dusrelimage

dusrelimage is one single package that does the everyday image jobs — compress, resize, crop, convert, watermark, annotate, colour picker, palette extraction, collage and pixelate — offline. No API keys, no network calls, no per-file fees, no data leaving your machine. One Rust core, with bindings for JavaScript/TypeScript, Python and Rust, plus a CLI.

These docs are in English for v0.1.0. The error messages in the package itself are bilingual: every error carries both message_en and message_hi.

Install

# JavaScript / TypeScript
npm install dusrelimage

# Python
pip install dusrelimage

# Rust
cargo add dusrelimage

Requires Node 18+, Python 3.9+, or Rust 1.85+. The Rust core is built on the image crate 0.25 for decode/encode, with text drawn via ab_glyph and a bundled DejaVu Sans font.

Your first operation

// Rust — compress a JPEG to quality 70
use dusrelimage::{compress_image, CompressOptions};

let input = std::fs::read("photo.jpg")?;
let out = compress_image(&input, &CompressOptions { quality: 70, ..Default::default() })?;
std::fs::write("photo-small.jpg", out)?;

// JavaScript / TypeScript — same operation, same Rust core (WASM).
// The npm wrapper uses camelCase names (compressImage, resizeImage…);
// the raw WASM exports keep the Rust snake_case spelling.
import { compressImage } from 'dusrelimage';

const input = new Uint8Array(await fs.readFile('photo.jpg'));
const out = compressImage(input, { quality: 70 }); // Uint8Array out

# Python — same operation, same Rust core
from dusrelimage import compress_image

with open("photo.jpg", "rb") as f:
    data = f.read()
out = compress_image(data, quality=70)  # bytes out

The 10 tools

Tool	What it does	Rust function
Compress	Shrink file size at a quality, or hit an exact target size	compress_image, compress_to_size
Resize	New width / height / percent, with fit modes and filters	resize_image
Crop	Cut a rectangle, or an anchor preset	crop_image
Convert	Change format (JPEG, PNG, WebP, AVIF, GIF, BMP)	convert_image
Watermark	Text or logo stamp, 9 anchors + tiled, rotation, opacity	add_watermark
Annotate	Arrows, rectangles, ellipses, lines and text, drawn in raster	annotate_image
Colour picker	The exact colour of one pixel, as hex + RGB + HSL	pick_colour
Palette	Dominant colours, most frequent first	extract_palette
Collage	Several images into one grid / row / column	make_collage
Pixelate	Mosaic-blur a rectangle with a block size	pixelate_image

Plus two helpers that go with them: get_image_info (format, dimensions, alpha, colour type, file size) and bulk_process (many inputs in one call, one bad input never kills the whole batch).

How it works

Every operation takes bytes in and returns bytes out. There is no streaming mode and no temp files: the whole image lives in memory, the operation runs, and you get the result back as bytes. That makes the API the same in every language — read a file, call the function, write the result.

Errors never panic on bad input. Every failure returns a stable, machine-readable error code (INVALID_IMAGE, CROP_OUT_OF_BOUNDS, …) plus a message in English and Hindi. See Error codes.

Honest limits — read this first

v0.1.0 has no HEIC support (libheif + licence), AVIF input does not decode (UNSUPPORTED_FORMAT — AVIF is encode-only), AVIF encode is lossy-only / 8-bit, and GIF encode is first-frame only for animated input. The whole image lives in memory; there is no streaming mode. Images larger than 120,000,000 pixels are refused with IMAGE_TOO_LARGE. Full list is on each tool page and in the changelog.

Try it in your browser

The playground runs the real Rust core compiled to WebAssembly. Pick a local image, run a tool, download the result — your files never leave the browser.