uploadFile(fileBuffer, 'my_uploaded_image.jpg');
console.log(`Uploaded file URL: ${uploadedUrl}`);
// Use the uploaded file in an img2img task
const result = await client.img2img({
generateParams: {
prompt: "Transform this into a beautiful oil painting",
// Other parameters...
sourceImage: uploadedUrl, // Use the uploaded image URL here
...
}
});// In browser environment using File/Blob
// Assuming you have a file input element in your HTML
const fileInput = document.getElementById('fileInput');
fileInput.addEventListener('change', async (event) => {
const file = event.target.files[0];
if (!file) return;
try {
// Upload the file
const uploadedUrl = await client.uploadFile(file);
console.log(`Uploaded file URL: ${uploadedUrl}`);
// You can now use uploadedUrl in other API calls
} catch (error) {
console.error('Error uploading file:', error);
}
});Submit a generation request without waiting for the result:
const submitResult = await client.submitText2Img({
"templateUuid": "e10adc3949ba59abbe56e057f20f883e", // 参数模板ID
"generateParams": {
// 基础参数
"checkPointId": "0ea388c7eb854be3ba3c6f65aac6bfd3", // 底模 modelVersionUUID
"vaeId": "",
"prompt": "Asian portrait,A young woman wearing a green baseball cap,covering one eye with her hand", // 选填
"negativePrompt": "ng_deepnegative_v1_75t,(badhandv4:1.2),EasyNegative,(worst quality:2),nsfw", //选填
"clipSkip": 2, // 1到12，正整数值
"sampler": 15, // 采样方法
"steps": 20, // 采样步数
"cfgScale": 7, // 提示词引导系数
"width": 768, // 宽
"height": 1024, // 高
"imgCount": 1, // 图片数量
"randnSource": 0, // 随机种子生成器 0 cpu，1 Gpu
"seed": -1, // 随机种子值，-1表示随机
"restoreFaces": 0, // 面部修复，0关闭，1开启
// Lora添加，最多5个
"additionalNetwork": [],
// 高分辨率修复
"hiResFixInfo": {},
// controlNet，最多4组
"controlNet": []
}
});
const generateUuid = submitResult.generateUuid;Check the status of a generation:
let status
while (true) {
status = await client.getStatus(generateUuid);
if ([GenerateStatus.SUCCESS, GenerateStatus.FAILED, GenerateStatus.TIMEOUT].includes(status.generateStatus)) {
break;
}
await new Promise(resolve => setTimeout(resolve, 3000));
}
console.log(status)Wait for a generation to complete:
const result = await client.waitResult(generateUuid);const status = await client.getComfyStatus(generateUuid);
const result = await client.waitAppResult(generateUuid)The SDK throws LiblibAIError for API-related errors:
try {
const result = await client.text2img({
prompt: 'a beautiful sunset',
});
console.log(result);
} catch (error) {
if (error instanceof LiblibAIError) {
console.error(`API Error (${error.statusCode}): ${error.message}`);
} else {
console.error('Unexpected error:', error);
}
}The API uses the following status codes for generation requests:
enum GenerateStatus {
PENDING = 1,
PROCESSING = 2,
GENERATED = 3,
AUDITING = 4,
SUCCESS = 5,
FAILED = 6,
TIMEOUT = 7
}This SDK is distributed under the MIT License. See the LICENSE file for more information.
For issues, questions, or feedback, please open an issue on our GitHub repository.