API Documentation

Introduction

Pixly's processing engine is built on WebAssembly (WASM) and browser-native APIs (Canvas, Web Workers, OffscreenCanvas). Rather than a traditional REST API, Pixly provides a client-side SDK that runs all image processing in the browser — no server roundtrips, no data uploads, complete privacy.

The SDK is available as an npm package and can also be loaded via CDN. It's compatible with all modern browsers that support WebAssembly and Web Workers.

Authentication

The Pixly SDK is completely free and does not require an API key. All tools — including AI-powered ones — are available without authentication. Simply install the SDK and start processing images.

// Initialize the SDK with your API key
import Pixly from '@pixly/sdk';

const pixly = new Pixly({
  apiKey: 'px_live_your_api_key_here',
  workerCount: 4, // number of Web Workers
});

Rate Limits

Since processing happens client-side, there are no traditional server-side rate limits. However, the following limits apply to API key validation and model downloads:

• All tools: Unlimited use of every tool, including AI-powered tools. No daily limits or usage caps.
• Model downloads: AI models are cached after first load (~50MB total) for instant subsequent use.
• No API key required: Everything runs client-side with no server validation needed.

API Endpoints (SDK Methods)

All methods accept a File, Blob, or ImageData object and return a Promise<Blob>.

const result = await pixly.removeBackground(imageFile, {
  model: 'u2net', // 'u2net' | 'u2netp' (lite)
  outputFormat: 'png',
  outputQuality: 0.92,
  edgeRefine: true,
});

// result is a Blob containing the processed PNG
const url = URL.createObjectURL(result);

const result = await pixly.upscale(imageFile, {
  scale: 4, // 2 or 4
  model: 'realesrgan-x4plus',
  tileMode: 'auto', // splits large images into tiles
  outputFormat: 'png',
});

const result = await pixly.compress(imageFile, {
  format: 'webp',
  quality: 80, // 1-100
  effort: 4, // encoding effort (1-6, higher = smaller file)
  keepMetadata: false,
});

const result = await pixly.convert(imageFile, {
  format: 'avif',
  quality: 85,
});

const results = await pixly.batchProcess(imageFiles, {
  operation: 'compress',
  format: 'webp',
  quality: 80,
  onProgress: (index, total) => {
    console.log(`Processed ${index}/${total}`);
  },
});

const meta = await pixly.getMetadata(imageFile);
// { width: 1920, height: 1080, format: 'jpeg',
//   exif: { camera: '...', gps: null, ... },
//   colorSpace: 'srgb', size: 245678 }

Error Handling

All SDK methods throw typed errors. Catch them with try/catch or .catch():

try {
  const result = await pixly.removeBackground(file);
} catch (err) {
  if (err.code === 'PIXLY_INVALID_KEY') {
    // API key is invalid or expired
  } else if (err.code === 'PIXLY_MODEL_LOAD_FAILED') {
    // ML model failed to download or initialize
  } else if (err.code === 'PIXLY_UNSUPPORTED_FORMAT') {
    // Image format not supported
  } else {
    // Other processing error
  }
}

SDKs & Libraries

The Pixly SDK is available for multiple platforms:

• JavaScript / TypeScript: npm install @pixly/sdk — Full SDK with all tools
• React: npm install @pixly/react — React hooks and components
• Vue: npm install @pixly/vue — Vue 3 composables
• CDN: https://cdn.pixly.app/sdk/v2/pixly.min.js — UMD bundle

Webhooks

Since Pixly processes images entirely client-side, there are no server-side webhooks. All processing results are returned directly to the calling code. For server-side workflows, consider running the Pixly WASM modules in a Node.js or Deno environment using the @pixly/node package.