Five free design tools that run entirely in your browser. Your files never leave your device — every pixel is processed locally using WebGL2, on-device AI, and multi-threaded workers. Built by Rhys Muirhead.
Background Remover
Remove any image background with Lucida, a state-of-the-art open AI matting model (MIT licence) — exceptionally good on hair, glass, text and tricky camouflage. Everything runs on your own device: no uploads, no third-party servers. Please note: the model is large (about 450 MB) — it downloads once on first use and is then cached by your browser, so later visits are instant. It needs a WebGPU-capable browser (current Chrome or Edge on desktop; recent Safari and Firefox also work) and the first run per visit can take a little while to warm up. The Strength slider fine-tunes the cut instantly without re-running the model: lower keeps more (softer edges), higher cuts harder. Model: Lucida by egeorcun (MIT), browser conversion by tomohisa — self-hosted here.
1 Add an image
Click to choose an image or drag & drop it here
2 Fine-tune
3 Export
Lucida is stored in this browser's site storage (an IndexedDB database named
lucida-model-cache, plus temporary browser cache) — about 450 MB, on this device only.
It is never uploaded anywhere and stays until you clear site data.
Point Cloud Generator
Turn text, any image, or a 3D model into a rotating point cloud, rendered live on the GPU with WebGL2. Type a word, drop an image (optionally with its background removed by on-device AI), or load a PLY / OBJ / STL / GLB model — including GLB vertex colours, materials and embedded textures — sampled across its surface or filled solid — then export the cloud as a PLY file for Blender, MeshLab, CloudCompare or any 3D tool.
1 Choose a source — text, image, or 3D model
Click to choose an image
or drag & drop it here
Click to choose a model
PLY · OBJ · STL · GLB
2 Tune the cloud
The cloud rotates automatically. PLY export matches the format used by this site's homepage hero (x y z r g b).
3 Export
LumiSpace
A Lightroom-style darkroom running on your GPU. Exposure, contrast, temperature, vibrance, highlights & shadows, eight-range HSL, a full curves editor, and local radial/linear gradient adjustments. Tip: double-click any slider to reset it to zero.
1 Add a photo
Click to choose a photo or drag & drop it here
2 Edit in the panels below
3 Export
Local adjustments: add a mask below, then drag on the photo — drag inside a mask to move it, drag outside to redraw the selected mask.
Color Palette Extractor
Pull a perceptually accurate palette out of any image — clustering happens in OKLCH space, so the swatches match how the image actually feels. Every value is selectable, and clicking any swatch or chip copies its hex code. The contrast matrix tells you which colour pairs are safe for text.
1 Add an image
Click to choose an image or drag & drop it here
2 Palette size
3 Export
Palette — click a swatch to make it the harmony base; use the ⧉ button or hex field to copy
Harmonies (from dominant colour) — click a chip to copy
WCAG 2.2 contrast matrix
Each cell is the contrast ratio between the row colour (as text) and the column colour (as background), e.g. 4.52 means 4.52:1. The badge is the WCAG 2.2 conformance level that pair meets: AAA ≥ 7:1 (best) · AA ≥ 4.5:1 (normal text) · AA-lg ≥ 3:1 (large text / UI only) · ✗ below 3:1 (avoid for text).
Batch Image Processor
Crop, resize, and convert whole folders of images at once — including next-gen WebP and AVIF output. Queue your images, frame each crop visually, type exact quality and size values, then let a pool of worker threads do the heavy lifting without freezing the page.
1 Add images
Click to choose images or drag & drop a whole batch here
2 Output settings
3 Run & export
Drag inside the frame to move it · drag a corner handle to resize · double-click the image to select the full frame