Chainable image processing for Node.js and the browser — with AI ops built in.

imgcraft converts, compresses, resizes, crops, removes backgrounds, upscales images and more — via a fluent chainable API. Runs natively in Node.js using sharp, and in the browser via WebAssembly. A hosted REST API is also available at https://imgcraft-api.imgcraft.workers.dev/docs.

Installation

npm install imgcraft

Examples

import { img, batch } from 'imgcraft'

// Compress — 3.5MB JPEG → 60KB WebP in one line
const buffer = await img('photo.jpg')
  .compress({ quality: 85 })
  .toBuffer()

// Bulk compress an entire folder
await batch(['hero.jpg', 'banner.jpg', 'thumb.jpg'])
  .compress({ quality: 80 })
  .toDir('./optimized')

// Resize and convert
const buffer = await img('photo.jpg')
  .resize(800, 600)
  .webp({ quality: 85 })
  .toBuffer()

// Remove background (no API key needed — runs locally)
const buffer = await img('portrait.jpg')
  .removeBackground()
  .png()
  .toBuffer()

// Chain multiple transforms
const buffer = await img('input.png')
  .resize(1200)
  .sharpen()
  .removeBackground()
  .upscale(2)
  .compress({ format: 'avif', quality: 60 })
  .toBuffer()

// Batch process with concurrency control
await batch(['a.jpg', 'b.jpg', 'c.jpg'], { concurrency: 4 })
  .resize(800)
  .compress()
  .toDir('./output')

// Browser (WASM — same API)
const result = await img(file)
  .resize(400)
  .compress({ quality: 80 })
  .toDataURL()

// Read metadata
const meta = await img('photo.jpg').meta()
// { width: 3024, height: 4032, format: 'jpeg', size: 3621944, hasAlpha: false }

// REST API
const form = new FormData()
form.append('image', file)
form.append('ops', JSON.stringify([
  { op: 'resize', options: { width: 800 } },
  { op: 'compress', options: { quality: 85 } }
]))
const res = await fetch('https://imgcraft-api.imgcraft.workers.dev/transform', {
  method: 'POST', body: form
})
const blob = await res.blob()

Supported operations

Compression & Format Conversion

• compress(options?) — compress to WebP/AVIF/JPEG/PNG (default: WebP q80)
• format(type) — convert between jpeg, png, webp, avif
• jpeg(options?) — JPEG with quality control
• png(options?) — PNG with quality control
• webp(options?) — WebP with quality control
• avif(options?) — AVIF with quality control (best compression ratio)
• quality(1–100) — set output quality

Transforms

• resize(width?, height?, options?) — fit modes: cover, contain, fill, inside, outside
• crop(left, top, width, height) — extract a region
• rotate(angle) — 0, 90, 180, 270 degrees
• flip() — vertical flip
• flop() — horizontal flip

Colour

• grayscale() — convert to greyscale
• tint(colour) — apply colour tint
• negate() — invert colours
• brightness(factor) — adjust brightness
• contrast(factor) — adjust contrast
• saturation(factor) — adjust saturation

Filters

• blur(sigma?) — Gaussian blur
• sharpen(options?) — unsharp mask
• median(size?) — median filter

Compositing

• composite(input, options?) — overlay/watermark (base64 data URI input)

Metadata

• meta() — read width, height, format, size, hasAlpha, channels
• stripMeta() — remove EXIF and ICC data

AI operations (Node.js only — lazy loaded, no API key required)

• removeBackground() — subject isolation via ONNX model
• smartCrop(options?) — subject-aware crop via TensorFlow coco-ssd
• upscale(2 | 4) — ESRGAN super-resolution upscaling

Output

• toBuffer() — Buffer (Node.js) / Uint8Array (browser)
• toFile(path) — write to disk (Node.js only)
• toStream() — ReadableStream<Uint8Array>
• toDataURL() — base64 data URI (browser)

Batch

• batch(inputs[], options?) — process multiple images
• .toBuffers() — array of results
• .toDir(path) — write all to directory
• .toFiles(paths[]) — 1:1 path mapping

REST API

A hosted Cloudflare Worker proxies requests to the Node.js processing server. No SDK required.

Base URL: https://imgcraft-api.imgcraft.workers.dev

| Method | Path | Description | |---|---|---| | POST | /transform | Process an image, returns transformed bytes | | GET | /info | Read metadata without processing | | GET | /health | Status check |

Try it live →

Rate limits: 10 requests/minute, 50 requests/day per IP. AI operations: 10/day.

Interactive API Reference: https://imgcraft-api.imgcraft.workers.dev/docs

curl -X POST https://imgcraft-api.imgcraft.workers.dev/transform \
  -F "[email protected]" \
  -F 'ops=[{"op":"compress","options":{"quality":85}}]' \
  --output result.webp

Full REST API reference: imgcraft-docs.vercel.app/docs/rest-api

Documentation

imgcraft-docs.vercel.app — getting started, full API reference, live playground.

License

MIT